modelId
stringlengths 5
139
| author
stringlengths 2
42
| last_modified
timestamp[us, tz=UTC]date 2020-02-15 11:33:14
2025-08-30 06:27:36
| downloads
int64 0
223M
| likes
int64 0
11.7k
| library_name
stringclasses 527
values | tags
listlengths 1
4.05k
| pipeline_tag
stringclasses 55
values | createdAt
timestamp[us, tz=UTC]date 2022-03-02 23:29:04
2025-08-30 06:27:12
| card
stringlengths 11
1.01M
|
|---|---|---|---|---|---|---|---|---|---|
WizardLMTeam/WizardMath-70B-V1.0
|
WizardLMTeam
| 2023-12-20T03:08:12Z | 207 | 119 |
transformers
|
[
"transformers",
"pytorch",
"llama",
"text-generation",
"arxiv:2304.12244",
"arxiv:2306.08568",
"arxiv:2308.09583",
"license:llama2",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2023-08-11T04:33:24Z |
---
license: llama2
---
## WizardMath: Empowering Mathematical Reasoning for Large Language Models via Reinforced Evol-Instruct (RLEIF)
<p style="font-size:28px;" align="center">
🏠 <a href="https://wizardlm.github.io/" target="_blank">Home Page</a> </p>
<p align="center">
<p align="center">
🤗 <a href="https://huggingface.co/WizardLM" target="_blank">HF Repo</a> •🐱 <a href="https://github.com/nlpxucan/WizardLM" target="_blank">Github Repo</a> • 🐦 <a href="https://twitter.com/WizardLM_AI" target="_blank">Twitter</a> </p>
<p align="center">
📃 <a href="https://arxiv.org/abs/2304.12244" target="_blank">[WizardLM]</a> • 📃 <a href="https://arxiv.org/abs/2306.08568" target="_blank">[WizardCoder]</a> • 📃 <a href="https://arxiv.org/abs/2308.09583" target="_blank">[WizardMath]</a> <br>
</p>
<p align="center">
👋 Join our <a href="https://discord.gg/VZjjHtWrKs" target="_blank">Discord</a>
</p>
## News
[12/19/2023] 🔥 We released **WizardMath-7B-V1.1** trained from Mistral-7B, the **SOTA 7B math LLM**, achieves **83.2 pass@1** on GSM8k, and **33.0 pass@1** on MATH.
[12/19/2023] 🔥 **WizardMath-7B-V1.1** outperforms **ChatGPT 3.5**, **Gemini Pro**, **Mixtral MOE**, and **Claude Instant** on GSM8K pass@1.
[12/19/2023] 🔥 **WizardMath-7B-V1.1** is comparable with **ChatGPT 3.5**, **Gemini Pro**, and surpasses **Mixtral MOE** on MATH pass@1.
| Model | Checkpoint | Paper | GSM8k | MATH |
| ----- |------| ---- |------|-------|
| **WizardMath-7B-V1.1** | 🤗 <a href="https://huggingface.co/WizardLM/WizardMath-7B-V1.1" target="_blank">HF Link</a> | 📃 <a href="https://arxiv.org/abs/2308.09583" target="_blank">[WizardMath]</a>| **83.2** | **33.0** |
| WizardMath-70B-V1.0 | 🤗 <a href="https://huggingface.co/WizardLM/WizardMath-70B-V1.0" target="_blank">HF Link</a> | 📃 <a href="https://arxiv.org/abs/2308.09583" target="_blank">[WizardMath]</a>| **81.6** | **22.7** |
| WizardMath-13B-V1.0 | 🤗 <a href="https://huggingface.co/WizardLM/WizardMath-13B-V1.0" target="_blank">HF Link</a> | 📃 <a href="https://arxiv.org/abs/2308.09583" target="_blank">[WizardMath]</a>| **63.9** | **14.0** |
| WizardMath-7B-V1.0 | 🤗 <a href="https://huggingface.co/WizardLM/WizardMath-7B-V1.0" target="_blank">HF Link</a> | 📃 <a href="https://arxiv.org/abs/2308.09583" target="_blank">[WizardMath]</a>| **54.9** | **10.7** |
## [12/19/2023] Comparing WizardMath-7B-V1.1 with other open source 7B size math LLMs.
| Model | GSM8k Pass@1 | MATH Pass@1 |
| ----- |------| ---- |
| MPT-7B | 6.8 | 3.0 |
|Llama 1-7B | 11.0 | 2.9 |
|Llama 2-7B|12.3 |2.8 |
|Yi-6b| 32.6 |5.8 |
|Mistral-7B|37.8 |9.1 |
|Qwen-7b|47.8 |9.3 |
| RFT-7B | 50.3 | -- |
| MAmmoTH-7B (COT) | 50.5 | 10.4 |
| WizardMath-7B-V1.0 | 54.9 | 10.7 |
|Abel-7B-001 |59.7 |13 |
| MetaMath-7B | 66.5 | 19.8 |
| Arithmo-Mistral-7B | 74.7 | 25.3 |
|MetaMath-Mistral-7B|77.7 |28.2 |
|Abel-7B-002 | 80.4 | 29.5 |
| **WizardMath-7B-V1.1** | **83.2** | **33.0** |
## [12/19/2023] Comparing WizardMath-7B-V1.1 with large open source (30B~70B) LLMs.
| Model | GSM8k Pass@1 | MATH Pass@1 |
| ----- |------| ---- |
| Llemma-34B | 51.5 | 25.0 |
| Minerva-62B | 52.4 | 27.6 |
| Llama 2-70B | 56.8 | 13.5 |
| DeepSeek 67B | 63.4 | -- |
| Gork 33B | 62.9 | 23.9 |
| MAmmoTH-70B | 72.4 | 21.1 |
| Yi-34B | 67.9 | 15.9 |
| Mixtral 8x7B | 74.4 | 28.4 |
| MetaMath-70B | 82.3 | 26.6 |
| **WizardMath-7B-V1.1** | **83.2** | **33.0** |
## ❗ Data Contamination Check:
Before model training, we carefully and rigorously checked all the training data, and used multiple deduplication methods to verify and prevent data leakage on GSM8k and MATH test set.
<font size=4>
| <sup>Model</sup> | <sup>Checkpoint</sup> | <sup>Paper</sup> |<sup>MT-Bench</sup> | <sup>AlpacaEval</sup> | <sup>GSM8k</sup> | <sup>HumanEval</sup> | <sup>License</sup>|
| ----- |------| ---- |------|-------| ----- | ----- | ----- |
| <sup>**WizardLM-70B-V1.0**</sup> | <sup>🤗 <a href="https://huggingface.co/WizardLM/WizardLM-70B-V1.0" target="_blank">HF Link</a> </sup>|<sup>📃**Coming Soon**</sup>| <sup>**7.78**</sup> | <sup>**92.91%**</sup> |<sup>**77.6%**</sup> | <sup> **50.6 pass@1**</sup>|<sup> <a href="https://ai.meta.com/resources/models-and-libraries/llama-downloads/" target="_blank">Llama 2 License </a></sup> |
| <sup>WizardLM-13B-V1.2</sup> | <sup>🤗 <a href="https://huggingface.co/WizardLM/WizardLM-13B-V1.2" target="_blank">HF Link</a> </sup>| | <sup>7.06</sup> | <sup>89.17%</sup> |<sup>55.3%</sup> | <sup>36.6 pass@1</sup>|<sup> <a href="https://ai.meta.com/resources/models-and-libraries/llama-downloads/" target="_blank">Llama 2 License </a></sup> |
| <sup>WizardLM-13B-V1.1</sup> |<sup> 🤗 <a href="https://huggingface.co/WizardLM/WizardLM-13B-V1.1" target="_blank">HF Link</a> </sup> | | <sup>6.76</sup> |<sup>86.32%</sup> | | <sup>25.0 pass@1</sup>| <sup>Non-commercial</sup>|
| <sup>WizardLM-30B-V1.0</sup> | <sup>🤗 <a href="https://huggingface.co/WizardLM/WizardLM-30B-V1.0" target="_blank">HF Link</a></sup> | | <sup>7.01</sup> | | | <sup>37.8 pass@1</sup>| <sup>Non-commercial</sup> |
| <sup>WizardLM-13B-V1.0</sup> | <sup>🤗 <a href="https://huggingface.co/WizardLM/WizardLM-13B-V1.0" target="_blank">HF Link</a> </sup> | | <sup>6.35</sup> | <sup>75.31%</sup> | | <sup> 24.0 pass@1 </sup> | <sup>Non-commercial</sup>|
| <sup>WizardLM-7B-V1.0 </sup>| <sup>🤗 <a href="https://huggingface.co/WizardLM/WizardLM-7B-V1.0" target="_blank">HF Link</a> </sup> |<sup> 📃 <a href="https://arxiv.org/abs/2304.12244" target="_blank">[WizardLM]</a> </sup>| | | |<sup>19.1 pass@1 </sup>|<sup> Non-commercial</sup>|
</font>
| Model | Checkpoint | Paper | HumanEval | MBPP | Demo | License |
| ----- |------| ---- |------|-------| ----- | ----- |
| WizardCoder-Python-34B-V1.0 | 🤗 <a href="https://huggingface.co/WizardLM/WizardCoder-Python-34B-V1.0" target="_blank">HF Link</a> | 📃 <a href="https://arxiv.org/abs/2306.08568" target="_blank">[WizardCoder]</a> | 73.2 | 61.2 | [Demo](http://47.103.63.15:50085/) | <a href="https://ai.meta.com/resources/models-and-libraries/llama-downloads/" target="_blank">Llama2</a> |
| WizardCoder-15B-V1.0 | 🤗 <a href="https://huggingface.co/WizardLM/WizardCoder-15B-V1.0" target="_blank">HF Link</a> | 📃 <a href="https://arxiv.org/abs/2306.08568" target="_blank">[WizardCoder]</a> | 59.8 |50.6 | -- | <a href="https://huggingface.co/spaces/bigcode/bigcode-model-license-agreement" target="_blank">OpenRAIL-M</a> |
| WizardCoder-Python-13B-V1.0 | 🤗 <a href="https://huggingface.co/WizardLM/WizardCoder-Python-13B-V1.0" target="_blank">HF Link</a> | 📃 <a href="https://arxiv.org/abs/2306.08568" target="_blank">[WizardCoder]</a> | 64.0 | 55.6 | -- | <a href="https://ai.meta.com/resources/models-and-libraries/llama-downloads/" target="_blank">Llama2</a> |
| WizardCoder-Python-7B-V1.0 | 🤗 <a href="https://huggingface.co/WizardLM/WizardCoder-Python-7B-V1.0" target="_blank">HF Link</a> | 📃 <a href="https://arxiv.org/abs/2306.08568" target="_blank">[WizardCoder]</a> | 55.5 | 51.6 | [Demo](http://47.103.63.15:50088/) | <a href="https://ai.meta.com/resources/models-and-libraries/llama-downloads/" target="_blank">Llama2</a> |
| WizardCoder-3B-V1.0 | 🤗 <a href="https://huggingface.co/WizardLM/WizardCoder-3B-V1.0" target="_blank">HF Link</a> | 📃 <a href="https://arxiv.org/abs/2306.08568" target="_blank">[WizardCoder]</a> | 34.8 |37.4 | -- | <a href="https://huggingface.co/spaces/bigcode/bigcode-model-license-agreement" target="_blank">OpenRAIL-M</a> |
| WizardCoder-1B-V1.0 | 🤗 <a href="https://huggingface.co/WizardLM/WizardCoder-1B-V1.0" target="_blank">HF Link</a> | 📃 <a href="https://arxiv.org/abs/2306.08568" target="_blank">[WizardCoder]</a> | 23.8 |28.6 | -- | <a href="https://huggingface.co/spaces/bigcode/bigcode-model-license-agreement" target="_blank">OpenRAIL-M</a> |
**Github Repo**: https://github.com/nlpxucan/WizardLM/tree/main/WizardMath
**Twitter**: https://twitter.com/WizardLM_AI/status/1689998428200112128
**Discord**: https://discord.gg/VZjjHtWrKs
## Comparing WizardMath-V1.0 with Other LLMs.
🔥 The following figure shows that our **WizardMath-70B-V1.0 attains the fifth position in this benchmark**, surpassing ChatGPT (81.6 vs. 80.8) , Claude Instant (81.6 vs. 80.9), PaLM 2 540B (81.6 vs. 80.7).
<p align="center" width="100%">
<a ><img src="https://raw.githubusercontent.com/nlpxucan/WizardLM/main/WizardMath/images/wizardmath_gsm8k.png" alt="WizardMath" style="width: 96%; min-width: 300px; display: block; margin: auto;"></a>
</p>
❗<b>Note for model system prompts usage:</b>
Please use **the same systems prompts strictly** with us, and we do not guarantee the accuracy of the **quantified versions**.
**Default version:**
```
"Below is an instruction that describes a task. Write a response that appropriately completes the request.\n\n### Instruction:\n{instruction}\n\n### Response:"
```
**CoT Version:** （❗For the **simple** math questions, we do NOT recommend to use the CoT prompt.）
```
"Below is an instruction that describes a task. Write a response that appropriately completes the request.\n\n### Instruction:\n{instruction}\n\n### Response: Let's think step by step."
```
## Inference WizardMath Demo Script
We provide the WizardMath inference demo code [here](https://github.com/nlpxucan/WizardLM/tree/main/demo).
❗<b>To commen concern about dataset:</b>
Recently, there have been clear changes in the open-source policy and regulations of our overall organization's code, data, and models.
Despite this, we have still worked hard to obtain opening the weights of the model first, but the data involves stricter auditing and is in review with our legal team .
Our researchers have no authority to publicly release them without authorization.
Thank you for your understanding.
## Citation
Please cite the repo if you use the data, method or code in this repo.
```
@article{luo2023wizardmath,
title={WizardMath: Empowering Mathematical Reasoning for Large Language Models via Reinforced Evol-Instruct},
author={Luo, Haipeng and Sun, Qingfeng and Xu, Can and Zhao, Pu and Lou, Jianguang and Tao, Chongyang and Geng, Xiubo and Lin, Qingwei and Chen, Shifeng and Zhang, Dongmei},
journal={arXiv preprint arXiv:2308.09583},
year={2023}
}
```
|
Riad/A2C
|
Riad
| 2023-12-20T02:53:03Z | 0 | 0 |
stable-baselines3
|
[
"stable-baselines3",
"FetchPickAndPlaceDense-v2",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-12-20T00:28:33Z |
---
library_name: stable-baselines3
tags:
- FetchPickAndPlaceDense-v2
- deep-reinforcement-learning
- reinforcement-learning
- stable-baselines3
model-index:
- name: A2C
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: FetchPickAndPlaceDense-v2
type: FetchPickAndPlaceDense-v2
metrics:
- type: mean_reward
value: -8.35 +/- 5.41
name: mean_reward
verified: false
---
# **A2C** Agent playing **FetchPickAndPlaceDense-v2**
This is a trained model of a **A2C** agent playing **FetchPickAndPlaceDense-v2**
using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3).
## Usage (with Stable-baselines3)
TODO: Add your code
```python
from stable_baselines3 import ...
from huggingface_sb3 import load_from_hub
...
```
|
Riad/A2D
|
Riad
| 2023-12-20T02:49:07Z | 0 | 0 |
stable-baselines3
|
[
"stable-baselines3",
"FetchPickAndPlaceDense-v2",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-12-20T02:49:00Z |
---
library_name: stable-baselines3
tags:
- FetchPickAndPlaceDense-v2
- deep-reinforcement-learning
- reinforcement-learning
- stable-baselines3
model-index:
- name: A2C
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: FetchPickAndPlaceDense-v2
type: FetchPickAndPlaceDense-v2
metrics:
- type: mean_reward
value: -12.01 +/- 6.17
name: mean_reward
verified: false
---
# **A2C** Agent playing **FetchPickAndPlaceDense-v2**
This is a trained model of a **A2C** agent playing **FetchPickAndPlaceDense-v2**
using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3).
## Usage (with Stable-baselines3)
TODO: Add your code
```python
from stable_baselines3 import ...
from huggingface_sb3 import load_from_hub
...
```
|
justinwangx/vicuna-robust-sft-lora
|
justinwangx
| 2023-12-20T02:41:41Z | 5 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"llama",
"text-generation",
"generated_from_trainer",
"conversational",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2023-12-20T02:05:21Z |
---
tags:
- generated_from_trainer
model-index:
- name: vicuna-robust-sft-lora
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# vicuna-robust-sft-lora
This model was trained from scratch on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 1.8328
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 4
- eval_batch_size: 8
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- gradient_accumulation_steps: 128
- total_train_batch_size: 2048
- total_eval_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- num_epochs: 2
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| No log | 0 | 0 | 1.8516 |
| No log | 0 | 0 | 1.8678 |
### Framework versions
- Transformers 4.35.0
- Pytorch 2.1.0a0+32f93b1
- Datasets 2.14.6
- Tokenizers 0.14.1
|
baltop/mistral-sql-finetune
|
baltop
| 2023-12-20T02:41:30Z | 10 | 0 |
peft
|
[
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mistralai/Mistral-7B-Instruct-v0.2",
"base_model:adapter:mistralai/Mistral-7B-Instruct-v0.2",
"license:apache-2.0",
"region:us"
] | null | 2023-12-19T01:46:25Z |
---
license: apache-2.0
library_name: peft
tags:
- generated_from_trainer
base_model: mistralai/Mistral-7B-Instruct-v0.2
model-index:
- name: mistral-sql-finetune
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# mistral-sql-finetune
This model is a fine-tuned version of [mistralai/Mistral-7B-Instruct-v0.2](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0307
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2.5e-05
- train_batch_size: 2
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 1
- training_steps: 300
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 1.1806 | 0.17 | 25 | 0.3104 |
| 0.1789 | 0.33 | 50 | 0.1074 |
| 0.0665 | 0.5 | 75 | 0.0420 |
| 0.0444 | 0.67 | 100 | 0.0414 |
| 0.05 | 0.83 | 125 | 0.0351 |
| 0.0322 | 1.0 | 150 | 0.0311 |
| 0.0361 | 1.17 | 175 | 0.0343 |
| 0.0338 | 1.33 | 200 | 0.0319 |
| 0.039 | 1.5 | 225 | 0.0322 |
| 0.0324 | 1.67 | 250 | 0.0304 |
| 0.0392 | 1.83 | 275 | 0.0309 |
| 0.0305 | 2.0 | 300 | 0.0307 |
### Framework versions
- PEFT 0.7.2.dev0
- Transformers 4.37.0.dev0
- Pytorch 2.1.2
- Datasets 2.15.0
- Tokenizers 0.15.0
|
hkivancoral/smids_10x_deit_tiny_adamax_001_fold2
|
hkivancoral
| 2023-12-20T02:40:14Z | 5 | 0 |
transformers
|
[
"transformers",
"pytorch",
"vit",
"image-classification",
"generated_from_trainer",
"dataset:imagefolder",
"base_model:facebook/deit-tiny-patch16-224",
"base_model:finetune:facebook/deit-tiny-patch16-224",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
image-classification
| 2023-12-20T00:35:11Z |
---
license: apache-2.0
base_model: facebook/deit-tiny-patch16-224
tags:
- generated_from_trainer
datasets:
- imagefolder
metrics:
- accuracy
model-index:
- name: smids_10x_deit_tiny_adamax_001_fold2
results:
- task:
name: Image Classification
type: image-classification
dataset:
name: imagefolder
type: imagefolder
config: default
split: test
args: default
metrics:
- name: Accuracy
type: accuracy
value: 0.9001663893510815
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# smids_10x_deit_tiny_adamax_001_fold2
This model is a fine-tuned version of [facebook/deit-tiny-patch16-224](https://huggingface.co/facebook/deit-tiny-patch16-224) on the imagefolder dataset.
It achieves the following results on the evaluation set:
- Loss: 1.0691
- Accuracy: 0.9002
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.001
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 50
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|
| 0.3362 | 1.0 | 750 | 0.3519 | 0.8652 |
| 0.2971 | 2.0 | 1500 | 0.3131 | 0.8918 |
| 0.1771 | 3.0 | 2250 | 0.2717 | 0.8885 |
| 0.2985 | 4.0 | 3000 | 0.3652 | 0.8652 |
| 0.1399 | 5.0 | 3750 | 0.3216 | 0.9018 |
| 0.1317 | 6.0 | 4500 | 0.3948 | 0.8802 |
| 0.1309 | 7.0 | 5250 | 0.3860 | 0.8902 |
| 0.1165 | 8.0 | 6000 | 0.4557 | 0.8852 |
| 0.0308 | 9.0 | 6750 | 0.5032 | 0.8686 |
| 0.0315 | 10.0 | 7500 | 0.4981 | 0.8769 |
| 0.0974 | 11.0 | 8250 | 0.6363 | 0.8769 |
| 0.1017 | 12.0 | 9000 | 0.5021 | 0.8869 |
| 0.0475 | 13.0 | 9750 | 0.5896 | 0.8885 |
| 0.0086 | 14.0 | 10500 | 0.6931 | 0.8918 |
| 0.0301 | 15.0 | 11250 | 0.6531 | 0.8902 |
| 0.0049 | 16.0 | 12000 | 0.7157 | 0.8819 |
| 0.0307 | 17.0 | 12750 | 0.7054 | 0.8935 |
| 0.0113 | 18.0 | 13500 | 0.7646 | 0.8869 |
| 0.0492 | 19.0 | 14250 | 0.7424 | 0.8885 |
| 0.0093 | 20.0 | 15000 | 0.6366 | 0.8952 |
| 0.011 | 21.0 | 15750 | 0.8426 | 0.8885 |
| 0.0191 | 22.0 | 16500 | 0.7557 | 0.8952 |
| 0.0047 | 23.0 | 17250 | 0.7578 | 0.8885 |
| 0.0163 | 24.0 | 18000 | 0.8275 | 0.8902 |
| 0.0001 | 25.0 | 18750 | 0.8176 | 0.8935 |
| 0.0023 | 26.0 | 19500 | 0.8054 | 0.8968 |
| 0.0181 | 27.0 | 20250 | 0.8270 | 0.8952 |
| 0.0 | 28.0 | 21000 | 0.8173 | 0.9035 |
| 0.0001 | 29.0 | 21750 | 0.8348 | 0.9018 |
| 0.0 | 30.0 | 22500 | 0.8105 | 0.9101 |
| 0.0 | 31.0 | 23250 | 0.7837 | 0.9118 |
| 0.0 | 32.0 | 24000 | 0.9929 | 0.8935 |
| 0.0 | 33.0 | 24750 | 0.8103 | 0.9085 |
| 0.0 | 34.0 | 25500 | 0.8769 | 0.9035 |
| 0.0 | 35.0 | 26250 | 0.8987 | 0.8985 |
| 0.0 | 36.0 | 27000 | 1.0129 | 0.9002 |
| 0.0053 | 37.0 | 27750 | 0.9506 | 0.9068 |
| 0.0 | 38.0 | 28500 | 1.0495 | 0.8935 |
| 0.0 | 39.0 | 29250 | 0.9869 | 0.9018 |
| 0.0 | 40.0 | 30000 | 1.0087 | 0.8968 |
| 0.0 | 41.0 | 30750 | 1.0348 | 0.8985 |
| 0.0 | 42.0 | 31500 | 1.0299 | 0.8985 |
| 0.0 | 43.0 | 32250 | 1.0437 | 0.8968 |
| 0.0 | 44.0 | 33000 | 1.0468 | 0.8985 |
| 0.0028 | 45.0 | 33750 | 1.0539 | 0.9002 |
| 0.0 | 46.0 | 34500 | 1.0588 | 0.9002 |
| 0.0 | 47.0 | 35250 | 1.0567 | 0.9002 |
| 0.0 | 48.0 | 36000 | 1.0631 | 0.9002 |
| 0.0 | 49.0 | 36750 | 1.0673 | 0.9002 |
| 0.0 | 50.0 | 37500 | 1.0691 | 0.9002 |
### Framework versions
- Transformers 4.32.1
- Pytorch 2.1.1+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2
|
mkoven/ppo-LunarLander-v2
|
mkoven
| 2023-12-20T02:16:40Z | 0 | 0 |
stable-baselines3
|
[
"stable-baselines3",
"LunarLander-v2",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-12-20T02:16:19Z |
---
library_name: stable-baselines3
tags:
- LunarLander-v2
- deep-reinforcement-learning
- reinforcement-learning
- stable-baselines3
model-index:
- name: PPO
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: LunarLander-v2
type: LunarLander-v2
metrics:
- type: mean_reward
value: 251.48 +/- 20.90
name: mean_reward
verified: false
---
# **PPO** Agent playing **LunarLander-v2**
This is a trained model of a **PPO** agent playing **LunarLander-v2**
using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3).
## Usage (with Stable-baselines3)
TODO: Add your code
```python
from stable_baselines3 import ...
from huggingface_sb3 import load_from_hub
...
```
|
Kekega/rut5-base-summ-dialogsum
|
Kekega
| 2023-12-20T02:11:53Z | 5 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"t5",
"text2text-generation",
"generated_from_trainer",
"base_model:d0rj/rut5-base-summ",
"base_model:finetune:d0rj/rut5-base-summ",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2023-12-20T02:09:14Z |
---
base_model: d0rj/rut5-base-summ
tags:
- generated_from_trainer
metrics:
- rouge
model-index:
- name: rut5-base-summ-dialogsum
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# rut5-base-summ-dialogsum
This model is a fine-tuned version of [d0rj/rut5-base-summ](https://huggingface.co/d0rj/rut5-base-summ) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.1263
- Rouge1: 33.5111
- Rouge2: 0.1696
- Rougel: 33.4559
- Rougelsum: 33.4934
- Gen Len: 4.1546
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 25
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|:-------------:|:-----:|:-----:|:---------------:|:-------:|:------:|:-------:|:---------:|:-------:|
| 2.0946 | 1.0 | 786 | 1.7462 | 45.4252 | 0.0 | 45.4009 | 45.4139 | 4.0464 |
| 1.7182 | 2.0 | 1572 | 1.5005 | 44.9295 | 0.0 | 44.9183 | 44.9108 | 4.1126 |
| 1.5304 | 3.0 | 2358 | 1.3826 | 39.5888 | 0.0 | 39.5811 | 39.5646 | 4.1698 |
| 1.4261 | 4.0 | 3144 | 1.3121 | 30.1735 | 0.0 | 30.1127 | 30.1415 | 4.1520 |
| 1.3252 | 5.0 | 3930 | 1.2641 | 35.7738 | 0.0 | 35.7408 | 35.7858 | 3.8791 |
| 1.2878 | 6.0 | 4716 | 1.2353 | 33.0773 | 0.0 | 32.9682 | 33.0551 | 3.7252 |
| 1.2068 | 7.0 | 5502 | 1.2051 | 34.4094 | 0.0 | 34.3902 | 34.3884 | 3.7729 |
| 1.1763 | 8.0 | 6288 | 1.1952 | 33.0914 | 0.1908 | 33.0267 | 33.0472 | 3.9739 |
| 1.1346 | 9.0 | 7074 | 1.1798 | 33.9606 | 0.0 | 33.9335 | 33.979 | 4.1768 |
| 1.1044 | 10.0 | 7860 | 1.1632 | 32.9529 | 0.0 | 32.9367 | 32.9396 | 4.1673 |
| 1.1073 | 11.0 | 8646 | 1.1499 | 34.0904 | 0.0 | 34.0659 | 34.1317 | 4.1934 |
| 1.0619 | 12.0 | 9432 | 1.1516 | 32.9502 | 0.0 | 32.9056 | 32.9376 | 4.0312 |
| 1.0365 | 13.0 | 10218 | 1.1478 | 31.68 | 0.0 | 31.6488 | 31.7003 | 4.0293 |
| 1.0161 | 14.0 | 11004 | 1.1427 | 32.6651 | 0.0424 | 32.6345 | 32.6538 | 4.1113 |
| 0.9805 | 15.0 | 11790 | 1.1343 | 34.0304 | 0.0636 | 33.9433 | 33.999 | 4.0674 |
| 0.9661 | 16.0 | 12576 | 1.1309 | 34.8704 | 0.0848 | 34.8014 | 34.8501 | 4.0681 |
| 0.9511 | 17.0 | 13362 | 1.1348 | 32.8744 | 0.0 | 32.8277 | 32.8547 | 4.1081 |
| 0.9392 | 18.0 | 14148 | 1.1326 | 32.9349 | 0.1908 | 32.8895 | 32.9376 | 4.2627 |
| 0.9341 | 19.0 | 14934 | 1.1263 | 33.5111 | 0.1696 | 33.4559 | 33.4934 | 4.1546 |
| 0.9396 | 20.0 | 15720 | 1.1349 | 33.9121 | 0.2545 | 33.8438 | 33.8993 | 4.1705 |
| 0.9314 | 21.0 | 16506 | 1.1276 | 33.0779 | 0.106 | 33.0546 | 33.0903 | 4.1399 |
| 0.8987 | 22.0 | 17292 | 1.1333 | 33.8566 | 0.1696 | 33.7943 | 33.843 | 4.1419 |
| 0.8895 | 23.0 | 18078 | 1.1343 | 33.6108 | 0.1484 | 33.5738 | 33.636 | 4.2328 |
| 0.8847 | 24.0 | 18864 | 1.1355 | 33.4257 | 0.2757 | 33.3804 | 33.4495 | 4.1711 |
| 0.8832 | 25.0 | 19650 | 1.1355 | 33.6211 | 0.3393 | 33.5937 | 33.636 | 4.1959 |
### Framework versions
- Transformers 4.35.2
- Pytorch 2.0.1+cu117
- Datasets 2.15.0
- Tokenizers 0.15.0
|
SimplCup/Dagon
|
SimplCup
| 2023-12-20T01:56:32Z | 0 | 0 | null |
[
"license:cc-by-nc-nd-4.0",
"region:us"
] | null | 2023-12-20T01:56:17Z |
---
license: cc-by-nc-nd-4.0
---
|
konapieces/VoidnoiseLoRA
|
konapieces
| 2023-12-20T01:53:32Z | 0 | 2 |
diffusers
|
[
"diffusers",
"art",
"artwork",
"girl",
"stable-diffusion",
"lora",
"ja",
"en",
"license:creativeml-openrail-m",
"region:us"
] | null | 2023-10-21T06:52:46Z |
---
license: creativeml-openrail-m
language:
- ja
- en
library_name: diffusers
tags:
- art
- artwork
- girl
- stable-diffusion
- lora
---
# VoidnoiseLoRA
# ▼ モデルの詳細 (Model Details)
<summary>AsianEyesEra</summary>
<details>
<summary>R1311_Rev2</summary>
<div>
# ▼ 本モデルの概要 (Overview of this model)
VoidnoiseLoRA - AsianEyesEra_BaseOn-R1311_Rev2は、SD1.5モデルをベースに、アジア系女性の目を美麗にするLoRAになります。<br>
非常に目力が強く表現されるので、AIフォトグラフィやAIグラビアに最適なLoRAです。<br>
"VoidnoiseLoRA - AsianEyesEra_BaseOn-R1311_Rev2" is a LoRA based on SD1.5 model to make Asian women's eyes beautiful.<br>
It is the best LoRA for AI Photography and AI Gravure because it expresses very strong eyesight.<br>
# ▼ 推奨設定及びトリガーワード (Recommended settings and Trigger Word)
- LoRA Strength: 0.6 - 0.8
- Trigger Word: AsianEyesEra
## ▼ 出力サンプル (Sample)
<div align="center">
<img src="https://huggingface.co/konapieces/VoidnoiseLoRA/resolve/main/AsianEyesEra/images/R1311_rev2/sample.png" width="512">
Summary Sample<br>
<img src="https://huggingface.co/konapieces/VoidnoiseLoRA/resolve/main/AsianEyesEra/images/R1311_rev2/sample1.png" width="512">
LoRA Strength : 0.8<br>
<img src="https://huggingface.co/konapieces/VoidnoiseLoRA/resolve/main/AsianEyesEra/images/R1311_rev2/sample2.png" width="512">
LoRA Strength : 1.0<br>
<img src="https://huggingface.co/konapieces/VoidnoiseLoRA/resolve/main/AsianEyesEra/images/R1311_rev2/sample3.png" width="512">
LoRA Strength : 0.8<br>
<img src="https://huggingface.co/konapieces/VoidnoiseLoRA/resolve/main/AsianEyesEra/images/R1311_rev2/sample4.png" width="512">
LoRA Strength : 0.8<br>
</div>
</div>
</details>
<details>
<summary>R1311r2 Alpha</summary>
<div>
# ▼ 本モデルの概要 (Overview of this model)
VoidnoiseLoRA - AsianEyesEra R1311r2 Alphaは、VoidnoiseCore R1311 Rev2をベースモデルとして、目の特徴をスライド式に変化させ、バリエーション豊かな目の美麗化LoRAを目的として開発しています。<br>
前作よりも、ハッキリとした目の大きさと、目力が強くなっています。<br>
スライドLoRAということもあり、マイナス強度で前作寄りの適用に、プラス強度で今回のLoRAが適用されます。<br>
"VoidnoiseLoRA - AsianEyesEra R1311r2 Alpha" is based on the "VoidnoiseCore R1311 Rev2" and is developed for the purpose of creating a LoRA with sliding eye feature changes and a rich variation of eye beautification. <br>
The eyes are clearly larger and more intense than in the previous version.<br>
As it is a sliding LoRA, the negative strength applies a LoRA closer to the previous work, and the positive strength applies this time's LoRA.<br>
# ▼ 学習パラメータ (Training Parameter)
- Learning rate : 1e-4
- Step : 2
- Epoch : 800
- Precision : bf16
- Dim : 32
<br>
# ▼ 推奨設定及びトリガーワード (Recommended settings and Trigger Word)
- LoRA Strength: -0.25 ～ 0.25
- Trigger Word: -
- Negative Prompt: freckles
<br>
- SD WebUIの拡張機能である、Tiled VAEとFreeUを使用して頂くと、さらに精細かつ美麗に描写することができます。
- 学習元のデータセットの関係上、そばかすやホクロなどが顔に描写される可能性がある為、ネガティブプロンプトへ「freckles（そばかす）」を追加することを推奨します。
<br>
- SD WebUI extensions, Tiled VAE and FreeU, can be used for more detailed and beautiful rendering.
- Due to the nature of the training dataset, freckles and moles may appear on the face, so it is recommended to add "freckles" to the negative prompt.
## ▼ 出力サンプル (Sample)
<div align="center">
<img src="https://huggingface.co/konapieces/VoidnoiseLoRA/resolve/main/AsianEyesEra/images/R1311r2_Alpha/sample1.png" width="512">
<img src="https://huggingface.co/konapieces/VoidnoiseLoRA/resolve/main/AsianEyesEra/images/R1311r2_Alpha/sample2.png" width="512">
<img src="https://huggingface.co/konapieces/VoidnoiseLoRA/resolve/main/AsianEyesEra/images/R1311r2_Alpha/sample3.png" width="512">
</div>
</div>
</details>
----
# ▼ 免責事項 (Disclaimer)
- 本モデルを使用して作成された画像に関しては、個々の利用者に委ねておりますので、生成された画像に関する如何なる問題や係争について、モデル製作者は一切の責任を負いません。
- 本モデルはアダルトコンテンツを目的とした用途を想定しておりません。成人向けコンテンツを生成し、発生した問題についてはモデル製作者は一切の責任を負いません。
- ライセンスに関して問題が発生した場合は、本モデルを予告なく削除させて頂く可能性があります。ご了承ください。
- 犯罪への利用や医療用などの専門的な用途への使用は禁止されております。ライセンス不履行による過失については、モデル製作者は一切の責任を負いません。
- CreativeML OpenRAIL ライセンスの特性上、モデル及び派生モデルにおける販売を許可しておりますが、現状オープンアクセスライセンスである為、モデルの販売は推奨致しません。著作者に無断でモデル販売を行った際に生じたいかなる問題もモデル製作者は一切責任を負いません。
<br>
- The model creator assumes no liability for any problems or disputes related to the images created using this model.
- This model is not intended for use with adult content. The model creator assumes no liability for any problems that may occur as a result of generating adult-oriented content.
- In the event of any licensing issues, this model may be removed without notice. We appreciate your understanding.
- Use for criminal offenses or for professional purposes such as medical use is prohibited. The model maker is not liable for any negligence due to non-fulfillment of the license.
- The CreativeML OpenRAIL license permits the sale of models and derivatives, but does not recommend the sale of models because it is currently an open access license. The creator of the model will not be held responsible for any problems that may arise from the sale of the model without the author's permission.
---
# ▼ モデルライセンス (Model License)
このモデルはオープンアクセスであり、すべての人が利用できます。CreativeML OpenRAIL-M ライセンスにより、権利と使用方法がさらに規定されています。<br>
CreativeML OpenRAIL ライセンスでは、次のことが規定されています。<br>
1. モデルを使用して、違法または有害な出力またはコンテンツを意図的に作成または共有することはできません。<br>
2. 作成者は、あなたが生成した出力に対していかなる権利も主張しません。あなたはそれらを自由に使用でき、ライセンスに設定された規定に違反してはならない使用について説明責任を負います。<br>
3. 重みを再配布し、モデルを商用および/またはサービスとして使用することができます。<br>
その場合、ライセンスに記載されているのと同じ使用制限を含め、<br>
CreativeML OpenRAIL-M のコピーをすべてのユーザーと共有する必要があることに注意してください。 (ライセンスを完全にかつ慎重にお読みください。) <br>
[こちら](https://huggingface.co/spaces/CompVis/stable-diffusion-license)からライセンス全文をお読みください。<br>
This model is open access and available to all, with a CreativeML OpenRAIL-M license further specifying rights and usage. The CreativeML OpenRAIL License specifies:<br>
1. You can't use the model to deliberately produce nor share illegal or harmful outputs or content<br>
2. The authors claims no rights on the outputs you generate, you are free to use them and are accountable for their use which must not go against the provisions set in the license<br>
3. You may re-distribute the weights and use the model commercially and/or as a service. <br>
If you do, please be aware you have to include the same use restrictions as the ones in the license and share a copy of the CreativeML OpenRAIL-M to all your users (please read the license entirely and carefully) <br>
Please read the full license [here](https://huggingface.co/spaces/CompVis/stable-diffusion-license)<br>
---
# ▼ 製作者 (The creator of this model)
とーふのかけら(konapieces)<br>
twitter: <a href="https://twitter.com/konapieces" target="_blank"> @konapieces</a><br>
Website: <a href="https://lit.link/konapieces" target="_blank">https://lit.link/konapieces</a>
---
|
ntc-ai/SDXL-LoRA-slider.colorful
|
ntc-ai
| 2023-12-20T01:36:58Z | 76 | 1 |
diffusers
|
[
"diffusers",
"text-to-image",
"stable-diffusion-xl",
"lora",
"template:sd-lora",
"template:sdxl-lora",
"sdxl-sliders",
"ntcai.xyz-sliders",
"concept",
"en",
"base_model:stabilityai/stable-diffusion-xl-base-1.0",
"base_model:adapter:stabilityai/stable-diffusion-xl-base-1.0",
"license:mit",
"region:us"
] |
text-to-image
| 2023-12-20T01:36:55Z |
---
language:
- en
thumbnail: "images/evaluate/colorful.../colorful_17_3.0.png"
widget:
- text: colorful
output:
url: images/colorful_17_3.0.png
- text: colorful
output:
url: images/colorful_19_3.0.png
- text: colorful
output:
url: images/colorful_20_3.0.png
- text: colorful
output:
url: images/colorful_21_3.0.png
- text: colorful
output:
url: images/colorful_22_3.0.png
tags:
- text-to-image
- stable-diffusion-xl
- lora
- template:sd-lora
- template:sdxl-lora
- sdxl-sliders
- ntcai.xyz-sliders
- concept
- diffusers
license: "mit"
inference: false
instance_prompt: "colorful"
base_model: "stabilityai/stable-diffusion-xl-base-1.0"
---
# ntcai.xyz slider - colorful (SDXL LoRA)
| Strength: -3 | Strength: 0 | Strength: 3 |
| --- | --- | --- |
| <img src="images/colorful_17_-3.0.png" width=256 height=256 /> | <img src="images/colorful_17_0.0.png" width=256 height=256 /> | <img src="images/colorful_17_3.0.png" width=256 height=256 /> |
| <img src="images/colorful_19_-3.0.png" width=256 height=256 /> | <img src="images/colorful_19_0.0.png" width=256 height=256 /> | <img src="images/colorful_19_3.0.png" width=256 height=256 /> |
| <img src="images/colorful_20_-3.0.png" width=256 height=256 /> | <img src="images/colorful_20_0.0.png" width=256 height=256 /> | <img src="images/colorful_20_3.0.png" width=256 height=256 /> |
## Download
Weights for this model are available in Safetensors format.
## Trigger words
You can apply this LoRA with trigger words for additional effect:
```
colorful
```
## Use in diffusers
```python
from diffusers import StableDiffusionXLPipeline
from diffusers import EulerAncestralDiscreteScheduler
import torch
pipe = StableDiffusionXLPipeline.from_single_file("https://huggingface.co/martyn/sdxl-turbo-mario-merge-top-rated/blob/main/topRatedTurboxlLCM_v10.safetensors")
pipe.to("cuda")
pipe.scheduler = EulerAncestralDiscreteScheduler.from_config(pipe.scheduler.config)
# Load the LoRA
pipe.load_lora_weights('ntc-ai/SDXL-LoRA-slider.colorful', weight_name='colorful.safetensors', adapter_name="colorful")
# Activate the LoRA
pipe.set_adapters(["colorful"], adapter_weights=[2.0])
prompt = "medieval rich kingpin sitting in a tavern, colorful"
negative_prompt = "nsfw"
width = 512
height = 512
num_inference_steps = 10
guidance_scale = 2
image = pipe(prompt, negative_prompt=negative_prompt, width=width, height=height, guidance_scale=guidance_scale, num_inference_steps=num_inference_steps).images[0]
image.save('result.png')
```
## Support the Patreon
If you like this model please consider [joining our Patreon](https://www.patreon.com/NTCAI).
By joining our Patreon, you'll gain access to an ever-growing library of over 480+ unique and diverse LoRAs, covering a wide range of styles and genres. You'll also receive early access to new models and updates, exclusive behind-the-scenes content, and the powerful LoRA slider creator, allowing you to craft your own custom LoRAs and experiment with endless possibilities.
Your support on Patreon will allow us to continue developing and refining new models.
## Other resources
- [CivitAI](https://civitai.com/user/ntc) - Follow ntc on Civit for even more LoRAs
- [ntcai.xyz](https://ntcai.xyz) - See ntcai.xyz to find more articles and LoRAs
|
toddwilson147/dqn-pong
|
toddwilson147
| 2023-12-20T01:34:11Z | 1 | 0 |
stable-baselines3
|
[
"stable-baselines3",
"PongNoFrameskip-v4",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-12-20T01:32:28Z |
---
library_name: stable-baselines3
tags:
- PongNoFrameskip-v4
- deep-reinforcement-learning
- reinforcement-learning
- stable-baselines3
model-index:
- name: DQN
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: PongNoFrameskip-v4
type: PongNoFrameskip-v4
metrics:
- type: mean_reward
value: 4.10 +/- 9.82
name: mean_reward
verified: false
---
# **DQN** Agent playing **PongNoFrameskip-v4**
This is a trained model of a **DQN** agent playing **PongNoFrameskip-v4**
using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3)
and the [RL Zoo](https://github.com/DLR-RM/rl-baselines3-zoo).
The RL Zoo is a training framework for Stable Baselines3
reinforcement learning agents,
with hyperparameter optimization and pre-trained agents included.
## Usage (with SB3 RL Zoo)
RL Zoo: https://github.com/DLR-RM/rl-baselines3-zoo<br/>
SB3: https://github.com/DLR-RM/stable-baselines3<br/>
SB3 Contrib: https://github.com/Stable-Baselines-Team/stable-baselines3-contrib
Install the RL Zoo (with SB3 and SB3-Contrib):
```bash
pip install rl_zoo3
```
```
# Download model and save it into the logs/ folder
python -m rl_zoo3.load_from_hub --algo dqn --env PongNoFrameskip-v4 -orga toddwilson147 -f logs/
python -m rl_zoo3.enjoy --algo dqn --env PongNoFrameskip-v4 -f logs/
```
If you installed the RL Zoo3 via pip (`pip install rl_zoo3`), from anywhere you can do:
```
python -m rl_zoo3.load_from_hub --algo dqn --env PongNoFrameskip-v4 -orga toddwilson147 -f logs/
python -m rl_zoo3.enjoy --algo dqn --env PongNoFrameskip-v4 -f logs/
```
## Training (with the RL Zoo)
```
python -m rl_zoo3.train --algo dqn --env PongNoFrameskip-v4 -f logs/
# Upload the model and generate video (when possible)
python -m rl_zoo3.push_to_hub --algo dqn --env PongNoFrameskip-v4 -f logs/ -orga toddwilson147
```
## Hyperparameters
```python
OrderedDict([('batch_size', 32),
('buffer_size', 100000),
('env_wrapper',
['stable_baselines3.common.atari_wrappers.AtariWrapper']),
('exploration_final_eps', 0.01),
('exploration_fraction', 0.1),
('frame_stack', 4),
('gradient_steps', 1),
('learning_rate', 0.0001),
('learning_starts', 100000),
('n_timesteps', 1000000.0),
('optimize_memory_usage', False),
('policy', 'CnnPolicy'),
('target_update_interval', 1000),
('train_freq', 4),
('normalize', False)])
```
# Environment Arguments
```python
{'render_mode': 'rgb_array'}
```
|
mp/ogow-vuc4-001w-0
|
mp
| 2023-12-20T01:26:30Z | 0 | 0 | null |
[
"safetensors",
"autotrain",
"text-generation",
"license:other",
"region:us"
] |
text-generation
| 2023-12-20T01:26:28Z |
---
tags:
- autotrain
- text-generation
widget:
- text: "I love AutoTrain because "
license: other
---
# Model Trained Using AutoTrain
This model was trained using AutoTrain. For more information, please visit [AutoTrain](https://hf.co/docs/autotrain).
# Usage
```python
from transformers import AutoModelForCausalLM, AutoTokenizer
model_path = "PATH_TO_THIS_REPO"
tokenizer = AutoTokenizer.from_pretrained(model_path)
model = AutoModelForCausalLM.from_pretrained(
model_path,
device_map="auto",
torch_dtype='auto'
).eval()
# Prompt content: "hi"
messages = [
{"role": "user", "content": "hi"}
]
input_ids = tokenizer.apply_chat_template(conversation=messages, tokenize=True, add_generation_prompt=True, return_tensors='pt')
output_ids = model.generate(input_ids.to('cuda'))
response = tokenizer.decode(output_ids[0][input_ids.shape[1]:], skip_special_tokens=True)
# Model response: "Hello! How can I assist you today?"
print(response)
```
|
mosmodels/sslmos
|
mosmodels
| 2023-12-20T01:16:49Z | 0 | 0 | null |
[
"license:bsd-3-clause",
"region:us"
] | null | 2023-12-20T01:09:43Z |
---
license: bsd-3-clause
---
https://github.com/nii-yamagishilab/mos-finetune-ssl
BSD 3-Clause License
Copyright (c) 2021, Yamagishi Laboratory, National Institute of Informatics
All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are met:
* Redistributions of source code must retain the above copyright notice, this
list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above copyright notice,
this list of conditions and the following disclaimer in the documentation
and/or other materials provided with the distribution.
* Neither the name of the copyright holder nor the names of its
contributors may be used to endorse or promote products derived from
this software without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE
FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
|
TheBloke/Swallow-70B-GPTQ
|
TheBloke
| 2023-12-20T01:12:18Z | 10 | 0 |
transformers
|
[
"transformers",
"safetensors",
"llama",
"text-generation",
"en",
"ja",
"base_model:tokyotech-llm/Swallow-70b-hf",
"base_model:quantized:tokyotech-llm/Swallow-70b-hf",
"license:llama2",
"autotrain_compatible",
"text-generation-inference",
"4-bit",
"gptq",
"region:us"
] |
text-generation
| 2023-12-19T21:54:33Z |
---
base_model: tokyotech-llm/Swallow-70b-hf
inference: false
language:
- en
- ja
library_name: transformers
license: llama2
model_creator: tokyotech-llm
model_name: Swallow 70B
model_type: llama
pipeline_tag: text-generation
prompt_template: '{prompt}
'
quantized_by: TheBloke
---
<!-- markdownlint-disable MD041 -->
<!-- header start -->
<!-- 200823 -->
<div style="width: auto; margin-left: auto; margin-right: auto">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<div style="text-align:center; margin-top: 0em; margin-bottom: 0em"><p style="margin-top: 0.25em; margin-bottom: 0em;">TheBloke's LLM work is generously supported by a grant from <a href="https://a16z.com">andreessen horowitz (a16z)</a></p></div>
<hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
<!-- header end -->
# Swallow 70B - GPTQ
- Model creator: [tokyotech-llm](https://huggingface.co/tokyotech-llm)
- Original model: [Swallow 70B](https://huggingface.co/tokyotech-llm/Swallow-70b-hf)
<!-- description start -->
# Description
This repo contains GPTQ model files for [tokyotech-llm's Swallow 70B](https://huggingface.co/tokyotech-llm/Swallow-70b-hf).
Multiple GPTQ parameter permutations are provided; see Provided Files below for details of the options provided, their parameters, and the software used to create them.
<!-- description end -->
<!-- repositories-available start -->
## Repositories available
* [AWQ model(s) for GPU inference.](https://huggingface.co/TheBloke/Swallow-70B-AWQ)
* [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/Swallow-70B-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/Swallow-70B-GGUF)
* [tokyotech-llm's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/tokyotech-llm/Swallow-70b-hf)
<!-- repositories-available end -->
<!-- prompt-template start -->
## Prompt template: None
```
{prompt}
```
<!-- prompt-template end -->
<!-- README_GPTQ.md-compatible clients start -->
## Known compatible clients / servers
GPTQ models are currently supported on Linux (NVidia/AMD) and Windows (NVidia only). macOS users: please use GGUF models.
These GPTQ models are known to work in the following inference servers/webuis.
- [text-generation-webui](https://github.com/oobabooga/text-generation-webui)
- [KoboldAI United](https://github.com/henk717/koboldai)
- [LoLLMS Web UI](https://github.com/ParisNeo/lollms-webui)
- [Hugging Face Text Generation Inference (TGI)](https://github.com/huggingface/text-generation-inference)
This may not be a complete list; if you know of others, please let me know!
<!-- README_GPTQ.md-compatible clients end -->
<!-- README_GPTQ.md-provided-files start -->
## Provided files, and GPTQ parameters
Multiple quantisation parameters are provided, to allow you to choose the best one for your hardware and requirements.
Each separate quant is in a different branch. See below for instructions on fetching from different branches.
Most GPTQ files are made with AutoGPTQ. Mistral models are currently made with Transformers.
<details>
<summary>Explanation of GPTQ parameters</summary>
- Bits: The bit size of the quantised model.
- GS: GPTQ group size. Higher numbers use less VRAM, but have lower quantisation accuracy. "None" is the lowest possible value.
- Act Order: True or False. Also known as `desc_act`. True results in better quantisation accuracy. Some GPTQ clients have had issues with models that use Act Order plus Group Size, but this is generally resolved now.
- Damp %: A GPTQ parameter that affects how samples are processed for quantisation. 0.01 is default, but 0.1 results in slightly better accuracy.
- GPTQ dataset: The calibration dataset used during quantisation. Using a dataset more appropriate to the model's training can improve quantisation accuracy. Note that the GPTQ calibration dataset is not the same as the dataset used to train the model - please refer to the original model repo for details of the training dataset(s).
- Sequence Length: The length of the dataset sequences used for quantisation. Ideally this is the same as the model sequence length. For some very long sequence models (16+K), a lower sequence length may have to be used. Note that a lower sequence length does not limit the sequence length of the quantised model. It only impacts the quantisation accuracy on longer inference sequences.
- ExLlama Compatibility: Whether this file can be loaded with ExLlama, which currently only supports Llama and Mistral models in 4-bit.
</details>
| Branch | Bits | GS | Act Order | Damp % | GPTQ Dataset | Seq Len | Size | ExLlama | Desc |
| ------ | ---- | -- | --------- | ------ | ------------ | ------- | ---- | ------- | ---- |
| [main](https://huggingface.co/TheBloke/Swallow-70B-GPTQ/tree/main) | 4 | None | Yes | 0.1 | [Alpaca Japanese](https://huggingface.co/datasets/fujiki/japanese_alpaca_data/viewer/) | 4096 | 35.70 GB | Yes | 4-bit, with Act Order. No group size, to lower VRAM requirements. |
| [gptq-4bit-128g-actorder_True](https://huggingface.co/TheBloke/Swallow-70B-GPTQ/tree/gptq-4bit-128g-actorder_True) | 4 | 128 | Yes | 0.1 | [Alpaca Japanese](https://huggingface.co/datasets/fujiki/japanese_alpaca_data/viewer/) | 4096 | 37.02 GB | Yes | 4-bit, with Act Order and group size 128g. Uses even less VRAM than 64g, but with slightly lower accuracy. |
| [gptq-4bit-32g-actorder_True](https://huggingface.co/TheBloke/Swallow-70B-GPTQ/tree/gptq-4bit-32g-actorder_True) | 4 | 32 | Yes | 0.1 | [Alpaca Japanese](https://huggingface.co/datasets/fujiki/japanese_alpaca_data/viewer/) | 4096 | 41.03 GB | Yes | 4-bit, with Act Order and group size 32g. Gives highest possible inference quality, with maximum VRAM usage. |
| [gptq-3bit--1g-actorder_True](https://huggingface.co/TheBloke/Swallow-70B-GPTQ/tree/gptq-3bit--1g-actorder_True) | 3 | None | Yes | 0.1 | [Alpaca Japanese](https://huggingface.co/datasets/fujiki/japanese_alpaca_data/viewer/) | 4096 | 27.14 GB | No | 3-bit, with Act Order and no group size. Lowest possible VRAM requirements. May be lower quality than 3-bit 128g. |
| [gptq-3bit-128g-actorder_True](https://huggingface.co/TheBloke/Swallow-70B-GPTQ/tree/gptq-3bit-128g-actorder_True) | 3 | 128 | Yes | 0.1 | [Alpaca Japanese](https://huggingface.co/datasets/fujiki/japanese_alpaca_data/viewer/) | 4096 | 28.40 GB | No | 3-bit, with group size 128g and act-order. Higher quality than 128g-False. |
| [gptq-3bit-32g-actorder_True](https://huggingface.co/TheBloke/Swallow-70B-GPTQ/tree/gptq-3bit-32g-actorder_True) | 3 | 32 | Yes | 0.1 | [Alpaca Japanese](https://huggingface.co/datasets/fujiki/japanese_alpaca_data/viewer/) | 4096 | 32.21 GB | No | 3-bit, with group size 64g and act-order. Highest quality 3-bit option. |
<!-- README_GPTQ.md-provided-files end -->
<!-- README_GPTQ.md-download-from-branches start -->
## How to download, including from branches
### In text-generation-webui
To download from the `main` branch, enter `TheBloke/Swallow-70B-GPTQ` in the "Download model" box.
To download from another branch, add `:branchname` to the end of the download name, eg `TheBloke/Swallow-70B-GPTQ:gptq-4bit-128g-actorder_True`
### From the command line
I recommend using the `huggingface-hub` Python library:
```shell
pip3 install huggingface-hub
```
To download the `main` branch to a folder called `Swallow-70B-GPTQ`:
```shell
mkdir Swallow-70B-GPTQ
huggingface-cli download TheBloke/Swallow-70B-GPTQ --local-dir Swallow-70B-GPTQ --local-dir-use-symlinks False
```
To download from a different branch, add the `--revision` parameter:
```shell
mkdir Swallow-70B-GPTQ
huggingface-cli download TheBloke/Swallow-70B-GPTQ --revision gptq-4bit-128g-actorder_True --local-dir Swallow-70B-GPTQ --local-dir-use-symlinks False
```
<details>
<summary>More advanced huggingface-cli download usage</summary>
If you remove the `--local-dir-use-symlinks False` parameter, the files will instead be stored in the central Hugging Face cache directory (default location on Linux is: `~/.cache/huggingface`), and symlinks will be added to the specified `--local-dir`, pointing to their real location in the cache. This allows for interrupted downloads to be resumed, and allows you to quickly clone the repo to multiple places on disk without triggering a download again. The downside, and the reason why I don't list that as the default option, is that the files are then hidden away in a cache folder and it's harder to know where your disk space is being used, and to clear it up if/when you want to remove a download model.
The cache location can be changed with the `HF_HOME` environment variable, and/or the `--cache-dir` parameter to `huggingface-cli`.
For more documentation on downloading with `huggingface-cli`, please see: [HF -> Hub Python Library -> Download files -> Download from the CLI](https://huggingface.co/docs/huggingface_hub/guides/download#download-from-the-cli).
To accelerate downloads on fast connections (1Gbit/s or higher), install `hf_transfer`:
```shell
pip3 install hf_transfer
```
And set environment variable `HF_HUB_ENABLE_HF_TRANSFER` to `1`:
```shell
mkdir Swallow-70B-GPTQ
HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download TheBloke/Swallow-70B-GPTQ --local-dir Swallow-70B-GPTQ --local-dir-use-symlinks False
```
Windows Command Line users: You can set the environment variable by running `set HF_HUB_ENABLE_HF_TRANSFER=1` before the download command.
</details>
### With `git` (**not** recommended)
To clone a specific branch with `git`, use a command like this:
```shell
git clone --single-branch --branch gptq-4bit-128g-actorder_True https://huggingface.co/TheBloke/Swallow-70B-GPTQ
```
Note that using Git with HF repos is strongly discouraged. It will be much slower than using `huggingface-hub`, and will use twice as much disk space as it has to store the model files twice (it stores every byte both in the intended target folder, and again in the `.git` folder as a blob.)
<!-- README_GPTQ.md-download-from-branches end -->
<!-- README_GPTQ.md-text-generation-webui start -->
## How to easily download and use this model in [text-generation-webui](https://github.com/oobabooga/text-generation-webui)
Please make sure you're using the latest version of [text-generation-webui](https://github.com/oobabooga/text-generation-webui).
It is strongly recommended to use the text-generation-webui one-click-installers unless you're sure you know how to make a manual install.
1. Click the **Model tab**.
2. Under **Download custom model or LoRA**, enter `TheBloke/Swallow-70B-GPTQ`.
- To download from a specific branch, enter for example `TheBloke/Swallow-70B-GPTQ:gptq-4bit-128g-actorder_True`
- see Provided Files above for the list of branches for each option.
3. Click **Download**.
4. The model will start downloading. Once it's finished it will say "Done".
5. In the top left, click the refresh icon next to **Model**.
6. In the **Model** dropdown, choose the model you just downloaded: `Swallow-70B-GPTQ`
7. The model will automatically load, and is now ready for use!
8. If you want any custom settings, set them and then click **Save settings for this model** followed by **Reload the Model** in the top right.
- Note that you do not need to and should not set manual GPTQ parameters any more. These are set automatically from the file `quantize_config.json`.
9. Once you're ready, click the **Text Generation** tab and enter a prompt to get started!
<!-- README_GPTQ.md-text-generation-webui end -->
<!-- README_GPTQ.md-use-from-tgi start -->
## Serving this model from Text Generation Inference (TGI)
It's recommended to use TGI version 1.1.0 or later. The official Docker container is: `ghcr.io/huggingface/text-generation-inference:1.1.0`
Example Docker parameters:
```shell
--model-id TheBloke/Swallow-70B-GPTQ --port 3000 --quantize gptq --max-input-length 3696 --max-total-tokens 4096 --max-batch-prefill-tokens 4096
```
Example Python code for interfacing with TGI (requires huggingface-hub 0.17.0 or later):
```shell
pip3 install huggingface-hub
```
```python
from huggingface_hub import InferenceClient
endpoint_url = "https://your-endpoint-url-here"
prompt = "Tell me about AI"
prompt_template=f'''{prompt}
'''
client = InferenceClient(endpoint_url)
response = client.text_generation(prompt,
max_new_tokens=128,
do_sample=True,
temperature=0.7,
top_p=0.95,
top_k=40,
repetition_penalty=1.1)
print(f"Model output: {response}")
```
<!-- README_GPTQ.md-use-from-tgi end -->
<!-- README_GPTQ.md-use-from-python start -->
## Python code example: inference from this GPTQ model
### Install the necessary packages
Requires: Transformers 4.33.0 or later, Optimum 1.12.0 or later, and AutoGPTQ 0.4.2 or later.
```shell
pip3 install --upgrade transformers optimum
# If using PyTorch 2.1 + CUDA 12.x:
pip3 install --upgrade auto-gptq
# or, if using PyTorch 2.1 + CUDA 11.x:
pip3 install --upgrade auto-gptq --extra-index-url https://huggingface.github.io/autogptq-index/whl/cu118/
```
If you are using PyTorch 2.0, you will need to install AutoGPTQ from source. Likewise if you have problems with the pre-built wheels, you should try building from source:
```shell
pip3 uninstall -y auto-gptq
git clone https://github.com/PanQiWei/AutoGPTQ
cd AutoGPTQ
git checkout v0.5.1
pip3 install .
```
### Example Python code
```python
from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
model_name_or_path = "TheBloke/Swallow-70B-GPTQ"
# To use a different branch, change revision
# For example: revision="gptq-4bit-128g-actorder_True"
model = AutoModelForCausalLM.from_pretrained(model_name_or_path,
device_map="auto",
trust_remote_code=False,
revision="main")
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
prompt = "Write a story about llamas"
system_message = "You are a story writing assistant"
prompt_template=f'''{prompt}
'''
print("\n\n*** Generate:")
input_ids = tokenizer(prompt_template, return_tensors='pt').input_ids.cuda()
output = model.generate(inputs=input_ids, temperature=0.7, do_sample=True, top_p=0.95, top_k=40, max_new_tokens=512)
print(tokenizer.decode(output[0]))
# Inference can also be done using transformers' pipeline
print("*** Pipeline:")
pipe = pipeline(
"text-generation",
model=model,
tokenizer=tokenizer,
max_new_tokens=512,
do_sample=True,
temperature=0.7,
top_p=0.95,
top_k=40,
repetition_penalty=1.1
)
print(pipe(prompt_template)[0]['generated_text'])
```
<!-- README_GPTQ.md-use-from-python end -->
<!-- README_GPTQ.md-compatibility start -->
## Compatibility
The files provided are tested to work with Transformers. For non-Mistral models, AutoGPTQ can also be used directly.
[ExLlama](https://github.com/turboderp/exllama) is compatible with Llama architecture models (including Mistral, Yi, DeepSeek, SOLAR, etc) in 4-bit. Please see the Provided Files table above for per-file compatibility.
For a list of clients/servers, please see "Known compatible clients / servers", above.
<!-- README_GPTQ.md-compatibility end -->
<!-- footer start -->
<!-- 200823 -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute
Thanks to the [chirper.ai](https://chirper.ai) team!
Thanks to Clay from [gpus.llm-utils.org](llm-utils)!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Aemon Algiz.
**Patreon special mentions**: Michael Levine, 阿明, Trailburnt, Nikolai Manek, John Detwiler, Randy H, Will Dee, Sebastain Graf, NimbleBox.ai, Eugene Pentland, Emad Mostaque, Ai Maven, Jim Angel, Jeff Scroggin, Michael Davis, Manuel Alberto Morcote, Stephen Murray, Robert, Justin Joy, Luke @flexchar, Brandon Frisco, Elijah Stavena, S_X, Dan Guido, Undi ., Komninos Chatzipapas, Shadi, theTransient, Lone Striker, Raven Klaugh, jjj, Cap'n Zoog, Michel-Marie MAUDET (LINAGORA), Matthew Berman, David, Fen Risland, Omer Bin Jawed, Luke Pendergrass, Kalila, OG, Erik Bjäreholt, Rooh Singh, Joseph William Delisle, Dan Lewis, TL, John Villwock, AzureBlack, Brad, Pedro Madruga, Caitlyn Gatomon, K, jinyuan sun, Mano Prime, Alex, Jeffrey Morgan, Alicia Loh, Illia Dulskyi, Chadd, transmissions 11, fincy, Rainer Wilmers, ReadyPlayerEmma, knownsqashed, Mandus, biorpg, Deo Leter, Brandon Phillips, SuperWojo, Sean Connelly, Iucharbius, Jack West, Harry Royden McLaughlin, Nicholas, terasurfer, Vitor Caleffi, Duane Dunston, Johann-Peter Hartmann, David Ziegler, Olakabola, Ken Nordquist, Trenton Dambrowitz, Tom X Nguyen, Vadim, Ajan Kanaga, Leonard Tan, Clay Pascal, Alexandros Triantafyllidis, JM33133, Xule, vamX, ya boyyy, subjectnull, Talal Aujan, Alps Aficionado, wassieverse, Ari Malik, James Bentley, Woland, Spencer Kim, Michael Dempsey, Fred von Graf, Elle, zynix, William Richards, Stanislav Ovsiannikov, Edmond Seymore, Jonathan Leane, Martin Kemka, usrbinkat, Enrico Ros
Thank you to all my generous patrons and donaters!
And thank you again to a16z for their generous grant.
<!-- footer end -->
# Original model card: tokyotech-llm's Swallow 70B
# Swallow
Our Swallow model has undergone continuous pre-training from the Llama 2 family, primarily with the addition of Japanese language data. The tuned versions use supervised fine-tuning (SFT).
Links to other models can be found in the index.
## Swallow Model Index
|Model|Swallow-hf|Swallow-instruct-hf|
|---|---|---|
|7B| [Link](https://huggingface.co/tokyotech-llm/Swallow-7b-hf) | [Link](https://huggingface.co/tokyotech-llm/Swallow-7b-instruct-hf)|
|13B| [Link](https://huggingface.co/tokyotech-llm/Swallow-13b-hf) | [Link](https://huggingface.co/tokyotech-llm/Swallow-13b-instruct-hf)|
|70B| [Link](https://huggingface.co/tokyotech-llm/Swallow-70b-hf) | [Link](https://huggingface.co/tokyotech-llm/Swallow-70b-instruct-hf)|
This repository provides large language models developed by [TokyoTech-LLM](https://tokyotech-llm.github.io/).
Read our [blog post](https://zenn.dev/tokyotech_lm/articles/d6cb3a8fdfc907) or our paper (preprint coming soon) for more details!
## Model Details
* **Model type**: Please refer to LLaMA-2 technical report for details on the model architecture.
* **Language(s)**: Japanese English
* **Library**: [Megatron-LM](https://github.com/rioyokotalab/Megatron-Llama2)
* **Tokenizer**: This model employs a tokenizer that features a broadened vocabulary based on Japanese data. This allows for a more efficient representation of text using fewer tokens, leading to a notably faster inference process.
* **Contact**: swallow[at]nlp.c.titech.ac.jp
## Base Model Performance
### Japanese version
|Model|Size|JCommonsenseQA|JEMHopQA|NIILC|JSQuAD|XL-Sum|MGSM|WMT20-en-ja|WMT20-ja-en|
|---|---|---|---|---|---|---|---|---|---|
| | |4-shot|4-shot|4-shot|4-shot|1-shot|4-shot|4-shot|4-shot|
|Llama 2|7B|0.3852|0.4240|0.3410|0.7917|0.1905|0.0760|0.1783|0.1738|
|Swallow|7B|0.4808|0.5078|0.5968|0.8573|0.1830|0.1240|0.2510|0.1511|
|Llama 2|13B|0.6997|0.4415|0.4170|0.8533|0.2139|0.1320|0.2146|0.1982|
|Swallow|13B|0.7837|0.5063|0.6398|0.9005|0.2168|0.2040|0.2720|0.1771|
|Llama 2|70B|0.8686|0.4656|0.5256|0.9080|**0.2361**|0.3560|0.2643|**0.2398**|
|Swallow|70B|**0.9348**|**0.6290**|**0.6960**|**0.9176**|0.2266|**0.4840**|**0.3043**|0.2298|
## Usage
First install additional dependencies in [requirements.txt](./requirements.txt):
```sh
pip install -r requirements.txt
```
### Use the instruct model
```python
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
model_name = "tokyotech-llm/Swallow-7b-instruct-hf"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.bfloat16, low_cpu_mem_usage=True, device_map="auto")
PROMPT_DICT = {
"prompt_input": (
"以下に、あるタスクを説明する指示があり、それに付随する入力が更なる文脈を提供しています。"
"リクエストを適切に完了するための回答を記述してください。\n\n"
"### 指示:\n{instruction}\n\n### 入力:\n{input}\n\n### 応答:"
),
"prompt_no_input": (
"以下に、あるタスクを説明する指示があります。"
"リクエストを適切に完了するための回答を記述してください。\n\n"
"### 指示:\n{instruction}\n\n### 応答:"
),
}
def create_prompt(instruction, input=None):
"""
Generates a prompt based on the given instruction and an optional input.
If input is provided, it uses the 'prompt_input' template from PROMPT_DICT.
If no input is provided, it uses the 'prompt_no_input' template.
Args:
instruction (str): The instruction describing the task.
input (str, optional): Additional input providing context for the task. Default is None.
Returns:
str: The generated prompt.
"""
if input:
# Use the 'prompt_input' template when additional input is provided
return PROMPT_DICT["prompt_input"].format(instruction=instruction, input=input)
else:
# Use the 'prompt_no_input' template when no additional input is provided
return PROMPT_DICT["prompt_no_input"].format(instruction=instruction)
# Example usage
instruction_example = "以下のトピックに関する詳細な情報を提供してください。"
input_example = "東京工業大学の主なキャンパスについて教えてください"
prompt = create_prompt(instruction_example, input_example)
input_ids = tokenizer.encode(
prompt,
add_special_tokens=False,
return_tensors="pt"
)
tokens = model.generate(
input_ids.to(device=model.device),
max_new_tokens=128,
temperature=0.99,
top_p=0.95,
do_sample=True,
)
out = tokenizer.decode(tokens[0], skip_special_tokens=True)
print(out)
```
### Use the base model
```python
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
model_name = "tokyotech-llm/Swallow-7b-hf"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.bfloat16, device_map="auto")
prompt = "東京工業大学の主なキャンパスは、"
input_ids = tokenizer.encode(
prompt,
add_special_tokens=False,
return_tensors="pt"
)
tokens = model.generate(
input_ids.to(device=model.device),
max_new_tokens=128,
temperature=0.99,
top_p=0.95,
do_sample=True,
)
out = tokenizer.decode(tokens[0], skip_special_tokens=True)
print(out)
```
## Training Datasets
### Continual Pre-Training
The following datasets were used for continual pre-training.
- [Japanese Wikipedia](https://dumps.wikimedia.org/other/cirrussearch)
- [RefinedWeb](https://huggingface.co/datasets/tiiuae/falcon-refinedweb)
- Swallow Corpus
- [The Pile](https://huggingface.co/datasets/EleutherAI/pile)
### Instruction Tuning
The following datasets were used for the instruction tuning.
- [Anthropic HH-RLHF](https://huggingface.co/datasets/kunishou/hh-rlhf-49k-ja)
- [Databricks Dolly 15-k](https://huggingface.co/datasets/kunishou/databricks-dolly-15k-ja)
- [OpenAssistant Conversations Dataset](https://huggingface.co/datasets/kunishou/oasst1-89k-ja)
## Risks and Limitations
The models released here are still in the early stages of our research and development and have not been tuned to ensure outputs align with human intent and safety considerations.
## Acknowledgements
We thank Meta Research for releasing Llama 2 under an open license for others to build on.
Our project is supported by the [ABCI Large-scale Language Model Building Support Program](https://abci.ai/en/link/llm_support_program.html) of the National Institute of Advanced Industrial Science and Technology.
## License
Llama 2 is licensed under the LLAMA 2 Community License, Copyright © Meta Platforms, Inc. All Rights Reserved.
## Authors
Here are the team members:
- From [Okazaki Laboratory](https://www.nlp.c.titech.ac.jp/index.en.html), the following members:
- [Naoaki Okazaki](https://www.chokkan.org/index.ja.html)
- [Sakae Mizuki](https://s-mizuki-nlp.github.io/)
- [Hiroki Iida](https://meshidenn.github.io/)
- [Mengsay Loem](https://loem-ms.github.io/)
- [Shota Hirai](https://huggingface.co/Kotemo428)
- [Kakeru Hattori](https://aya-se.vercel.app/)
- [Masanari Ohi](https://twitter.com/stjohn2007)
- From [YOKOTA Laboratory](https://www.rio.gsic.titech.ac.jp/en/index.html), the following members:
- [Rio Yokota](https://twitter.com/rioyokota)
- [Kazuki Fujii](https://twitter.com/okoge_kaz)
- [Taishi Nakamura](https://twitter.com/Setuna7777_2)
|
jindig/t5-large_PREFIX_TUNING_SEQ2SEQ_es_data_abstracts
|
jindig
| 2023-12-20T01:11:08Z | 5 | 0 |
peft
|
[
"peft",
"safetensors",
"arxiv:1910.09700",
"base_model:google-t5/t5-large",
"base_model:adapter:google-t5/t5-large",
"region:us"
] | null | 2023-12-20T01:11:07Z |
---
library_name: peft
base_model: t5-large
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
### Framework versions
- PEFT 0.7.1
|
mosmodels/wav2vec_small
|
mosmodels
| 2023-12-20T01:10:21Z | 0 | 0 | null |
[
"license:mit",
"region:us"
] | null | 2023-12-20T00:54:29Z |
---
license: mit
---
https://github.com/facebookresearch/fairseq
MIT License
Copyright (c) Facebook, Inc. and its affiliates.
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.
|
zxh4546/unsup-wr-s64-bs128-lr6
|
zxh4546
| 2023-12-20T01:01:11Z | 3 | 0 |
transformers
|
[
"transformers",
"pytorch",
"pixel",
"text-classification",
"generated_from_trainer",
"dataset:unsup-wr",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2023-12-20T00:32:44Z |
---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- unsup-wr
model-index:
- name: contrastive-unsup-wr-pixel-base-mean-64-128-1-3e-6-7600-42-eval
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# contrastive-unsup-wr-pixel-base-mean-64-128-1-3e-6-7600-42-eval
This model is a fine-tuned version of [Team-PIXEL/pixel-base](https://huggingface.co/Team-PIXEL/pixel-base) on the UNSUP-WR dataset.
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 3e-06
- train_batch_size: 128
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 7600
### Training results
### Framework versions
- Transformers 4.17.0
- Pytorch 1.11.0
- Datasets 2.14.7.dev0
- Tokenizers 0.14.1
|
hkivancoral/smids_10x_deit_small_rms_001_fold3
|
hkivancoral
| 2023-12-20T00:58:55Z | 3 | 0 |
transformers
|
[
"transformers",
"pytorch",
"vit",
"image-classification",
"generated_from_trainer",
"dataset:imagefolder",
"base_model:facebook/deit-small-patch16-224",
"base_model:finetune:facebook/deit-small-patch16-224",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
image-classification
| 2023-12-20T00:03:41Z |
---
license: apache-2.0
base_model: facebook/deit-small-patch16-224
tags:
- generated_from_trainer
datasets:
- imagefolder
metrics:
- accuracy
model-index:
- name: smids_10x_deit_small_rms_001_fold3
results:
- task:
name: Image Classification
type: image-classification
dataset:
name: imagefolder
type: imagefolder
config: default
split: test
args: default
metrics:
- name: Accuracy
type: accuracy
value: 0.7766666666666666
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# smids_10x_deit_small_rms_001_fold3
This model is a fine-tuned version of [facebook/deit-small-patch16-224](https://huggingface.co/facebook/deit-small-patch16-224) on the imagefolder dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5590
- Accuracy: 0.7767
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.001
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 50
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|
| 0.8839 | 1.0 | 750 | 0.8956 | 0.4917 |
| 0.8402 | 2.0 | 1500 | 0.8459 | 0.5383 |
| 0.827 | 3.0 | 2250 | 0.8365 | 0.5417 |
| 0.7595 | 4.0 | 3000 | 0.8404 | 0.5617 |
| 0.8496 | 5.0 | 3750 | 0.9112 | 0.505 |
| 0.7825 | 6.0 | 4500 | 0.8246 | 0.6233 |
| 0.8185 | 7.0 | 5250 | 0.7843 | 0.6233 |
| 0.7863 | 8.0 | 6000 | 0.7862 | 0.6183 |
| 0.7304 | 9.0 | 6750 | 0.7478 | 0.6433 |
| 0.7486 | 10.0 | 7500 | 0.7941 | 0.625 |
| 0.7979 | 11.0 | 8250 | 0.7438 | 0.6817 |
| 0.6928 | 12.0 | 9000 | 0.8898 | 0.58 |
| 0.683 | 13.0 | 9750 | 0.7126 | 0.68 |
| 0.7194 | 14.0 | 10500 | 0.7634 | 0.6367 |
| 0.7001 | 15.0 | 11250 | 0.6906 | 0.68 |
| 0.7209 | 16.0 | 12000 | 0.6988 | 0.675 |
| 0.693 | 17.0 | 12750 | 0.7227 | 0.6733 |
| 0.6594 | 18.0 | 13500 | 0.7119 | 0.675 |
| 0.6733 | 19.0 | 14250 | 0.6769 | 0.695 |
| 0.6368 | 20.0 | 15000 | 0.6310 | 0.7183 |
| 0.5529 | 21.0 | 15750 | 0.6379 | 0.73 |
| 0.674 | 22.0 | 16500 | 0.6200 | 0.7233 |
| 0.6173 | 23.0 | 17250 | 0.6390 | 0.7117 |
| 0.7017 | 24.0 | 18000 | 0.6234 | 0.7217 |
| 0.6672 | 25.0 | 18750 | 0.6159 | 0.7117 |
| 0.6143 | 26.0 | 19500 | 0.6119 | 0.7133 |
| 0.5447 | 27.0 | 20250 | 0.6511 | 0.7 |
| 0.616 | 28.0 | 21000 | 0.5943 | 0.7317 |
| 0.6257 | 29.0 | 21750 | 0.6135 | 0.7417 |
| 0.5784 | 30.0 | 22500 | 0.6236 | 0.7383 |
| 0.5488 | 31.0 | 23250 | 0.5814 | 0.7483 |
| 0.5683 | 32.0 | 24000 | 0.6409 | 0.725 |
| 0.5657 | 33.0 | 24750 | 0.6193 | 0.7583 |
| 0.7061 | 34.0 | 25500 | 0.7958 | 0.6533 |
| 0.5815 | 35.0 | 26250 | 0.6092 | 0.7467 |
| 0.545 | 36.0 | 27000 | 0.5902 | 0.7567 |
| 0.574 | 37.0 | 27750 | 0.5865 | 0.7483 |
| 0.5654 | 38.0 | 28500 | 0.6161 | 0.7467 |
| 0.5393 | 39.0 | 29250 | 0.5677 | 0.7667 |
| 0.6213 | 40.0 | 30000 | 0.5702 | 0.7633 |
| 0.5565 | 41.0 | 30750 | 0.5675 | 0.75 |
| 0.5323 | 42.0 | 31500 | 0.5645 | 0.7583 |
| 0.5444 | 43.0 | 32250 | 0.5820 | 0.76 |
| 0.4988 | 44.0 | 33000 | 0.5588 | 0.765 |
| 0.5249 | 45.0 | 33750 | 0.5669 | 0.7583 |
| 0.5246 | 46.0 | 34500 | 0.5504 | 0.7733 |
| 0.4975 | 47.0 | 35250 | 0.5697 | 0.7717 |
| 0.5083 | 48.0 | 36000 | 0.5554 | 0.7717 |
| 0.4948 | 49.0 | 36750 | 0.5551 | 0.775 |
| 0.4147 | 50.0 | 37500 | 0.5590 | 0.7767 |
### Framework versions
- Transformers 4.32.1
- Pytorch 2.1.0+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2
|
zxh4546/unsup-wc-s144-bs128-lr6
|
zxh4546
| 2023-12-20T00:57:07Z | 1 | 0 |
transformers
|
[
"transformers",
"pytorch",
"pixel",
"text-classification",
"generated_from_trainer",
"dataset:unsup-wc",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2023-12-20T00:33:05Z |
---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- unsup-wc
model-index:
- name: contrastive-unsup-wc-pixel-base-mean-144-128-1-3e-6-7600-42-eval
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# contrastive-unsup-wc-pixel-base-mean-144-128-1-3e-6-7600-42-eval
This model is a fine-tuned version of [Team-PIXEL/pixel-base](https://huggingface.co/Team-PIXEL/pixel-base) on the UNSUP-WC dataset.
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 3e-06
- train_batch_size: 128
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 7600
### Training results
### Framework versions
- Transformers 4.17.0
- Pytorch 1.11.0
- Datasets 2.14.7.dev0
- Tokenizers 0.14.1
|
kitsonr/ppo-BipedalWalker-v3
|
kitsonr
| 2023-12-20T00:54:37Z | 1 | 0 |
stable-baselines3
|
[
"stable-baselines3",
"BipedalWalker-v3",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-12-20T00:54:01Z |
---
library_name: stable-baselines3
tags:
- BipedalWalker-v3
- deep-reinforcement-learning
- reinforcement-learning
- stable-baselines3
model-index:
- name: PPO
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: BipedalWalker-v3
type: BipedalWalker-v3
metrics:
- type: mean_reward
value: 71.26 +/- 39.69
name: mean_reward
verified: false
---
# **PPO** Agent playing **BipedalWalker-v3**
This is a trained model of a **PPO** agent playing **BipedalWalker-v3**
using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3).
## Usage (with Stable-baselines3)
TODO: Add your code
```python
from stable_baselines3 import ...
from huggingface_sb3 import load_from_hub
...
```
|
Osquery/alberta-te-pos
|
Osquery
| 2023-12-20T00:34:31Z | 4 | 0 |
transformers
|
[
"transformers",
"safetensors",
"t5",
"text2text-generation",
"generated_from_trainer",
"dataset:universal_dependencies",
"base_model:google-t5/t5-small",
"base_model:finetune:google-t5/t5-small",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2023-12-20T00:34:09Z |
---
license: apache-2.0
base_model: t5-small
tags:
- generated_from_trainer
datasets:
- universal_dependencies
model-index:
- name: alberta-te-pos
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# alberta-te-pos
This model is a fine-tuned version of [t5-small](https://huggingface.co/t5-small) on the universal_dependencies dataset.
It achieves the following results on the evaluation set:
- Loss: 0.7232
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| No log | 1.0 | 252 | 1.1029 |
| 2.4271 | 2.0 | 504 | 0.7232 |
### Framework versions
- Transformers 4.36.0.dev0
- Pytorch 2.1.0+cu118
- Datasets 2.15.0
- Tokenizers 0.15.0
|
matthewchung74/distilbert-base-uncased-finetuned-cola
|
matthewchung74
| 2023-12-20T00:28:34Z | 7 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"distilbert",
"text-classification",
"generated_from_trainer",
"dataset:glue",
"base_model:distilbert/distilbert-base-uncased",
"base_model:finetune:distilbert/distilbert-base-uncased",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2023-12-19T23:45:16Z |
---
license: apache-2.0
base_model: distilbert-base-uncased
tags:
- generated_from_trainer
datasets:
- glue
metrics:
- matthews_correlation
model-index:
- name: distilbert-base-uncased-finetuned-cola
results:
- task:
name: Text Classification
type: text-classification
dataset:
name: glue
type: glue
config: cola
split: validation
args: cola
metrics:
- name: Matthews Correlation
type: matthews_correlation
value: 0.43482765994004224
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# distilbert-base-uncased-finetuned-cola
This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the glue dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4872
- Matthews Correlation: 0.4348
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 1
### Training results
| Training Loss | Epoch | Step | Validation Loss | Matthews Correlation |
|:-------------:|:-----:|:----:|:---------------:|:--------------------:|
| 0.5163 | 1.0 | 535 | 0.4872 | 0.4348 |
### Framework versions
- Transformers 4.36.2
- Pytorch 2.1.0+cu121
- Datasets 2.15.0
- Tokenizers 0.15.0
|
hkivancoral/smids_10x_deit_small_rms_001_fold2
|
hkivancoral
| 2023-12-20T00:03:09Z | 3 | 0 |
transformers
|
[
"transformers",
"pytorch",
"vit",
"image-classification",
"generated_from_trainer",
"dataset:imagefolder",
"base_model:facebook/deit-small-patch16-224",
"base_model:finetune:facebook/deit-small-patch16-224",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
image-classification
| 2023-12-19T23:08:02Z |
---
license: apache-2.0
base_model: facebook/deit-small-patch16-224
tags:
- generated_from_trainer
datasets:
- imagefolder
metrics:
- accuracy
model-index:
- name: smids_10x_deit_small_rms_001_fold2
results:
- task:
name: Image Classification
type: image-classification
dataset:
name: imagefolder
type: imagefolder
config: default
split: test
args: default
metrics:
- name: Accuracy
type: accuracy
value: 0.8369384359400999
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# smids_10x_deit_small_rms_001_fold2
This model is a fine-tuned version of [facebook/deit-small-patch16-224](https://huggingface.co/facebook/deit-small-patch16-224) on the imagefolder dataset.
It achieves the following results on the evaluation set:
- Loss: 0.7042
- Accuracy: 0.8369
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.001
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 50
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|
| 0.8864 | 1.0 | 750 | 0.8248 | 0.5474 |
| 0.7722 | 2.0 | 1500 | 0.8988 | 0.5092 |
| 0.7721 | 3.0 | 2250 | 0.7604 | 0.6456 |
| 0.7086 | 4.0 | 3000 | 0.6560 | 0.7371 |
| 0.6588 | 5.0 | 3750 | 0.6906 | 0.7088 |
| 0.5657 | 6.0 | 4500 | 0.5964 | 0.7654 |
| 0.5826 | 7.0 | 5250 | 0.5186 | 0.7854 |
| 0.5637 | 8.0 | 6000 | 0.5513 | 0.7737 |
| 0.5395 | 9.0 | 6750 | 0.5704 | 0.7537 |
| 0.5342 | 10.0 | 7500 | 0.4931 | 0.7987 |
| 0.5349 | 11.0 | 8250 | 0.5109 | 0.7937 |
| 0.596 | 12.0 | 9000 | 0.5425 | 0.7804 |
| 0.5878 | 13.0 | 9750 | 0.4766 | 0.8103 |
| 0.4609 | 14.0 | 10500 | 0.7520 | 0.7022 |
| 0.4491 | 15.0 | 11250 | 0.5442 | 0.7504 |
| 0.4637 | 16.0 | 12000 | 0.5054 | 0.8136 |
| 0.4699 | 17.0 | 12750 | 0.4927 | 0.8037 |
| 0.4528 | 18.0 | 13500 | 0.4576 | 0.8120 |
| 0.4797 | 19.0 | 14250 | 0.4748 | 0.7970 |
| 0.4704 | 20.0 | 15000 | 0.4438 | 0.8070 |
| 0.4406 | 21.0 | 15750 | 0.4383 | 0.8153 |
| 0.4289 | 22.0 | 16500 | 0.4522 | 0.8120 |
| 0.4219 | 23.0 | 17250 | 0.4457 | 0.8286 |
| 0.3979 | 24.0 | 18000 | 0.4791 | 0.8203 |
| 0.476 | 25.0 | 18750 | 0.4867 | 0.8136 |
| 0.4039 | 26.0 | 19500 | 0.4638 | 0.8319 |
| 0.4302 | 27.0 | 20250 | 0.4222 | 0.8303 |
| 0.4091 | 28.0 | 21000 | 0.4516 | 0.8270 |
| 0.3603 | 29.0 | 21750 | 0.5085 | 0.8170 |
| 0.4414 | 30.0 | 22500 | 0.4568 | 0.8353 |
| 0.3768 | 31.0 | 23250 | 0.4984 | 0.8253 |
| 0.3126 | 32.0 | 24000 | 0.4428 | 0.8436 |
| 0.3269 | 33.0 | 24750 | 0.4871 | 0.8236 |
| 0.3283 | 34.0 | 25500 | 0.4708 | 0.8253 |
| 0.3471 | 35.0 | 26250 | 0.4869 | 0.8353 |
| 0.3619 | 36.0 | 27000 | 0.5210 | 0.8153 |
| 0.4176 | 37.0 | 27750 | 0.4744 | 0.8353 |
| 0.3395 | 38.0 | 28500 | 0.5334 | 0.8386 |
| 0.2458 | 39.0 | 29250 | 0.5218 | 0.8286 |
| 0.3331 | 40.0 | 30000 | 0.5874 | 0.8186 |
| 0.3063 | 41.0 | 30750 | 0.5488 | 0.8236 |
| 0.2956 | 42.0 | 31500 | 0.5739 | 0.8220 |
| 0.3105 | 43.0 | 32250 | 0.5441 | 0.8369 |
| 0.2918 | 44.0 | 33000 | 0.6039 | 0.8303 |
| 0.2418 | 45.0 | 33750 | 0.6214 | 0.8303 |
| 0.2859 | 46.0 | 34500 | 0.6601 | 0.8286 |
| 0.2507 | 47.0 | 35250 | 0.6435 | 0.8369 |
| 0.2443 | 48.0 | 36000 | 0.6789 | 0.8336 |
| 0.2825 | 49.0 | 36750 | 0.6931 | 0.8336 |
| 0.1845 | 50.0 | 37500 | 0.7042 | 0.8369 |
### Framework versions
- Transformers 4.32.1
- Pytorch 2.1.0+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2
|
LoftQ/Mistral-7B-v0.1-4bit-32rank
|
LoftQ
| 2023-12-20T00:01:17Z | 51 | 0 |
transformers
|
[
"transformers",
"safetensors",
"mistral",
"text-generation",
"quantization ",
"lora",
"en",
"arxiv:2310.08659",
"license:mit",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2023-11-22T05:05:56Z |
---
license: mit
language:
- en
pipeline_tag: text-generation
tags:
- 'quantization '
- lora
---
# LoftQ Initialization
| [Paper](https://arxiv.org/abs/2310.08659) | [Code](https://github.com/yxli2123/LoftQ) | [PEFT Example](https://github.com/huggingface/peft/tree/main/examples/loftq_finetuning) |
LoftQ (LoRA-fine-tuning-aware Quantization) provides a quantized backbone Q and LoRA adapters A and B, given a full-precision pre-trained weight W.
This model, `Mistral-7B-v0.1-4bit-32rank`, is obtained from [Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1).
The backbone is under `LoftQ/Mistral-7B-v0.1-4bit-32rank` and LoRA adapters are under the `subfolder='loftq_init'`.
## Model Info
### Backbone
- Stored format: `torch.bfloat16`
- Size: ~ 14 GiB
- Loaded format: bitsandbytes nf4
- Size loaded on GPU: ~3.5 GiB
### LoRA adapters
- rank: 32
- lora_alpha: 16
- target_modules: ["down_proj", "up_proj", "q_proj", "k_proj", "v_proj", "o_proj", "gate_proj"]
## Usage
**Training.** Here's an example of loading this model and preparing for the LoRA fine-tuning.
```python
import torch
from transformers import AutoModelForCausalLM, BitsAndBytesConfig
from peft import PeftModel
MODEL_ID = "LoftQ/Mistral-7B-v0.1-4bit-32rank"
base_model = AutoModelForCausalLM.from_pretrained(
MODEL_ID,
torch_dtype=torch.bfloat16, # you may change it with different models
quantization_config=BitsAndBytesConfig(
load_in_4bit=True,
bnb_4bit_compute_dtype=torch.bfloat16, # bfloat16 is recommended
bnb_4bit_use_double_quant=False,
bnb_4bit_quant_type='nf4',
),
)
peft_model = PeftModel.from_pretrained(
base_model,
MODEL_ID,
subfolder="loftq_init",
is_trainable=True,
)
# Do training with peft_model ...
```
See the full code at our [Github Repo]((https://github.com/yxli2123/LoftQ))
## Citation
```bibtex
@article{li2023loftq,
title={Loftq: Lora-fine-tuning-aware quantization for large language models},
author={Li, Yixiao and Yu, Yifan and Liang, Chen and He, Pengcheng and Karampatziakis, Nikos and Chen, Weizhu and Zhao, Tuo},
journal={arXiv preprint arXiv:2310.08659},
year={2023}
}
```
|
LoftQ/Llama-2-13b-hf-4bit-64rank
|
LoftQ
| 2023-12-19T23:54:28Z | 35 | 2 |
transformers
|
[
"transformers",
"safetensors",
"llama",
"text-generation",
"quantization ",
"lora",
"en",
"arxiv:2310.08659",
"license:mit",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2023-11-21T07:47:20Z |
---
license: mit
language:
- en
pipeline_tag: text-generation
tags:
- 'quantization '
- lora
---
# LoftQ Initialization
| [Paper](https://arxiv.org/abs/2310.08659) | [Code](https://github.com/yxli2123/LoftQ) | [PEFT Example](https://github.com/huggingface/peft/tree/main/examples/loftq_finetuning) |
LoftQ (LoRA-fine-tuning-aware Quantization) provides a quantized backbone Q and LoRA adapters A and B, given a full-precision pre-trained weight W.
This model, `Llama-2-13b-hf-4bit-64rank`, is obtained from [LLAMA-2-13b](https://huggingface.co/meta-llama/Llama-2-13b-hf).
The backbone is under `LoftQ/Llama-2-13b-hf-4bit-64rank` and LoRA adapters are under the `subfolder='loftq_init'`.
## Model Info
### Backbone
- Stored format: `torch.bfloat16`
- Size: ~ 26 GiB
- Loaded format: bitsandbytes nf4
- Size loaded on GPU: ~6.5 GiB
### LoRA adapters
- rank: 64
- lora_alpha: 64
- target_modules: ["down_proj", "up_proj", "q_proj", "k_proj", "v_proj", "o_proj", "gate_proj"]
## Usage
**Training** Here's an example of loading this model and preparing for the LoRA fine-tuning.
```python
import torch
from transformers import AutoModelForCausalLM, BitsAndBytesConfig
from peft import PeftModel
MODEL_ID = "LoftQ/Llama-2-13b-hf-4bit-64rank"
base_model = AutoModelForCausalLM.from_pretrained(
MODEL_ID,
torch_dtype=torch.bfloat16, # you may change it with different models
quantization_config=BitsAndBytesConfig(
load_in_4bit=True,
bnb_4bit_compute_dtype=torch.bfloat16, # bfloat16 is recommended
bnb_4bit_use_double_quant=False,
bnb_4bit_quant_type='nf4',
),
)
peft_model = PeftModel.from_pretrained(
base_model,
MODEL_ID,
subfolder="loftq_init",
is_trainable=True,
)
# Do training with peft_model ...
```
## Experiment Results
We have conducted experiments on supervised fine-tuning of [GSM8K](https://huggingface.co/datasets/gsm8k)
and [WikiText-2](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-raw-v1).
| Model | Bits | Rank | LoRA Initial | GSM8K | WikiText-2 |
| -------------- | ---- | ---- | -------------------- | ----- | ---------- |
| LLAMA-2-13b | 16 | 64 | Gaussian + 0 | 45.3 | 5.12 |
| LLAMA-2-13b | 4 | 64 | Gaussian + 0 (QLoRA) | 39.9 | 5.22 |
| **LLAMA-2-13b** | 4 | 64 | LoftQ | 45.0 | 5.16 |
**Inference** Here is an example code for inference after the model has been fine-tuned on [GSM8K](https://huggingface.co/datasets/gsm8k).
```python
import torch
from transformers import AutoModelForCausalLM, BitsAndBytesConfig
from peft import PeftModel
MODEL_ID = "LoftQ/Llama-2-13b-hf-4bit-64rank"
base_model = AutoModelForCausalLM.from_pretrained(
MODEL_ID,
torch_dtype=torch.bfloat16, # you may change it with different models
quantization_config=BitsAndBytesConfig(
load_in_4bit=True,
bnb_4bit_compute_dtype=torch.bfloat16, # bfloat16 is recommended
bnb_4bit_use_double_quant=False,
bnb_4bit_quant_type='nf4',
),
)
peft_model = PeftModel.from_pretrained(
base_model,
MODEL_ID,
subfolder="gsm8k",
is_trainable=True,
)
# Do inference with peft_model ...
```
See the full code at our [Github Repo]((https://github.com/yxli2123/LoftQ))
## Citation
```bibtex
@article{li2023loftq,
title={Loftq: Lora-fine-tuning-aware quantization for large language models},
author={Li, Yixiao and Yu, Yifan and Liang, Chen and He, Pengcheng and Karampatziakis, Nikos and Chen, Weizhu and Zhao, Tuo},
journal={arXiv preprint arXiv:2310.08659},
year={2023}
}
```
|
TheBloke/Swallow-7B-Instruct-AWQ
|
TheBloke
| 2023-12-19T23:44:07Z | 8 | 1 |
transformers
|
[
"transformers",
"safetensors",
"llama",
"text-generation",
"en",
"ja",
"base_model:tokyotech-llm/Swallow-7b-instruct-hf",
"base_model:quantized:tokyotech-llm/Swallow-7b-instruct-hf",
"license:llama2",
"autotrain_compatible",
"text-generation-inference",
"4-bit",
"awq",
"region:us"
] |
text-generation
| 2023-12-19T23:29:43Z |
---
base_model: tokyotech-llm/Swallow-7b-instruct-hf
inference: false
language:
- en
- ja
library_name: transformers
license: llama2
model_creator: tokyotech-llm
model_name: Swallow 7B Instruct
model_type: llama
pipeline_tag: text-generation
prompt_template: "\u4EE5\u4E0B\u306B\u3001\u3042\u308B\u30BF\u30B9\u30AF\u3092\u8AAC\
\u660E\u3059\u308B\u6307\u793A\u304C\u3042\u308A\u307E\u3059\u3002\u30EA\u30AF\u30A8\
\u30B9\u30C8\u3092\u9069\u5207\u306B\u5B8C\u4E86\u3059\u308B\u305F\u3081\u306E\u56DE\
\u7B54\u3092\u8A18\u8FF0\u3057\u3066\u304F\u3060\u3055\u3044\u3002\\n\\n### \u6307\
\u793A:\\n{prompt}\\n\\n### \u5FDC\u7B54:\n"
quantized_by: TheBloke
---
<!-- markdownlint-disable MD041 -->
<!-- header start -->
<!-- 200823 -->
<div style="width: auto; margin-left: auto; margin-right: auto">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<div style="text-align:center; margin-top: 0em; margin-bottom: 0em"><p style="margin-top: 0.25em; margin-bottom: 0em;">TheBloke's LLM work is generously supported by a grant from <a href="https://a16z.com">andreessen horowitz (a16z)</a></p></div>
<hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
<!-- header end -->
# Swallow 7B Instruct - AWQ
- Model creator: [tokyotech-llm](https://huggingface.co/tokyotech-llm)
- Original model: [Swallow 7B Instruct](https://huggingface.co/tokyotech-llm/Swallow-7b-instruct-hf)
<!-- description start -->
## Description
This repo contains AWQ model files for [tokyotech-llm's Swallow 7B Instruct](https://huggingface.co/tokyotech-llm/Swallow-7b-instruct-hf).
These files were quantised using hardware kindly provided by [Massed Compute](https://massedcompute.com/).
### About AWQ
AWQ is an efficient, accurate and blazing-fast low-bit weight quantization method, currently supporting 4-bit quantization. Compared to GPTQ, it offers faster Transformers-based inference with equivalent or better quality compared to the most commonly used GPTQ settings.
AWQ models are currently supported on Linux and Windows, with NVidia GPUs only. macOS users: please use GGUF models instead.
It is supported by:
- [Text Generation Webui](https://github.com/oobabooga/text-generation-webui) - using Loader: AutoAWQ
- [vLLM](https://github.com/vllm-project/vllm) - version 0.2.2 or later for support for all model types.
- [Hugging Face Text Generation Inference (TGI)](https://github.com/huggingface/text-generation-inference)
- [Transformers](https://huggingface.co/docs/transformers) version 4.35.0 and later, from any code or client that supports Transformers
- [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) - for use from Python code
<!-- description end -->
<!-- repositories-available start -->
## Repositories available
* [AWQ model(s) for GPU inference.](https://huggingface.co/TheBloke/Swallow-7B-Instruct-AWQ)
* [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/Swallow-7B-Instruct-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/Swallow-7B-Instruct-GGUF)
* [tokyotech-llm's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/tokyotech-llm/Swallow-7b-instruct-hf)
<!-- repositories-available end -->
<!-- prompt-template start -->
## Prompt template: Swallow-Instruct
```
以下に、あるタスクを説明する指示があります。リクエストを適切に完了するための回答を記述してください。\n\n### 指示:\n{prompt}\n\n### 応答:
```
<!-- prompt-template end -->
<!-- README_AWQ.md-provided-files start -->
## Provided files, and AWQ parameters
I currently release 128g GEMM models only. The addition of group_size 32 models, and GEMV kernel models, is being actively considered.
Models are released as sharded safetensors files.
| Branch | Bits | GS | AWQ Dataset | Seq Len | Size |
| ------ | ---- | -- | ----------- | ------- | ---- |
| [main](https://huggingface.co/TheBloke/Swallow-7B-Instruct-AWQ/tree/main) | 4 | 128 | [Alpaca Japanese](https://huggingface.co/datasets/fujiki/japanese_alpaca_data/viewer/) | 4096 | 4.07 GB
<!-- README_AWQ.md-provided-files end -->
<!-- README_AWQ.md-text-generation-webui start -->
## How to easily download and use this model in [text-generation-webui](https://github.com/oobabooga/text-generation-webui)
Please make sure you're using the latest version of [text-generation-webui](https://github.com/oobabooga/text-generation-webui).
It is strongly recommended to use the text-generation-webui one-click-installers unless you're sure you know how to make a manual install.
1. Click the **Model tab**.
2. Under **Download custom model or LoRA**, enter `TheBloke/Swallow-7B-Instruct-AWQ`.
3. Click **Download**.
4. The model will start downloading. Once it's finished it will say "Done".
5. In the top left, click the refresh icon next to **Model**.
6. In the **Model** dropdown, choose the model you just downloaded: `Swallow-7B-Instruct-AWQ`
7. Select **Loader: AutoAWQ**.
8. Click Load, and the model will load and is now ready for use.
9. If you want any custom settings, set them and then click **Save settings for this model** followed by **Reload the Model** in the top right.
10. Once you're ready, click the **Text Generation** tab and enter a prompt to get started!
<!-- README_AWQ.md-text-generation-webui end -->
<!-- README_AWQ.md-use-from-vllm start -->
## Multi-user inference server: vLLM
Documentation on installing and using vLLM [can be found here](https://vllm.readthedocs.io/en/latest/).
- Please ensure you are using vLLM version 0.2 or later.
- When using vLLM as a server, pass the `--quantization awq` parameter.
For example:
```shell
python3 -m vllm.entrypoints.api_server --model TheBloke/Swallow-7B-Instruct-AWQ --quantization awq --dtype auto
```
- When using vLLM from Python code, again set `quantization=awq`.
For example:
```python
from vllm import LLM, SamplingParams
prompts = [
"Tell me about AI",
"Write a story about llamas",
"What is 291 - 150?",
"How much wood would a woodchuck chuck if a woodchuck could chuck wood?",
]
prompt_template=f'''以下に、あるタスクを説明する指示があります。リクエストを適切に完了するための回答を記述してください。\n\n### 指示:\n{prompt}\n\n### 応答:
'''
prompts = [prompt_template.format(prompt=prompt) for prompt in prompts]
sampling_params = SamplingParams(temperature=0.8, top_p=0.95)
llm = LLM(model="TheBloke/Swallow-7B-Instruct-AWQ", quantization="awq", dtype="auto")
outputs = llm.generate(prompts, sampling_params)
# Print the outputs.
for output in outputs:
prompt = output.prompt
generated_text = output.outputs[0].text
print(f"Prompt: {prompt!r}, Generated text: {generated_text!r}")
```
<!-- README_AWQ.md-use-from-vllm start -->
<!-- README_AWQ.md-use-from-tgi start -->
## Multi-user inference server: Hugging Face Text Generation Inference (TGI)
Use TGI version 1.1.0 or later. The official Docker container is: `ghcr.io/huggingface/text-generation-inference:1.1.0`
Example Docker parameters:
```shell
--model-id TheBloke/Swallow-7B-Instruct-AWQ --port 3000 --quantize awq --max-input-length 3696 --max-total-tokens 4096 --max-batch-prefill-tokens 4096
```
Example Python code for interfacing with TGI (requires [huggingface-hub](https://github.com/huggingface/huggingface_hub) 0.17.0 or later):
```shell
pip3 install huggingface-hub
```
```python
from huggingface_hub import InferenceClient
endpoint_url = "https://your-endpoint-url-here"
prompt = "Tell me about AI"
prompt_template=f'''以下に、あるタスクを説明する指示があります。リクエストを適切に完了するための回答を記述してください。\n\n### 指示:\n{prompt}\n\n### 応答:
'''
client = InferenceClient(endpoint_url)
response = client.text_generation(prompt,
max_new_tokens=128,
do_sample=True,
temperature=0.7,
top_p=0.95,
top_k=40,
repetition_penalty=1.1)
print(f"Model output: ", response)
```
<!-- README_AWQ.md-use-from-tgi end -->
<!-- README_AWQ.md-use-from-python start -->
## Inference from Python code using Transformers
### Install the necessary packages
- Requires: [Transformers](https://huggingface.co/docs/transformers) 4.35.0 or later.
- Requires: [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) 0.1.6 or later.
```shell
pip3 install --upgrade "autoawq>=0.1.6" "transformers>=4.35.0"
```
Note that if you are using PyTorch 2.0.1, the above AutoAWQ command will automatically upgrade you to PyTorch 2.1.0.
If you are using CUDA 11.8 and wish to continue using PyTorch 2.0.1, instead run this command:
```shell
pip3 install https://github.com/casper-hansen/AutoAWQ/releases/download/v0.1.6/autoawq-0.1.6+cu118-cp310-cp310-linux_x86_64.whl
```
If you have problems installing [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) using the pre-built wheels, install it from source instead:
```shell
pip3 uninstall -y autoawq
git clone https://github.com/casper-hansen/AutoAWQ
cd AutoAWQ
pip3 install .
```
### Transformers example code (requires Transformers 4.35.0 and later)
```python
from transformers import AutoModelForCausalLM, AutoTokenizer, TextStreamer
model_name_or_path = "TheBloke/Swallow-7B-Instruct-AWQ"
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path)
model = AutoModelForCausalLM.from_pretrained(
model_name_or_path,
low_cpu_mem_usage=True,
device_map="cuda:0"
)
# Using the text streamer to stream output one token at a time
streamer = TextStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True)
prompt = "Tell me about AI"
prompt_template=f'''以下に、あるタスクを説明する指示があります。リクエストを適切に完了するための回答を記述してください。\n\n### 指示:\n{prompt}\n\n### 応答:
'''
# Convert prompt to tokens
tokens = tokenizer(
prompt_template,
return_tensors='pt'
).input_ids.cuda()
generation_params = {
"do_sample": True,
"temperature": 0.7,
"top_p": 0.95,
"top_k": 40,
"max_new_tokens": 512,
"repetition_penalty": 1.1
}
# Generate streamed output, visible one token at a time
generation_output = model.generate(
tokens,
streamer=streamer,
**generation_params
)
# Generation without a streamer, which will include the prompt in the output
generation_output = model.generate(
tokens,
**generation_params
)
# Get the tokens from the output, decode them, print them
token_output = generation_output[0]
text_output = tokenizer.decode(token_output)
print("model.generate output: ", text_output)
# Inference is also possible via Transformers' pipeline
from transformers import pipeline
pipe = pipeline(
"text-generation",
model=model,
tokenizer=tokenizer,
**generation_params
)
pipe_output = pipe(prompt_template)[0]['generated_text']
print("pipeline output: ", pipe_output)
```
<!-- README_AWQ.md-use-from-python end -->
<!-- README_AWQ.md-compatibility start -->
## Compatibility
The files provided are tested to work with:
- [text-generation-webui](https://github.com/oobabooga/text-generation-webui) using `Loader: AutoAWQ`.
- [vLLM](https://github.com/vllm-project/vllm) version 0.2.0 and later.
- [Hugging Face Text Generation Inference (TGI)](https://github.com/huggingface/text-generation-inference) version 1.1.0 and later.
- [Transformers](https://huggingface.co/docs/transformers) version 4.35.0 and later.
- [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) version 0.1.1 and later.
<!-- README_AWQ.md-compatibility end -->
<!-- footer start -->
<!-- 200823 -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute
Thanks to the [chirper.ai](https://chirper.ai) team!
Thanks to Clay from [gpus.llm-utils.org](llm-utils)!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Aemon Algiz.
**Patreon special mentions**: Michael Levine, 阿明, Trailburnt, Nikolai Manek, John Detwiler, Randy H, Will Dee, Sebastain Graf, NimbleBox.ai, Eugene Pentland, Emad Mostaque, Ai Maven, Jim Angel, Jeff Scroggin, Michael Davis, Manuel Alberto Morcote, Stephen Murray, Robert, Justin Joy, Luke @flexchar, Brandon Frisco, Elijah Stavena, S_X, Dan Guido, Undi ., Komninos Chatzipapas, Shadi, theTransient, Lone Striker, Raven Klaugh, jjj, Cap'n Zoog, Michel-Marie MAUDET (LINAGORA), Matthew Berman, David, Fen Risland, Omer Bin Jawed, Luke Pendergrass, Kalila, OG, Erik Bjäreholt, Rooh Singh, Joseph William Delisle, Dan Lewis, TL, John Villwock, AzureBlack, Brad, Pedro Madruga, Caitlyn Gatomon, K, jinyuan sun, Mano Prime, Alex, Jeffrey Morgan, Alicia Loh, Illia Dulskyi, Chadd, transmissions 11, fincy, Rainer Wilmers, ReadyPlayerEmma, knownsqashed, Mandus, biorpg, Deo Leter, Brandon Phillips, SuperWojo, Sean Connelly, Iucharbius, Jack West, Harry Royden McLaughlin, Nicholas, terasurfer, Vitor Caleffi, Duane Dunston, Johann-Peter Hartmann, David Ziegler, Olakabola, Ken Nordquist, Trenton Dambrowitz, Tom X Nguyen, Vadim, Ajan Kanaga, Leonard Tan, Clay Pascal, Alexandros Triantafyllidis, JM33133, Xule, vamX, ya boyyy, subjectnull, Talal Aujan, Alps Aficionado, wassieverse, Ari Malik, James Bentley, Woland, Spencer Kim, Michael Dempsey, Fred von Graf, Elle, zynix, William Richards, Stanislav Ovsiannikov, Edmond Seymore, Jonathan Leane, Martin Kemka, usrbinkat, Enrico Ros
Thank you to all my generous patrons and donaters!
And thank you again to a16z for their generous grant.
<!-- footer end -->
# Original model card: tokyotech-llm's Swallow 7B Instruct
# Swallow
Our Swallow model has undergone continuous pre-training from the Llama 2 family, primarily with the addition of Japanese language data. The tuned versions use supervised fine-tuning (SFT).
Links to other models can be found in the index.
## Swallow Model Index
|Model|Swallow-hf|Swallow-instruct-hf|
|---|---|---|
|7B| [Link](https://huggingface.co/tokyotech-llm/Swallow-7b-hf) | [Link](https://huggingface.co/tokyotech-llm/Swallow-7b-instruct-hf)|
|13B| [Link](https://huggingface.co/tokyotech-llm/Swallow-13b-hf) | [Link](https://huggingface.co/tokyotech-llm/Swallow-13b-instruct-hf)|
|70B| [Link](https://huggingface.co/tokyotech-llm/Swallow-70b-hf) | [Link](https://huggingface.co/tokyotech-llm/Swallow-70b-instruct-hf)|
This repository provides large language models developed by [TokyoTech-LLM](https://tokyotech-llm.github.io/).
Read our [blog post](https://zenn.dev/tokyotech_lm/articles/d6cb3a8fdfc907) or our paper (preprint coming soon) for more details!
## Model Details
* **Model type**: Please refer to LLaMA-2 technical report for details on the model architecture.
* **Language(s)**: Japanese English
* **Library**: [Megatron-LM](https://github.com/rioyokotalab/Megatron-Llama2)
* **Tokenizer**: This model employs a tokenizer that features a broadened vocabulary based on Japanese data. This allows for a more efficient representation of text using fewer tokens, leading to a notably faster inference process.
* **Contact**: swallow[at]nlp.c.titech.ac.jp
## Base Model Performance
### Japanese version
|Model|Size|JCommonsenseQA|JEMHopQA|NIILC|JSQuAD|XL-Sum|MGSM|WMT20-en-ja|WMT20-ja-en|
|---|---|---|---|---|---|---|---|---|---|
| | |4-shot|4-shot|4-shot|4-shot|1-shot|4-shot|4-shot|4-shot|
|Llama 2|7B|0.3852|0.4240|0.3410|0.7917|0.1905|0.0760|0.1783|0.1738|
|Swallow|7B|0.4808|0.5078|0.5968|0.8573|0.1830|0.1240|0.2510|0.1511|
|Llama 2|13B|0.6997|0.4415|0.4170|0.8533|0.2139|0.1320|0.2146|0.1982|
|Swallow|13B|0.7837|0.5063|0.6398|0.9005|0.2168|0.2040|0.2720|0.1771|
|Llama 2|70B|0.8686|0.4656|0.5256|0.9080|**0.2361**|0.3560|0.2643|**0.2398**|
|Swallow|70B|**0.9348**|**0.6290**|**0.6960**|**0.9176**|0.2266|**0.4840**|**0.3043**|0.2298|
## Usage
First install additional dependencies in [requirements.txt](./requirements.txt):
```sh
pip install -r requirements.txt
```
### Use the instruct model
```python
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
model_name = "tokyotech-llm/Swallow-7b-instruct-hf"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.bfloat16, low_cpu_mem_usage=True, device_map="auto")
PROMPT_DICT = {
"prompt_input": (
"以下に、あるタスクを説明する指示があり、それに付随する入力が更なる文脈を提供しています。"
"リクエストを適切に完了するための回答を記述してください。\n\n"
"### 指示:\n{instruction}\n\n### 入力:\n{input}\n\n### 応答:"
),
"prompt_no_input": (
"以下に、あるタスクを説明する指示があります。"
"リクエストを適切に完了するための回答を記述してください。\n\n"
"### 指示:\n{instruction}\n\n### 応答:"
),
}
def create_prompt(instruction, input=None):
"""
Generates a prompt based on the given instruction and an optional input.
If input is provided, it uses the 'prompt_input' template from PROMPT_DICT.
If no input is provided, it uses the 'prompt_no_input' template.
Args:
instruction (str): The instruction describing the task.
input (str, optional): Additional input providing context for the task. Default is None.
Returns:
str: The generated prompt.
"""
if input:
# Use the 'prompt_input' template when additional input is provided
return PROMPT_DICT["prompt_input"].format(instruction=instruction, input=input)
else:
# Use the 'prompt_no_input' template when no additional input is provided
return PROMPT_DICT["prompt_no_input"].format(instruction=instruction)
# Example usage
instruction_example = "以下のトピックに関する詳細な情報を提供してください。"
input_example = "東京工業大学の主なキャンパスについて教えてください"
prompt = create_prompt(instruction_example, input_example)
input_ids = tokenizer.encode(
prompt,
add_special_tokens=False,
return_tensors="pt"
)
tokens = model.generate(
input_ids.to(device=model.device),
max_new_tokens=128,
temperature=0.99,
top_p=0.95,
do_sample=True,
)
out = tokenizer.decode(tokens[0], skip_special_tokens=True)
print(out)
```
### Use the base model
```python
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
model_name = "tokyotech-llm/Swallow-7b-hf"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.bfloat16, device_map="auto")
prompt = "東京工業大学の主なキャンパスは、"
input_ids = tokenizer.encode(
prompt,
add_special_tokens=False,
return_tensors="pt"
)
tokens = model.generate(
input_ids.to(device=model.device),
max_new_tokens=128,
temperature=0.99,
top_p=0.95,
do_sample=True,
)
out = tokenizer.decode(tokens[0], skip_special_tokens=True)
print(out)
```
## Training Datasets
### Continual Pre-Training
The following datasets were used for continual pre-training.
- [Japanese Wikipedia](https://dumps.wikimedia.org/other/cirrussearch)
- [RefinedWeb](https://huggingface.co/datasets/tiiuae/falcon-refinedweb)
- Swallow Corpus
- [The Pile](https://huggingface.co/datasets/EleutherAI/pile)
### Instruction Tuning
The following datasets were used for the instruction tuning.
- [Anthropic HH-RLHF](https://huggingface.co/datasets/kunishou/hh-rlhf-49k-ja)
- [Databricks Dolly 15-k](https://huggingface.co/datasets/kunishou/databricks-dolly-15k-ja)
- [OpenAssistant Conversations Dataset](https://huggingface.co/datasets/kunishou/oasst1-89k-ja)
## Risks and Limitations
The models released here are still in the early stages of our research and development and have not been tuned to ensure outputs align with human intent and safety considerations.
## Acknowledgements
We thank Meta Research for releasing Llama 2 under an open license for others to build on.
Our project is supported by the [ABCI Large-scale Language Model Building Support Program](https://abci.ai/en/link/llm_support_program.html) of the National Institute of Advanced Industrial Science and Technology.
## License
Llama 2 is licensed under the LLAMA 2 Community License, Copyright © Meta Platforms, Inc. All Rights Reserved.
## Authors
Here are the team members:
- From [Okazaki Laboratory](https://www.nlp.c.titech.ac.jp/index.en.html), the following members:
- [Naoaki Okazaki](https://www.chokkan.org/index.ja.html)
- [Sakae Mizuki](https://s-mizuki-nlp.github.io/)
- [Hiroki Iida](https://meshidenn.github.io/)
- [Mengsay Loem](https://loem-ms.github.io/)
- [Shota Hirai](https://huggingface.co/Kotemo428)
- [Kakeru Hattori](https://aya-se.vercel.app/)
- [Masanari Ohi](https://twitter.com/stjohn2007)
- From [YOKOTA Laboratory](https://www.rio.gsic.titech.ac.jp/en/index.html), the following members:
- [Rio Yokota](https://twitter.com/rioyokota)
- [Kazuki Fujii](https://twitter.com/okoge_kaz)
- [Taishi Nakamura](https://twitter.com/Setuna7777_2)
|
ep150de/linglenet
|
ep150de
| 2023-12-19T23:37:48Z | 1,903 | 4 |
diffusers
|
[
"diffusers",
"text-to-image",
"stable-diffusion",
"lora",
"template:sd-lora",
"base_model:stabilityai/stable-diffusion-xl-base-1.0",
"base_model:adapter:stabilityai/stable-diffusion-xl-base-1.0",
"license:apache-2.0",
"region:us"
] |
text-to-image
| 2023-12-06T01:35:17Z |
---
tags:
- text-to-image
- stable-diffusion
- lora
- diffusers
- template:sd-lora
widget:
- text: lingle wearing a fedora
output:
url: images/image.png
base_model: stabilityai/stable-diffusion-xl-base-1.0
instance_prompt: lingle
license: apache-2.0
---
# linglenet
<Gallery />
## Trigger words
You should use `lingle` to trigger the image generation.
## Download model
Weights for this model are available in Safetensors format.
[Download](/ep150de/linglenet/tree/main) them in the Files & versions tab.
|
badokorach/afriqa_afroxlmr_squad_v2-191223
|
badokorach
| 2023-12-19T23:15:00Z | 4 | 0 |
transformers
|
[
"transformers",
"tf",
"xlm-roberta",
"question-answering",
"generated_from_keras_callback",
"base_model:masakhane/afriqa_afroxlmr_squad_v2",
"base_model:finetune:masakhane/afriqa_afroxlmr_squad_v2",
"license:mit",
"endpoints_compatible",
"region:us"
] |
question-answering
| 2023-12-19T19:59:29Z |
---
license: mit
base_model: masakhane/afriqa_afroxlmr_squad_v2
tags:
- generated_from_keras_callback
model-index:
- name: badokorach/afriqa_afroxlmr_squad_v2-191223
results: []
---
<!-- This model card has been generated automatically according to the information Keras had access to. You should
probably proofread and complete it, then remove this comment. -->
# badokorach/afriqa_afroxlmr_squad_v2-191223
This model is a fine-tuned version of [masakhane/afriqa_afroxlmr_squad_v2](https://huggingface.co/masakhane/afriqa_afroxlmr_squad_v2) on an unknown dataset.
It achieves the following results on the evaluation set:
- Train Loss: 1.2750
- Validation Loss: 0.0
- Epoch: 19
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- optimizer: {'name': 'AdamWeightDecay', 'learning_rate': {'module': 'keras.optimizers.schedules', 'class_name': 'PolynomialDecay', 'config': {'initial_learning_rate': 1e-05, 'decay_steps': 9840, 'end_learning_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}, 'registered_name': None}, 'decay': 0.0, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False, 'weight_decay_rate': 0.02}
- training_precision: mixed_float16
### Training results
| Train Loss | Validation Loss | Epoch |
|:----------:|:---------------:|:-----:|
| 3.6857 | 0.0 | 0 |
| 3.5750 | 0.0 | 1 |
| 3.4376 | 0.0 | 2 |
| 3.2725 | 0.0 | 3 |
| 3.0996 | 0.0 | 4 |
| 2.8985 | 0.0 | 5 |
| 2.6869 | 0.0 | 6 |
| 2.4815 | 0.0 | 7 |
| 2.3027 | 0.0 | 8 |
| 2.1286 | 0.0 | 9 |
| 1.9690 | 0.0 | 10 |
| 1.8468 | 0.0 | 11 |
| 1.7192 | 0.0 | 12 |
| 1.6282 | 0.0 | 13 |
| 1.5134 | 0.0 | 14 |
| 1.4472 | 0.0 | 15 |
| 1.3944 | 0.0 | 16 |
| 1.3467 | 0.0 | 17 |
| 1.2940 | 0.0 | 18 |
| 1.2750 | 0.0 | 19 |
### Framework versions
- Transformers 4.35.2
- TensorFlow 2.15.0
- Datasets 2.15.0
- Tokenizers 0.15.0
|
TheBloke/Swallow-13B-Instruct-GPTQ
|
TheBloke
| 2023-12-19T23:03:31Z | 9 | 2 |
transformers
|
[
"transformers",
"safetensors",
"llama",
"text-generation",
"en",
"ja",
"base_model:tokyotech-llm/Swallow-13b-instruct-hf",
"base_model:quantized:tokyotech-llm/Swallow-13b-instruct-hf",
"license:llama2",
"autotrain_compatible",
"text-generation-inference",
"4-bit",
"gptq",
"region:us"
] |
text-generation
| 2023-12-19T22:17:43Z |
---
base_model: tokyotech-llm/Swallow-13b-instruct-hf
inference: false
language:
- en
- ja
library_name: transformers
license: llama2
model_creator: tokyotech-llm
model_name: Swallow 13B Instruct
model_type: llama
pipeline_tag: text-generation
prompt_template: "\u4EE5\u4E0B\u306B\u3001\u3042\u308B\u30BF\u30B9\u30AF\u3092\u8AAC\
\u660E\u3059\u308B\u6307\u793A\u304C\u3042\u308A\u307E\u3059\u3002\u30EA\u30AF\u30A8\
\u30B9\u30C8\u3092\u9069\u5207\u306B\u5B8C\u4E86\u3059\u308B\u305F\u3081\u306E\u56DE\
\u7B54\u3092\u8A18\u8FF0\u3057\u3066\u304F\u3060\u3055\u3044\u3002\\n\\n### \u6307\
\u793A:\\n{prompt}\\n\\n### \u5FDC\u7B54:\n"
quantized_by: TheBloke
---
<!-- markdownlint-disable MD041 -->
<!-- header start -->
<!-- 200823 -->
<div style="width: auto; margin-left: auto; margin-right: auto">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<div style="text-align:center; margin-top: 0em; margin-bottom: 0em"><p style="margin-top: 0.25em; margin-bottom: 0em;">TheBloke's LLM work is generously supported by a grant from <a href="https://a16z.com">andreessen horowitz (a16z)</a></p></div>
<hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
<!-- header end -->
# Swallow 13B Instruct - GPTQ
- Model creator: [tokyotech-llm](https://huggingface.co/tokyotech-llm)
- Original model: [Swallow 13B Instruct](https://huggingface.co/tokyotech-llm/Swallow-13b-instruct-hf)
<!-- description start -->
# Description
This repo contains GPTQ model files for [tokyotech-llm's Swallow 13B Instruct](https://huggingface.co/tokyotech-llm/Swallow-13b-instruct-hf).
Multiple GPTQ parameter permutations are provided; see Provided Files below for details of the options provided, their parameters, and the software used to create them.
These files were quantised using hardware kindly provided by [Massed Compute](https://massedcompute.com/).
<!-- description end -->
<!-- repositories-available start -->
## Repositories available
* [AWQ model(s) for GPU inference.](https://huggingface.co/TheBloke/Swallow-13B-Instruct-AWQ)
* [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/Swallow-13B-Instruct-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/Swallow-13B-Instruct-GGUF)
* [tokyotech-llm's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/tokyotech-llm/Swallow-13b-instruct-hf)
<!-- repositories-available end -->
<!-- prompt-template start -->
## Prompt template: Swallow-Instruct
```
以下に、あるタスクを説明する指示があります。リクエストを適切に完了するための回答を記述してください。\n\n### 指示:\n{prompt}\n\n### 応答:
```
<!-- prompt-template end -->
<!-- README_GPTQ.md-compatible clients start -->
## Known compatible clients / servers
GPTQ models are currently supported on Linux (NVidia/AMD) and Windows (NVidia only). macOS users: please use GGUF models.
These GPTQ models are known to work in the following inference servers/webuis.
- [text-generation-webui](https://github.com/oobabooga/text-generation-webui)
- [KoboldAI United](https://github.com/henk717/koboldai)
- [LoLLMS Web UI](https://github.com/ParisNeo/lollms-webui)
- [Hugging Face Text Generation Inference (TGI)](https://github.com/huggingface/text-generation-inference)
This may not be a complete list; if you know of others, please let me know!
<!-- README_GPTQ.md-compatible clients end -->
<!-- README_GPTQ.md-provided-files start -->
## Provided files, and GPTQ parameters
Multiple quantisation parameters are provided, to allow you to choose the best one for your hardware and requirements.
Each separate quant is in a different branch. See below for instructions on fetching from different branches.
Most GPTQ files are made with AutoGPTQ. Mistral models are currently made with Transformers.
<details>
<summary>Explanation of GPTQ parameters</summary>
- Bits: The bit size of the quantised model.
- GS: GPTQ group size. Higher numbers use less VRAM, but have lower quantisation accuracy. "None" is the lowest possible value.
- Act Order: True or False. Also known as `desc_act`. True results in better quantisation accuracy. Some GPTQ clients have had issues with models that use Act Order plus Group Size, but this is generally resolved now.
- Damp %: A GPTQ parameter that affects how samples are processed for quantisation. 0.01 is default, but 0.1 results in slightly better accuracy.
- GPTQ dataset: The calibration dataset used during quantisation. Using a dataset more appropriate to the model's training can improve quantisation accuracy. Note that the GPTQ calibration dataset is not the same as the dataset used to train the model - please refer to the original model repo for details of the training dataset(s).
- Sequence Length: The length of the dataset sequences used for quantisation. Ideally this is the same as the model sequence length. For some very long sequence models (16+K), a lower sequence length may have to be used. Note that a lower sequence length does not limit the sequence length of the quantised model. It only impacts the quantisation accuracy on longer inference sequences.
- ExLlama Compatibility: Whether this file can be loaded with ExLlama, which currently only supports Llama and Mistral models in 4-bit.
</details>
| Branch | Bits | GS | Act Order | Damp % | GPTQ Dataset | Seq Len | Size | ExLlama | Desc |
| ------ | ---- | -- | --------- | ------ | ------------ | ------- | ---- | ------- | ---- |
| [main](https://huggingface.co/TheBloke/Swallow-13B-Instruct-GPTQ/tree/main) | 4 | 128 | Yes | 0.1 | [Alpaca Japanese](https://huggingface.co/datasets/fujiki/japanese_alpaca_data/viewer/) | 4096 | 7.49 GB | Yes | 4-bit, with Act Order and group size 128g. Uses even less VRAM than 64g, but with slightly lower accuracy. |
| [gptq-4bit-32g-actorder_True](https://huggingface.co/TheBloke/Swallow-13B-Instruct-GPTQ/tree/gptq-4bit-32g-actorder_True) | 4 | 32 | Yes | 0.1 | [Alpaca Japanese](https://huggingface.co/datasets/fujiki/japanese_alpaca_data/viewer/) | 4096 | 8.23 GB | Yes | 4-bit, with Act Order and group size 32g. Gives highest possible inference quality, with maximum VRAM usage. |
| [gptq-8bit--1g-actorder_True](https://huggingface.co/TheBloke/Swallow-13B-Instruct-GPTQ/tree/gptq-8bit--1g-actorder_True) | 8 | None | Yes | 0.1 | [Alpaca Japanese](https://huggingface.co/datasets/fujiki/japanese_alpaca_data/viewer/) | 4096 | 13.59 GB | No | 8-bit, with Act Order. No group size, to lower VRAM requirements. |
| [gptq-8bit-128g-actorder_True](https://huggingface.co/TheBloke/Swallow-13B-Instruct-GPTQ/tree/gptq-8bit-128g-actorder_True) | 8 | 128 | Yes | 0.1 | [Alpaca Japanese](https://huggingface.co/datasets/fujiki/japanese_alpaca_data/viewer/) | 4096 | 13.88 GB | No | 8-bit, with group size 128g for higher inference quality and with Act Order for even higher accuracy. |
| [gptq-8bit-32g-actorder_True](https://huggingface.co/TheBloke/Swallow-13B-Instruct-GPTQ/tree/gptq-8bit-32g-actorder_True) | 8 | 32 | Yes | 0.1 | [Alpaca Japanese](https://huggingface.co/datasets/fujiki/japanese_alpaca_data/viewer/) | 4096 | 14.77 GB | No | 8-bit, with group size 32g and Act Order for maximum inference quality. |
| [gptq-4bit-64g-actorder_True](https://huggingface.co/TheBloke/Swallow-13B-Instruct-GPTQ/tree/gptq-4bit-64g-actorder_True) | 4 | 64 | Yes | 0.1 | [Alpaca Japanese](https://huggingface.co/datasets/fujiki/japanese_alpaca_data/viewer/) | 4096 | 7.74 GB | Yes | 4-bit, with Act Order and group size 64g. Uses less VRAM than 32g, but with slightly lower accuracy. |
<!-- README_GPTQ.md-provided-files end -->
<!-- README_GPTQ.md-download-from-branches start -->
## How to download, including from branches
### In text-generation-webui
To download from the `main` branch, enter `TheBloke/Swallow-13B-Instruct-GPTQ` in the "Download model" box.
To download from another branch, add `:branchname` to the end of the download name, eg `TheBloke/Swallow-13B-Instruct-GPTQ:gptq-4bit-32g-actorder_True`
### From the command line
I recommend using the `huggingface-hub` Python library:
```shell
pip3 install huggingface-hub
```
To download the `main` branch to a folder called `Swallow-13B-Instruct-GPTQ`:
```shell
mkdir Swallow-13B-Instruct-GPTQ
huggingface-cli download TheBloke/Swallow-13B-Instruct-GPTQ --local-dir Swallow-13B-Instruct-GPTQ --local-dir-use-symlinks False
```
To download from a different branch, add the `--revision` parameter:
```shell
mkdir Swallow-13B-Instruct-GPTQ
huggingface-cli download TheBloke/Swallow-13B-Instruct-GPTQ --revision gptq-4bit-32g-actorder_True --local-dir Swallow-13B-Instruct-GPTQ --local-dir-use-symlinks False
```
<details>
<summary>More advanced huggingface-cli download usage</summary>
If you remove the `--local-dir-use-symlinks False` parameter, the files will instead be stored in the central Hugging Face cache directory (default location on Linux is: `~/.cache/huggingface`), and symlinks will be added to the specified `--local-dir`, pointing to their real location in the cache. This allows for interrupted downloads to be resumed, and allows you to quickly clone the repo to multiple places on disk without triggering a download again. The downside, and the reason why I don't list that as the default option, is that the files are then hidden away in a cache folder and it's harder to know where your disk space is being used, and to clear it up if/when you want to remove a download model.
The cache location can be changed with the `HF_HOME` environment variable, and/or the `--cache-dir` parameter to `huggingface-cli`.
For more documentation on downloading with `huggingface-cli`, please see: [HF -> Hub Python Library -> Download files -> Download from the CLI](https://huggingface.co/docs/huggingface_hub/guides/download#download-from-the-cli).
To accelerate downloads on fast connections (1Gbit/s or higher), install `hf_transfer`:
```shell
pip3 install hf_transfer
```
And set environment variable `HF_HUB_ENABLE_HF_TRANSFER` to `1`:
```shell
mkdir Swallow-13B-Instruct-GPTQ
HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download TheBloke/Swallow-13B-Instruct-GPTQ --local-dir Swallow-13B-Instruct-GPTQ --local-dir-use-symlinks False
```
Windows Command Line users: You can set the environment variable by running `set HF_HUB_ENABLE_HF_TRANSFER=1` before the download command.
</details>
### With `git` (**not** recommended)
To clone a specific branch with `git`, use a command like this:
```shell
git clone --single-branch --branch gptq-4bit-32g-actorder_True https://huggingface.co/TheBloke/Swallow-13B-Instruct-GPTQ
```
Note that using Git with HF repos is strongly discouraged. It will be much slower than using `huggingface-hub`, and will use twice as much disk space as it has to store the model files twice (it stores every byte both in the intended target folder, and again in the `.git` folder as a blob.)
<!-- README_GPTQ.md-download-from-branches end -->
<!-- README_GPTQ.md-text-generation-webui start -->
## How to easily download and use this model in [text-generation-webui](https://github.com/oobabooga/text-generation-webui)
Please make sure you're using the latest version of [text-generation-webui](https://github.com/oobabooga/text-generation-webui).
It is strongly recommended to use the text-generation-webui one-click-installers unless you're sure you know how to make a manual install.
1. Click the **Model tab**.
2. Under **Download custom model or LoRA**, enter `TheBloke/Swallow-13B-Instruct-GPTQ`.
- To download from a specific branch, enter for example `TheBloke/Swallow-13B-Instruct-GPTQ:gptq-4bit-32g-actorder_True`
- see Provided Files above for the list of branches for each option.
3. Click **Download**.
4. The model will start downloading. Once it's finished it will say "Done".
5. In the top left, click the refresh icon next to **Model**.
6. In the **Model** dropdown, choose the model you just downloaded: `Swallow-13B-Instruct-GPTQ`
7. The model will automatically load, and is now ready for use!
8. If you want any custom settings, set them and then click **Save settings for this model** followed by **Reload the Model** in the top right.
- Note that you do not need to and should not set manual GPTQ parameters any more. These are set automatically from the file `quantize_config.json`.
9. Once you're ready, click the **Text Generation** tab and enter a prompt to get started!
<!-- README_GPTQ.md-text-generation-webui end -->
<!-- README_GPTQ.md-use-from-tgi start -->
## Serving this model from Text Generation Inference (TGI)
It's recommended to use TGI version 1.1.0 or later. The official Docker container is: `ghcr.io/huggingface/text-generation-inference:1.1.0`
Example Docker parameters:
```shell
--model-id TheBloke/Swallow-13B-Instruct-GPTQ --port 3000 --quantize gptq --max-input-length 3696 --max-total-tokens 4096 --max-batch-prefill-tokens 4096
```
Example Python code for interfacing with TGI (requires huggingface-hub 0.17.0 or later):
```shell
pip3 install huggingface-hub
```
```python
from huggingface_hub import InferenceClient
endpoint_url = "https://your-endpoint-url-here"
prompt = "Tell me about AI"
prompt_template=f'''以下に、あるタスクを説明する指示があります。リクエストを適切に完了するための回答を記述してください。\n\n### 指示:\n{prompt}\n\n### 応答:
'''
client = InferenceClient(endpoint_url)
response = client.text_generation(prompt,
max_new_tokens=128,
do_sample=True,
temperature=0.7,
top_p=0.95,
top_k=40,
repetition_penalty=1.1)
print(f"Model output: {response}")
```
<!-- README_GPTQ.md-use-from-tgi end -->
<!-- README_GPTQ.md-use-from-python start -->
## Python code example: inference from this GPTQ model
### Install the necessary packages
Requires: Transformers 4.33.0 or later, Optimum 1.12.0 or later, and AutoGPTQ 0.4.2 or later.
```shell
pip3 install --upgrade transformers optimum
# If using PyTorch 2.1 + CUDA 12.x:
pip3 install --upgrade auto-gptq
# or, if using PyTorch 2.1 + CUDA 11.x:
pip3 install --upgrade auto-gptq --extra-index-url https://huggingface.github.io/autogptq-index/whl/cu118/
```
If you are using PyTorch 2.0, you will need to install AutoGPTQ from source. Likewise if you have problems with the pre-built wheels, you should try building from source:
```shell
pip3 uninstall -y auto-gptq
git clone https://github.com/PanQiWei/AutoGPTQ
cd AutoGPTQ
git checkout v0.5.1
pip3 install .
```
### Example Python code
```python
from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
model_name_or_path = "TheBloke/Swallow-13B-Instruct-GPTQ"
# To use a different branch, change revision
# For example: revision="gptq-4bit-32g-actorder_True"
model = AutoModelForCausalLM.from_pretrained(model_name_or_path,
device_map="auto",
trust_remote_code=False,
revision="main")
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
prompt = "Write a story about llamas"
system_message = "You are a story writing assistant"
prompt_template=f'''以下に、あるタスクを説明する指示があります。リクエストを適切に完了するための回答を記述してください。\n\n### 指示:\n{prompt}\n\n### 応答:
'''
print("\n\n*** Generate:")
input_ids = tokenizer(prompt_template, return_tensors='pt').input_ids.cuda()
output = model.generate(inputs=input_ids, temperature=0.7, do_sample=True, top_p=0.95, top_k=40, max_new_tokens=512)
print(tokenizer.decode(output[0]))
# Inference can also be done using transformers' pipeline
print("*** Pipeline:")
pipe = pipeline(
"text-generation",
model=model,
tokenizer=tokenizer,
max_new_tokens=512,
do_sample=True,
temperature=0.7,
top_p=0.95,
top_k=40,
repetition_penalty=1.1
)
print(pipe(prompt_template)[0]['generated_text'])
```
<!-- README_GPTQ.md-use-from-python end -->
<!-- README_GPTQ.md-compatibility start -->
## Compatibility
The files provided are tested to work with Transformers. For non-Mistral models, AutoGPTQ can also be used directly.
[ExLlama](https://github.com/turboderp/exllama) is compatible with Llama architecture models (including Mistral, Yi, DeepSeek, SOLAR, etc) in 4-bit. Please see the Provided Files table above for per-file compatibility.
For a list of clients/servers, please see "Known compatible clients / servers", above.
<!-- README_GPTQ.md-compatibility end -->
<!-- footer start -->
<!-- 200823 -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute
Thanks to the [chirper.ai](https://chirper.ai) team!
Thanks to Clay from [gpus.llm-utils.org](llm-utils)!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Aemon Algiz.
**Patreon special mentions**: Michael Levine, 阿明, Trailburnt, Nikolai Manek, John Detwiler, Randy H, Will Dee, Sebastain Graf, NimbleBox.ai, Eugene Pentland, Emad Mostaque, Ai Maven, Jim Angel, Jeff Scroggin, Michael Davis, Manuel Alberto Morcote, Stephen Murray, Robert, Justin Joy, Luke @flexchar, Brandon Frisco, Elijah Stavena, S_X, Dan Guido, Undi ., Komninos Chatzipapas, Shadi, theTransient, Lone Striker, Raven Klaugh, jjj, Cap'n Zoog, Michel-Marie MAUDET (LINAGORA), Matthew Berman, David, Fen Risland, Omer Bin Jawed, Luke Pendergrass, Kalila, OG, Erik Bjäreholt, Rooh Singh, Joseph William Delisle, Dan Lewis, TL, John Villwock, AzureBlack, Brad, Pedro Madruga, Caitlyn Gatomon, K, jinyuan sun, Mano Prime, Alex, Jeffrey Morgan, Alicia Loh, Illia Dulskyi, Chadd, transmissions 11, fincy, Rainer Wilmers, ReadyPlayerEmma, knownsqashed, Mandus, biorpg, Deo Leter, Brandon Phillips, SuperWojo, Sean Connelly, Iucharbius, Jack West, Harry Royden McLaughlin, Nicholas, terasurfer, Vitor Caleffi, Duane Dunston, Johann-Peter Hartmann, David Ziegler, Olakabola, Ken Nordquist, Trenton Dambrowitz, Tom X Nguyen, Vadim, Ajan Kanaga, Leonard Tan, Clay Pascal, Alexandros Triantafyllidis, JM33133, Xule, vamX, ya boyyy, subjectnull, Talal Aujan, Alps Aficionado, wassieverse, Ari Malik, James Bentley, Woland, Spencer Kim, Michael Dempsey, Fred von Graf, Elle, zynix, William Richards, Stanislav Ovsiannikov, Edmond Seymore, Jonathan Leane, Martin Kemka, usrbinkat, Enrico Ros
Thank you to all my generous patrons and donaters!
And thank you again to a16z for their generous grant.
<!-- footer end -->
# Original model card: tokyotech-llm's Swallow 13B Instruct
# Swallow
Our Swallow model has undergone continuous pre-training from the Llama 2 family, primarily with the addition of Japanese language data. The tuned versions use supervised fine-tuning (SFT).
Links to other models can be found in the index.
## Swallow Model Index
|Model|Swallow-hf|Swallow-instruct-hf|
|---|---|---|
|7B| [Link](https://huggingface.co/tokyotech-llm/Swallow-7b-hf) | [Link](https://huggingface.co/tokyotech-llm/Swallow-7b-instruct-hf)|
|13B| [Link](https://huggingface.co/tokyotech-llm/Swallow-13b-hf) | [Link](https://huggingface.co/tokyotech-llm/Swallow-13b-instruct-hf)|
|70B| [Link](https://huggingface.co/tokyotech-llm/Swallow-70b-hf) | [Link](https://huggingface.co/tokyotech-llm/Swallow-70b-instruct-hf)|
This repository provides large language models developed by [TokyoTech-LLM](https://tokyotech-llm.github.io/).
Read our [blog post](https://zenn.dev/tokyotech_lm/articles/d6cb3a8fdfc907) or our paper (preprint coming soon) for more details!
## Model Details
* **Model type**: Please refer to LLaMA-2 technical report for details on the model architecture.
* **Language(s)**: Japanese English
* **Library**: [Megatron-LM](https://github.com/rioyokotalab/Megatron-Llama2)
* **Tokenizer**: This model employs a tokenizer that features a broadened vocabulary based on Japanese data. This allows for a more efficient representation of text using fewer tokens, leading to a notably faster inference process.
* **Contact**: swallow[at]nlp.c.titech.ac.jp
## Base Model Performance
### Japanese version
|Model|Size|JCommonsenseQA|JEMHopQA|NIILC|JSQuAD|XL-Sum|MGSM|WMT20-en-ja|WMT20-ja-en|
|---|---|---|---|---|---|---|---|---|---|
| | |4-shot|4-shot|4-shot|4-shot|1-shot|4-shot|4-shot|4-shot|
|Llama 2|7B|0.3852|0.4240|0.3410|0.7917|0.1905|0.0760|0.1783|0.1738|
|Swallow|7B|0.4808|0.5078|0.5968|0.8573|0.1830|0.1240|0.2510|0.1511|
|Llama 2|13B|0.6997|0.4415|0.4170|0.8533|0.2139|0.1320|0.2146|0.1982|
|Swallow|13B|0.7837|0.5063|0.6398|0.9005|0.2168|0.2040|0.2720|0.1771|
|Llama 2|70B|0.8686|0.4656|0.5256|0.9080|**0.2361**|0.3560|0.2643|**0.2398**|
|Swallow|70B|**0.9348**|**0.6290**|**0.6960**|**0.9176**|0.2266|**0.4840**|**0.3043**|0.2298|
## Usage
First install additional dependencies in [requirements.txt](./requirements.txt):
```sh
pip install -r requirements.txt
```
### Use the instruct model
```python
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
model_name = "tokyotech-llm/Swallow-7b-instruct-hf"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.bfloat16, low_cpu_mem_usage=True, device_map="auto")
PROMPT_DICT = {
"prompt_input": (
"以下に、あるタスクを説明する指示があり、それに付随する入力が更なる文脈を提供しています。"
"リクエストを適切に完了するための回答を記述してください。\n\n"
"### 指示:\n{instruction}\n\n### 入力:\n{input}\n\n### 応答:"
),
"prompt_no_input": (
"以下に、あるタスクを説明する指示があります。"
"リクエストを適切に完了するための回答を記述してください。\n\n"
"### 指示:\n{instruction}\n\n### 応答:"
),
}
def create_prompt(instruction, input=None):
"""
Generates a prompt based on the given instruction and an optional input.
If input is provided, it uses the 'prompt_input' template from PROMPT_DICT.
If no input is provided, it uses the 'prompt_no_input' template.
Args:
instruction (str): The instruction describing the task.
input (str, optional): Additional input providing context for the task. Default is None.
Returns:
str: The generated prompt.
"""
if input:
# Use the 'prompt_input' template when additional input is provided
return PROMPT_DICT["prompt_input"].format(instruction=instruction, input=input)
else:
# Use the 'prompt_no_input' template when no additional input is provided
return PROMPT_DICT["prompt_no_input"].format(instruction=instruction)
# Example usage
instruction_example = "以下のトピックに関する詳細な情報を提供してください。"
input_example = "東京工業大学の主なキャンパスについて教えてください"
prompt = create_prompt(instruction_example, input_example)
input_ids = tokenizer.encode(
prompt,
add_special_tokens=False,
return_tensors="pt"
)
tokens = model.generate(
input_ids.to(device=model.device),
max_new_tokens=128,
temperature=0.99,
top_p=0.95,
do_sample=True,
)
out = tokenizer.decode(tokens[0], skip_special_tokens=True)
print(out)
```
### Use the base model
```python
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
model_name = "tokyotech-llm/Swallow-7b-hf"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.bfloat16, device_map="auto")
prompt = "東京工業大学の主なキャンパスは、"
input_ids = tokenizer.encode(
prompt,
add_special_tokens=False,
return_tensors="pt"
)
tokens = model.generate(
input_ids.to(device=model.device),
max_new_tokens=128,
temperature=0.99,
top_p=0.95,
do_sample=True,
)
out = tokenizer.decode(tokens[0], skip_special_tokens=True)
print(out)
```
## Training Datasets
### Continual Pre-Training
The following datasets were used for continual pre-training.
- [Japanese Wikipedia](https://dumps.wikimedia.org/other/cirrussearch)
- [RefinedWeb](https://huggingface.co/datasets/tiiuae/falcon-refinedweb)
- Swallow Corpus
- [The Pile](https://huggingface.co/datasets/EleutherAI/pile)
### Instruction Tuning
The following datasets were used for the instruction tuning.
- [Anthropic HH-RLHF](https://huggingface.co/datasets/kunishou/hh-rlhf-49k-ja)
- [Databricks Dolly 15-k](https://huggingface.co/datasets/kunishou/databricks-dolly-15k-ja)
- [OpenAssistant Conversations Dataset](https://huggingface.co/datasets/kunishou/oasst1-89k-ja)
## Risks and Limitations
The models released here are still in the early stages of our research and development and have not been tuned to ensure outputs align with human intent and safety considerations.
## Acknowledgements
We thank Meta Research for releasing Llama 2 under an open license for others to build on.
Our project is supported by the [ABCI Large-scale Language Model Building Support Program](https://abci.ai/en/link/llm_support_program.html) of the National Institute of Advanced Industrial Science and Technology.
## License
Llama 2 is licensed under the LLAMA 2 Community License, Copyright © Meta Platforms, Inc. All Rights Reserved.
## Authors
Here are the team members:
- From [Okazaki Laboratory](https://www.nlp.c.titech.ac.jp/index.en.html), the following members:
- [Naoaki Okazaki](https://www.chokkan.org/index.ja.html)
- [Sakae Mizuki](https://s-mizuki-nlp.github.io/)
- [Hiroki Iida](https://meshidenn.github.io/)
- [Mengsay Loem](https://loem-ms.github.io/)
- [Shota Hirai](https://huggingface.co/Kotemo428)
- [Kakeru Hattori](https://aya-se.vercel.app/)
- [Masanari Ohi](https://twitter.com/stjohn2007)
- From [YOKOTA Laboratory](https://www.rio.gsic.titech.ac.jp/en/index.html), the following members:
- [Rio Yokota](https://twitter.com/rioyokota)
- [Kazuki Fujii](https://twitter.com/okoge_kaz)
- [Taishi Nakamura](https://twitter.com/Setuna7777_2)
|
Blackroot/SDXL-DPO-Actual-Safetensors
|
Blackroot
| 2023-12-19T22:54:14Z | 0 | 1 | null |
[
"region:us"
] | null | 2023-12-19T22:49:27Z |
Just a converted version of <https://huggingface.co/mhdang/dpo-sdxl-text2image-v1> from HF format to something you can run in comfy/auto
|
jak414/liedetect_fold3
|
jak414
| 2023-12-19T22:34:21Z | 0 | 0 |
peft
|
[
"peft",
"region:us"
] | null | 2023-12-19T22:34:20Z |
---
library_name: peft
---
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- load_in_8bit: True
- load_in_4bit: False
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: fp4
- bnb_4bit_use_double_quant: False
- bnb_4bit_compute_dtype: float32
### Framework versions
- PEFT 0.4.0
|
XinHun/Macross_F
|
XinHun
| 2023-12-19T22:21:28Z | 0 | 1 | null |
[
"license:other",
"region:us"
] | null | 2023-12-15T19:49:06Z |
---
license: other
license_name: '1'
license_link: LICENSE
---
|
TheBloke/Swallow-70B-GGUF
|
TheBloke
| 2023-12-19T22:19:44Z | 250 | 0 |
transformers
|
[
"transformers",
"gguf",
"llama",
"text-generation",
"en",
"ja",
"base_model:tokyotech-llm/Swallow-70b-hf",
"base_model:quantized:tokyotech-llm/Swallow-70b-hf",
"license:llama2",
"region:us"
] |
text-generation
| 2023-12-19T21:54:33Z |
---
base_model: tokyotech-llm/Swallow-70b-hf
inference: false
language:
- en
- ja
library_name: transformers
license: llama2
model_creator: tokyotech-llm
model_name: Swallow 70B
model_type: llama
pipeline_tag: text-generation
prompt_template: '{prompt}
'
quantized_by: TheBloke
---
<!-- markdownlint-disable MD041 -->
<!-- header start -->
<!-- 200823 -->
<div style="width: auto; margin-left: auto; margin-right: auto">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<div style="text-align:center; margin-top: 0em; margin-bottom: 0em"><p style="margin-top: 0.25em; margin-bottom: 0em;">TheBloke's LLM work is generously supported by a grant from <a href="https://a16z.com">andreessen horowitz (a16z)</a></p></div>
<hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
<!-- header end -->
# Swallow 70B - GGUF
- Model creator: [tokyotech-llm](https://huggingface.co/tokyotech-llm)
- Original model: [Swallow 70B](https://huggingface.co/tokyotech-llm/Swallow-70b-hf)
<!-- description start -->
## Description
This repo contains GGUF format model files for [tokyotech-llm's Swallow 70B](https://huggingface.co/tokyotech-llm/Swallow-70b-hf).
<!-- description end -->
<!-- README_GGUF.md-about-gguf start -->
### About GGUF
GGUF is a new format introduced by the llama.cpp team on August 21st 2023. It is a replacement for GGML, which is no longer supported by llama.cpp.
Here is an incomplete list of clients and libraries that are known to support GGUF:
* [llama.cpp](https://github.com/ggerganov/llama.cpp). The source project for GGUF. Offers a CLI and a server option.
* [text-generation-webui](https://github.com/oobabooga/text-generation-webui), the most widely used web UI, with many features and powerful extensions. Supports GPU acceleration.
* [KoboldCpp](https://github.com/LostRuins/koboldcpp), a fully featured web UI, with GPU accel across all platforms and GPU architectures. Especially good for story telling.
* [GPT4All](https://gpt4all.io/index.html), a free and open source local running GUI, supporting Windows, Linux and macOS with full GPU accel.
* [LM Studio](https://lmstudio.ai/), an easy-to-use and powerful local GUI for Windows and macOS (Silicon), with GPU acceleration. Linux available, in beta as of 27/11/2023.
* [LoLLMS Web UI](https://github.com/ParisNeo/lollms-webui), a great web UI with many interesting and unique features, including a full model library for easy model selection.
* [Faraday.dev](https://faraday.dev/), an attractive and easy to use character-based chat GUI for Windows and macOS (both Silicon and Intel), with GPU acceleration.
* [llama-cpp-python](https://github.com/abetlen/llama-cpp-python), a Python library with GPU accel, LangChain support, and OpenAI-compatible API server.
* [candle](https://github.com/huggingface/candle), a Rust ML framework with a focus on performance, including GPU support, and ease of use.
* [ctransformers](https://github.com/marella/ctransformers), a Python library with GPU accel, LangChain support, and OpenAI-compatible AI server. Note, as of time of writing (November 27th 2023), ctransformers has not been updated in a long time and does not support many recent models.
<!-- README_GGUF.md-about-gguf end -->
<!-- repositories-available start -->
## Repositories available
* [AWQ model(s) for GPU inference.](https://huggingface.co/TheBloke/Swallow-70B-AWQ)
* [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/Swallow-70B-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/Swallow-70B-GGUF)
* [tokyotech-llm's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/tokyotech-llm/Swallow-70b-hf)
<!-- repositories-available end -->
<!-- prompt-template start -->
## Prompt template: None
```
{prompt}
```
<!-- prompt-template end -->
<!-- compatibility_gguf start -->
## Compatibility
These quantised GGUFv2 files are compatible with llama.cpp from August 27th onwards, as of commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221)
They are also compatible with many third party UIs and libraries - please see the list at the top of this README.
## Explanation of quantisation methods
<details>
<summary>Click to see details</summary>
The new methods available are:
* GGML_TYPE_Q2_K - "type-1" 2-bit quantization in super-blocks containing 16 blocks, each block having 16 weight. Block scales and mins are quantized with 4 bits. This ends up effectively using 2.5625 bits per weight (bpw)
* GGML_TYPE_Q3_K - "type-0" 3-bit quantization in super-blocks containing 16 blocks, each block having 16 weights. Scales are quantized with 6 bits. This end up using 3.4375 bpw.
* GGML_TYPE_Q4_K - "type-1" 4-bit quantization in super-blocks containing 8 blocks, each block having 32 weights. Scales and mins are quantized with 6 bits. This ends up using 4.5 bpw.
* GGML_TYPE_Q5_K - "type-1" 5-bit quantization. Same super-block structure as GGML_TYPE_Q4_K resulting in 5.5 bpw
* GGML_TYPE_Q6_K - "type-0" 6-bit quantization. Super-blocks with 16 blocks, each block having 16 weights. Scales are quantized with 8 bits. This ends up using 6.5625 bpw
Refer to the Provided Files table below to see what files use which methods, and how.
</details>
<!-- compatibility_gguf end -->
<!-- README_GGUF.md-provided-files start -->
## Provided files
| Name | Quant method | Bits | Size | Max RAM required | Use case |
| ---- | ---- | ---- | ---- | ---- | ----- |
| [swallow-70b.Q2_K.gguf](https://huggingface.co/TheBloke/Swallow-70B-GGUF/blob/main/swallow-70b.Q2_K.gguf) | Q2_K | 2 | 29.38 GB| 31.88 GB | smallest, significant quality loss - not recommended for most purposes |
| [swallow-70b.Q3_K_S.gguf](https://huggingface.co/TheBloke/Swallow-70B-GGUF/blob/main/swallow-70b.Q3_K_S.gguf) | Q3_K_S | 3 | 30.03 GB| 32.53 GB | very small, high quality loss |
| [swallow-70b.Q3_K_M.gguf](https://huggingface.co/TheBloke/Swallow-70B-GGUF/blob/main/swallow-70b.Q3_K_M.gguf) | Q3_K_M | 3 | 33.30 GB| 35.80 GB | very small, high quality loss |
| [swallow-70b.Q3_K_L.gguf](https://huggingface.co/TheBloke/Swallow-70B-GGUF/blob/main/swallow-70b.Q3_K_L.gguf) | Q3_K_L | 3 | 36.26 GB| 38.76 GB | small, substantial quality loss |
| [swallow-70b.Q4_0.gguf](https://huggingface.co/TheBloke/Swallow-70B-GGUF/blob/main/swallow-70b.Q4_0.gguf) | Q4_0 | 4 | 39.00 GB| 41.50 GB | legacy; small, very high quality loss - prefer using Q3_K_M |
| [swallow-70b.Q4_K_S.gguf](https://huggingface.co/TheBloke/Swallow-70B-GGUF/blob/main/swallow-70b.Q4_K_S.gguf) | Q4_K_S | 4 | 39.20 GB| 41.70 GB | small, greater quality loss |
| [swallow-70b.Q4_K_M.gguf](https://huggingface.co/TheBloke/Swallow-70B-GGUF/blob/main/swallow-70b.Q4_K_M.gguf) | Q4_K_M | 4 | 41.55 GB| 44.05 GB | medium, balanced quality - recommended |
| [swallow-70b.Q5_0.gguf](https://huggingface.co/TheBloke/Swallow-70B-GGUF/blob/main/swallow-70b.Q5_0.gguf) | Q5_0 | 5 | 47.60 GB| 50.10 GB | legacy; medium, balanced quality - prefer using Q4_K_M |
| [swallow-70b.Q5_K_S.gguf](https://huggingface.co/TheBloke/Swallow-70B-GGUF/blob/main/swallow-70b.Q5_K_S.gguf) | Q5_K_S | 5 | 47.60 GB| 50.10 GB | large, low quality loss - recommended |
| [swallow-70b.Q5_K_M.gguf](https://huggingface.co/TheBloke/Swallow-70B-GGUF/blob/main/swallow-70b.Q5_K_M.gguf) | Q5_K_M | 5 | 48.89 GB| 51.39 GB | large, very low quality loss - recommended |
| swallow-70b.Q6_K.gguf | Q6_K | 6 | 56.74 GB| 59.24 GB | very large, extremely low quality loss |
| swallow-70b.Q8_0.gguf | Q8_0 | 8 | 73.49 GB| 75.99 GB | very large, extremely low quality loss - not recommended |
**Note**: the above RAM figures assume no GPU offloading. If layers are offloaded to the GPU, this will reduce RAM usage and use VRAM instead.
### Q6_K and Q8_0 files are split and require joining
**Note:** HF does not support uploading files larger than 50GB. Therefore I have uploaded the Q6_K and Q8_0 files as split files.
<details>
<summary>Click for instructions regarding Q6_K and Q8_0 files</summary>
### q6_K
Please download:
* `swallow-70b.Q6_K.gguf-split-a`
* `swallow-70b.Q6_K.gguf-split-b`
### q8_0
Please download:
* `swallow-70b.Q8_0.gguf-split-a`
* `swallow-70b.Q8_0.gguf-split-b`
To join the files, do the following:
Linux and macOS:
```
cat swallow-70b.Q6_K.gguf-split-* > swallow-70b.Q6_K.gguf && rm swallow-70b.Q6_K.gguf-split-*
cat swallow-70b.Q8_0.gguf-split-* > swallow-70b.Q8_0.gguf && rm swallow-70b.Q8_0.gguf-split-*
```
Windows command line:
```
COPY /B swallow-70b.Q6_K.gguf-split-a + swallow-70b.Q6_K.gguf-split-b swallow-70b.Q6_K.gguf
del swallow-70b.Q6_K.gguf-split-a swallow-70b.Q6_K.gguf-split-b
COPY /B swallow-70b.Q8_0.gguf-split-a + swallow-70b.Q8_0.gguf-split-b swallow-70b.Q8_0.gguf
del swallow-70b.Q8_0.gguf-split-a swallow-70b.Q8_0.gguf-split-b
```
</details>
<!-- README_GGUF.md-provided-files end -->
<!-- README_GGUF.md-how-to-download start -->
## How to download GGUF files
**Note for manual downloaders:** You almost never want to clone the entire repo! Multiple different quantisation formats are provided, and most users only want to pick and download a single file.
The following clients/libraries will automatically download models for you, providing a list of available models to choose from:
* LM Studio
* LoLLMS Web UI
* Faraday.dev
### In `text-generation-webui`
Under Download Model, you can enter the model repo: TheBloke/Swallow-70B-GGUF and below it, a specific filename to download, such as: swallow-70b.Q4_K_M.gguf.
Then click Download.
### On the command line, including multiple files at once
I recommend using the `huggingface-hub` Python library:
```shell
pip3 install huggingface-hub
```
Then you can download any individual model file to the current directory, at high speed, with a command like this:
```shell
huggingface-cli download TheBloke/Swallow-70B-GGUF swallow-70b.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False
```
<details>
<summary>More advanced huggingface-cli download usage (click to read)</summary>
You can also download multiple files at once with a pattern:
```shell
huggingface-cli download TheBloke/Swallow-70B-GGUF --local-dir . --local-dir-use-symlinks False --include='*Q4_K*gguf'
```
For more documentation on downloading with `huggingface-cli`, please see: [HF -> Hub Python Library -> Download files -> Download from the CLI](https://huggingface.co/docs/huggingface_hub/guides/download#download-from-the-cli).
To accelerate downloads on fast connections (1Gbit/s or higher), install `hf_transfer`:
```shell
pip3 install hf_transfer
```
And set environment variable `HF_HUB_ENABLE_HF_TRANSFER` to `1`:
```shell
HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download TheBloke/Swallow-70B-GGUF swallow-70b.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False
```
Windows Command Line users: You can set the environment variable by running `set HF_HUB_ENABLE_HF_TRANSFER=1` before the download command.
</details>
<!-- README_GGUF.md-how-to-download end -->
<!-- README_GGUF.md-how-to-run start -->
## Example `llama.cpp` command
Make sure you are using `llama.cpp` from commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221) or later.
```shell
./main -ngl 35 -m swallow-70b.Q4_K_M.gguf --color -c 4096 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "{prompt}"
```
Change `-ngl 32` to the number of layers to offload to GPU. Remove it if you don't have GPU acceleration.
Change `-c 4096` to the desired sequence length. For extended sequence models - eg 8K, 16K, 32K - the necessary RoPE scaling parameters are read from the GGUF file and set by llama.cpp automatically. Note that longer sequence lengths require much more resources, so you may need to reduce this value.
If you want to have a chat-style conversation, replace the `-p <PROMPT>` argument with `-i -ins`
For other parameters and how to use them, please refer to [the llama.cpp documentation](https://github.com/ggerganov/llama.cpp/blob/master/examples/main/README.md)
## How to run in `text-generation-webui`
Further instructions can be found in the text-generation-webui documentation, here: [text-generation-webui/docs/04 ‐ Model Tab.md](https://github.com/oobabooga/text-generation-webui/blob/main/docs/04%20%E2%80%90%20Model%20Tab.md#llamacpp).
## How to run from Python code
You can use GGUF models from Python using the [llama-cpp-python](https://github.com/abetlen/llama-cpp-python) or [ctransformers](https://github.com/marella/ctransformers) libraries. Note that at the time of writing (Nov 27th 2023), ctransformers has not been updated for some time and is not compatible with some recent models. Therefore I recommend you use llama-cpp-python.
### How to load this model in Python code, using llama-cpp-python
For full documentation, please see: [llama-cpp-python docs](https://abetlen.github.io/llama-cpp-python/).
#### First install the package
Run one of the following commands, according to your system:
```shell
# Base ctransformers with no GPU acceleration
pip install llama-cpp-python
# With NVidia CUDA acceleration
CMAKE_ARGS="-DLLAMA_CUBLAS=on" pip install llama-cpp-python
# Or with OpenBLAS acceleration
CMAKE_ARGS="-DLLAMA_BLAS=ON -DLLAMA_BLAS_VENDOR=OpenBLAS" pip install llama-cpp-python
# Or with CLBLast acceleration
CMAKE_ARGS="-DLLAMA_CLBLAST=on" pip install llama-cpp-python
# Or with AMD ROCm GPU acceleration (Linux only)
CMAKE_ARGS="-DLLAMA_HIPBLAS=on" pip install llama-cpp-python
# Or with Metal GPU acceleration for macOS systems only
CMAKE_ARGS="-DLLAMA_METAL=on" pip install llama-cpp-python
# In windows, to set the variables CMAKE_ARGS in PowerShell, follow this format; eg for NVidia CUDA:
$env:CMAKE_ARGS = "-DLLAMA_OPENBLAS=on"
pip install llama-cpp-python
```
#### Simple llama-cpp-python example code
```python
from llama_cpp import Llama
# Set gpu_layers to the number of layers to offload to GPU. Set to 0 if no GPU acceleration is available on your system.
llm = Llama(
model_path="./swallow-70b.Q4_K_M.gguf", # Download the model file first
n_ctx=4096, # The max sequence length to use - note that longer sequence lengths require much more resources
n_threads=8, # The number of CPU threads to use, tailor to your system and the resulting performance
n_gpu_layers=35 # The number of layers to offload to GPU, if you have GPU acceleration available
)
# Simple inference example
output = llm(
"{prompt}", # Prompt
max_tokens=512, # Generate up to 512 tokens
stop=["</s>"], # Example stop token - not necessarily correct for this specific model! Please check before using.
echo=True # Whether to echo the prompt
)
# Chat Completion API
llm = Llama(model_path="./swallow-70b.Q4_K_M.gguf", chat_format="llama-2") # Set chat_format according to the model you are using
llm.create_chat_completion(
messages = [
{"role": "system", "content": "You are a story writing assistant."},
{
"role": "user",
"content": "Write a story about llamas."
}
]
)
```
## How to use with LangChain
Here are guides on using llama-cpp-python and ctransformers with LangChain:
* [LangChain + llama-cpp-python](https://python.langchain.com/docs/integrations/llms/llamacpp)
* [LangChain + ctransformers](https://python.langchain.com/docs/integrations/providers/ctransformers)
<!-- README_GGUF.md-how-to-run end -->
<!-- footer start -->
<!-- 200823 -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute
Thanks to the [chirper.ai](https://chirper.ai) team!
Thanks to Clay from [gpus.llm-utils.org](llm-utils)!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Aemon Algiz.
**Patreon special mentions**: Michael Levine, 阿明, Trailburnt, Nikolai Manek, John Detwiler, Randy H, Will Dee, Sebastain Graf, NimbleBox.ai, Eugene Pentland, Emad Mostaque, Ai Maven, Jim Angel, Jeff Scroggin, Michael Davis, Manuel Alberto Morcote, Stephen Murray, Robert, Justin Joy, Luke @flexchar, Brandon Frisco, Elijah Stavena, S_X, Dan Guido, Undi ., Komninos Chatzipapas, Shadi, theTransient, Lone Striker, Raven Klaugh, jjj, Cap'n Zoog, Michel-Marie MAUDET (LINAGORA), Matthew Berman, David, Fen Risland, Omer Bin Jawed, Luke Pendergrass, Kalila, OG, Erik Bjäreholt, Rooh Singh, Joseph William Delisle, Dan Lewis, TL, John Villwock, AzureBlack, Brad, Pedro Madruga, Caitlyn Gatomon, K, jinyuan sun, Mano Prime, Alex, Jeffrey Morgan, Alicia Loh, Illia Dulskyi, Chadd, transmissions 11, fincy, Rainer Wilmers, ReadyPlayerEmma, knownsqashed, Mandus, biorpg, Deo Leter, Brandon Phillips, SuperWojo, Sean Connelly, Iucharbius, Jack West, Harry Royden McLaughlin, Nicholas, terasurfer, Vitor Caleffi, Duane Dunston, Johann-Peter Hartmann, David Ziegler, Olakabola, Ken Nordquist, Trenton Dambrowitz, Tom X Nguyen, Vadim, Ajan Kanaga, Leonard Tan, Clay Pascal, Alexandros Triantafyllidis, JM33133, Xule, vamX, ya boyyy, subjectnull, Talal Aujan, Alps Aficionado, wassieverse, Ari Malik, James Bentley, Woland, Spencer Kim, Michael Dempsey, Fred von Graf, Elle, zynix, William Richards, Stanislav Ovsiannikov, Edmond Seymore, Jonathan Leane, Martin Kemka, usrbinkat, Enrico Ros
Thank you to all my generous patrons and donaters!
And thank you again to a16z for their generous grant.
<!-- footer end -->
<!-- original-model-card start -->
# Original model card: tokyotech-llm's Swallow 70B
# Swallow
Our Swallow model has undergone continuous pre-training from the Llama 2 family, primarily with the addition of Japanese language data. The tuned versions use supervised fine-tuning (SFT).
Links to other models can be found in the index.
## Swallow Model Index
|Model|Swallow-hf|Swallow-instruct-hf|
|---|---|---|
|7B| [Link](https://huggingface.co/tokyotech-llm/Swallow-7b-hf) | [Link](https://huggingface.co/tokyotech-llm/Swallow-7b-instruct-hf)|
|13B| [Link](https://huggingface.co/tokyotech-llm/Swallow-13b-hf) | [Link](https://huggingface.co/tokyotech-llm/Swallow-13b-instruct-hf)|
|70B| [Link](https://huggingface.co/tokyotech-llm/Swallow-70b-hf) | [Link](https://huggingface.co/tokyotech-llm/Swallow-70b-instruct-hf)|
This repository provides large language models developed by [TokyoTech-LLM](https://tokyotech-llm.github.io/).
Read our [blog post](https://zenn.dev/tokyotech_lm/articles/d6cb3a8fdfc907) or our paper (preprint coming soon) for more details!
## Model Details
* **Model type**: Please refer to LLaMA-2 technical report for details on the model architecture.
* **Language(s)**: Japanese English
* **Library**: [Megatron-LM](https://github.com/rioyokotalab/Megatron-Llama2)
* **Tokenizer**: This model employs a tokenizer that features a broadened vocabulary based on Japanese data. This allows for a more efficient representation of text using fewer tokens, leading to a notably faster inference process.
* **Contact**: swallow[at]nlp.c.titech.ac.jp
## Base Model Performance
### Japanese version
|Model|Size|JCommonsenseQA|JEMHopQA|NIILC|JSQuAD|XL-Sum|MGSM|WMT20-en-ja|WMT20-ja-en|
|---|---|---|---|---|---|---|---|---|---|
| | |4-shot|4-shot|4-shot|4-shot|1-shot|4-shot|4-shot|4-shot|
|Llama 2|7B|0.3852|0.4240|0.3410|0.7917|0.1905|0.0760|0.1783|0.1738|
|Swallow|7B|0.4808|0.5078|0.5968|0.8573|0.1830|0.1240|0.2510|0.1511|
|Llama 2|13B|0.6997|0.4415|0.4170|0.8533|0.2139|0.1320|0.2146|0.1982|
|Swallow|13B|0.7837|0.5063|0.6398|0.9005|0.2168|0.2040|0.2720|0.1771|
|Llama 2|70B|0.8686|0.4656|0.5256|0.9080|**0.2361**|0.3560|0.2643|**0.2398**|
|Swallow|70B|**0.9348**|**0.6290**|**0.6960**|**0.9176**|0.2266|**0.4840**|**0.3043**|0.2298|
## Usage
First install additional dependencies in [requirements.txt](./requirements.txt):
```sh
pip install -r requirements.txt
```
### Use the instruct model
```python
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
model_name = "tokyotech-llm/Swallow-7b-instruct-hf"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.bfloat16, low_cpu_mem_usage=True, device_map="auto")
PROMPT_DICT = {
"prompt_input": (
"以下に、あるタスクを説明する指示があり、それに付随する入力が更なる文脈を提供しています。"
"リクエストを適切に完了するための回答を記述してください。\n\n"
"### 指示:\n{instruction}\n\n### 入力:\n{input}\n\n### 応答:"
),
"prompt_no_input": (
"以下に、あるタスクを説明する指示があります。"
"リクエストを適切に完了するための回答を記述してください。\n\n"
"### 指示:\n{instruction}\n\n### 応答:"
),
}
def create_prompt(instruction, input=None):
"""
Generates a prompt based on the given instruction and an optional input.
If input is provided, it uses the 'prompt_input' template from PROMPT_DICT.
If no input is provided, it uses the 'prompt_no_input' template.
Args:
instruction (str): The instruction describing the task.
input (str, optional): Additional input providing context for the task. Default is None.
Returns:
str: The generated prompt.
"""
if input:
# Use the 'prompt_input' template when additional input is provided
return PROMPT_DICT["prompt_input"].format(instruction=instruction, input=input)
else:
# Use the 'prompt_no_input' template when no additional input is provided
return PROMPT_DICT["prompt_no_input"].format(instruction=instruction)
# Example usage
instruction_example = "以下のトピックに関する詳細な情報を提供してください。"
input_example = "東京工業大学の主なキャンパスについて教えてください"
prompt = create_prompt(instruction_example, input_example)
input_ids = tokenizer.encode(
prompt,
add_special_tokens=False,
return_tensors="pt"
)
tokens = model.generate(
input_ids.to(device=model.device),
max_new_tokens=128,
temperature=0.99,
top_p=0.95,
do_sample=True,
)
out = tokenizer.decode(tokens[0], skip_special_tokens=True)
print(out)
```
### Use the base model
```python
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
model_name = "tokyotech-llm/Swallow-7b-hf"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.bfloat16, device_map="auto")
prompt = "東京工業大学の主なキャンパスは、"
input_ids = tokenizer.encode(
prompt,
add_special_tokens=False,
return_tensors="pt"
)
tokens = model.generate(
input_ids.to(device=model.device),
max_new_tokens=128,
temperature=0.99,
top_p=0.95,
do_sample=True,
)
out = tokenizer.decode(tokens[0], skip_special_tokens=True)
print(out)
```
## Training Datasets
### Continual Pre-Training
The following datasets were used for continual pre-training.
- [Japanese Wikipedia](https://dumps.wikimedia.org/other/cirrussearch)
- [RefinedWeb](https://huggingface.co/datasets/tiiuae/falcon-refinedweb)
- Swallow Corpus
- [The Pile](https://huggingface.co/datasets/EleutherAI/pile)
### Instruction Tuning
The following datasets were used for the instruction tuning.
- [Anthropic HH-RLHF](https://huggingface.co/datasets/kunishou/hh-rlhf-49k-ja)
- [Databricks Dolly 15-k](https://huggingface.co/datasets/kunishou/databricks-dolly-15k-ja)
- [OpenAssistant Conversations Dataset](https://huggingface.co/datasets/kunishou/oasst1-89k-ja)
## Risks and Limitations
The models released here are still in the early stages of our research and development and have not been tuned to ensure outputs align with human intent and safety considerations.
## Acknowledgements
We thank Meta Research for releasing Llama 2 under an open license for others to build on.
Our project is supported by the [ABCI Large-scale Language Model Building Support Program](https://abci.ai/en/link/llm_support_program.html) of the National Institute of Advanced Industrial Science and Technology.
## License
Llama 2 is licensed under the LLAMA 2 Community License, Copyright © Meta Platforms, Inc. All Rights Reserved.
## Authors
Here are the team members:
- From [Okazaki Laboratory](https://www.nlp.c.titech.ac.jp/index.en.html), the following members:
- [Naoaki Okazaki](https://www.chokkan.org/index.ja.html)
- [Sakae Mizuki](https://s-mizuki-nlp.github.io/)
- [Hiroki Iida](https://meshidenn.github.io/)
- [Mengsay Loem](https://loem-ms.github.io/)
- [Shota Hirai](https://huggingface.co/Kotemo428)
- [Kakeru Hattori](https://aya-se.vercel.app/)
- [Masanari Ohi](https://twitter.com/stjohn2007)
- From [YOKOTA Laboratory](https://www.rio.gsic.titech.ac.jp/en/index.html), the following members:
- [Rio Yokota](https://twitter.com/rioyokota)
- [Kazuki Fujii](https://twitter.com/okoge_kaz)
- [Taishi Nakamura](https://twitter.com/Setuna7777_2)
<!-- original-model-card end -->
|
LKGSR/Mistral-7B-Instruct-v0.1-GuteFrage
|
LKGSR
| 2023-12-19T22:00:37Z | 4 | 0 |
peft
|
[
"peft",
"safetensors",
"arxiv:1910.09700",
"base_model:mistralai/Mistral-7B-Instruct-v0.1",
"base_model:adapter:mistralai/Mistral-7B-Instruct-v0.1",
"region:us"
] | null | 2023-12-19T20:31:42Z |
---
library_name: peft
base_model: mistralai/Mistral-7B-Instruct-v0.1
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
### Framework versions
- PEFT 0.7.1
|
mdaffarudiyanto/t5-small-finetuned-xsum-updated
|
mdaffarudiyanto
| 2023-12-19T21:57:04Z | 5 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"t5",
"text2text-generation",
"generated_from_trainer",
"dataset:xsum",
"base_model:google-t5/t5-small",
"base_model:finetune:google-t5/t5-small",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2023-12-19T11:45:04Z |
---
license: apache-2.0
base_model: t5-small
tags:
- generated_from_trainer
datasets:
- xsum
metrics:
- rouge
model-index:
- name: t5-small-finetuned-xsum-updated
results:
- task:
name: Sequence-to-sequence Language Modeling
type: text2text-generation
dataset:
name: xsum
type: xsum
config: default
split: validation
args: default
metrics:
- name: Rouge1
type: rouge
value: 33.2945
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# t5-small-finetuned-xsum-updated
This model is a fine-tuned version of [t5-small](https://huggingface.co/t5-small) on the xsum dataset.
It achieves the following results on the evaluation set:
- Loss: 2.0767
- Rouge1: 33.2945
- Rouge2: 12.0165
- Rougel: 26.9804
- Rougelsum: 26.9729
- Gen Len: 18.7853
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0001
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 15
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|:-------------:|:-----:|:------:|:---------------:|:-------:|:-------:|:-------:|:---------:|:-------:|
| 2.5219 | 1.0 | 12753 | 2.3054 | 30.4745 | 9.435 | 24.263 | 24.2522 | 18.823 |
| 2.4191 | 2.0 | 25506 | 2.2385 | 31.2305 | 10.0552 | 24.9345 | 24.9254 | 18.7562 |
| 2.3564 | 3.0 | 38259 | 2.1961 | 31.8234 | 10.6556 | 25.6109 | 25.6023 | 18.7708 |
| 2.3028 | 4.0 | 51012 | 2.1692 | 32.2053 | 11.0513 | 26.0184 | 26.0056 | 18.772 |
| 2.2737 | 5.0 | 63765 | 2.1452 | 32.3716 | 11.1779 | 26.1423 | 26.1363 | 18.7731 |
| 2.2432 | 6.0 | 76518 | 2.1304 | 32.5413 | 11.2517 | 26.2119 | 26.2098 | 18.8007 |
| 2.2266 | 7.0 | 89271 | 2.1193 | 32.8983 | 11.5683 | 26.5995 | 26.5958 | 18.8108 |
| 2.1863 | 8.0 | 102024 | 2.1058 | 32.9046 | 11.6564 | 26.6466 | 26.6473 | 18.8008 |
| 2.1583 | 9.0 | 114777 | 2.0987 | 32.9622 | 11.7285 | 26.7161 | 26.7116 | 18.7798 |
| 2.1653 | 10.0 | 127530 | 2.0900 | 33.1259 | 11.8525 | 26.8461 | 26.8419 | 18.7999 |
| 2.1403 | 11.0 | 140283 | 2.0880 | 33.0949 | 11.8135 | 26.7863 | 26.7765 | 18.7629 |
| 2.1212 | 12.0 | 153036 | 2.0825 | 33.1671 | 11.8939 | 26.9072 | 26.8982 | 18.7825 |
| 2.1021 | 13.0 | 165789 | 2.0793 | 33.1375 | 11.9119 | 26.8466 | 26.8386 | 18.8076 |
| 2.0877 | 14.0 | 178542 | 2.0774 | 33.2516 | 11.9574 | 26.9391 | 26.9327 | 18.7989 |
| 2.0984 | 15.0 | 191295 | 2.0767 | 33.2945 | 12.0165 | 26.9804 | 26.9729 | 18.7853 |
### Framework versions
- Transformers 4.35.2
- Pytorch 2.1.0+cu121
- Datasets 2.15.0
- Tokenizers 0.15.0
|
Bilic/NeuralChat-finetuned-for-fraud-detection
|
Bilic
| 2023-12-19T21:55:11Z | 54 | 2 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"safetensors",
"mistral",
"text-generation",
"generated_from_trainer",
"base_model:Intel/neural-chat-7b-v3-1",
"base_model:finetune:Intel/neural-chat-7b-v3-1",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2023-12-06T12:49:36Z |
---
license: apache-2.0
base_model: Intel/neural-chat-7b-v3-1
tags:
- generated_from_trainer
model-index:
- name: neural-chat-finetuned-bilic-v1
results: []
---
# neural-chat-finetuned-bilic-v1
This model is a fine-tuned version of [Intel/neural-chat-7b-v3-1](https://huggingface.co/Intel/neural-chat-7b-v3-1) on our custom dataset.
## Model description
This is a fine tuned version of the intel's Neuralchat model, specifically trained on a carefully curated dataset on fraud detection. We implemented a contextual based architecture to enable the model learn and be adept at understanding context within a conversation as opposed to the traditional rule based approach.
## Intended uses & limitations
- detecting fraudulent conversations in real-time
- Giving a summary of conversations and suggestions
- Understanding with high accuracy the context in a conversation to make better predictions
## Training
50,000 synthetically conversations
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0002
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- training_steps: 250
- mixed_precision_training: Native AMP
### Framework versions
- Transformers 4.36.0.dev0
- Pytorch 2.1.0+cu118
- Datasets 2.15.0
- Tokenizers 0.15.0
|
TheBloke/Swallow-70B-instruct-GPTQ
|
TheBloke
| 2023-12-19T21:50:15Z | 25 | 3 |
transformers
|
[
"transformers",
"safetensors",
"llama",
"text-generation",
"en",
"ja",
"base_model:tokyotech-llm/Swallow-70b-instruct-hf",
"base_model:quantized:tokyotech-llm/Swallow-70b-instruct-hf",
"license:llama2",
"autotrain_compatible",
"text-generation-inference",
"4-bit",
"gptq",
"region:us"
] |
text-generation
| 2023-12-19T18:37:03Z |
---
base_model: tokyotech-llm/Swallow-70b-instruct-hf
inference: false
language:
- en
- ja
library_name: transformers
license: llama2
model_creator: tokyotech-llm
model_name: Swallow 70B Instruct
model_type: llama
pipeline_tag: text-generation
prompt_template: "\u4EE5\u4E0B\u306B\u3001\u3042\u308B\u30BF\u30B9\u30AF\u3092\u8AAC\
\u660E\u3059\u308B\u6307\u793A\u304C\u3042\u308A\u307E\u3059\u3002\u30EA\u30AF\u30A8\
\u30B9\u30C8\u3092\u9069\u5207\u306B\u5B8C\u4E86\u3059\u308B\u305F\u3081\u306E\u56DE\
\u7B54\u3092\u8A18\u8FF0\u3057\u3066\u304F\u3060\u3055\u3044\u3002\\n\\n### \u6307\
\u793A:\\n{prompt}\\n\\n### \u5FDC\u7B54:\n"
quantized_by: TheBloke
---
<!-- markdownlint-disable MD041 -->
<!-- header start -->
<!-- 200823 -->
<div style="width: auto; margin-left: auto; margin-right: auto">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<div style="text-align:center; margin-top: 0em; margin-bottom: 0em"><p style="margin-top: 0.25em; margin-bottom: 0em;">TheBloke's LLM work is generously supported by a grant from <a href="https://a16z.com">andreessen horowitz (a16z)</a></p></div>
<hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
<!-- header end -->
# Swallow 70B Instruct - GPTQ
- Model creator: [tokyotech-llm](https://huggingface.co/tokyotech-llm)
- Original model: [Swallow 70B Instruct](https://huggingface.co/tokyotech-llm/Swallow-70b-instruct-hf)
<!-- description start -->
# Description
This repo contains GPTQ model files for [tokyotech-llm's Swallow 70B Instruct](https://huggingface.co/tokyotech-llm/Swallow-70b-instruct-hf).
Multiple GPTQ parameter permutations are provided; see Provided Files below for details of the options provided, their parameters, and the software used to create them.
<!-- description end -->
<!-- repositories-available start -->
## Repositories available
* [AWQ model(s) for GPU inference.](https://huggingface.co/TheBloke/Swallow-70B-instruct-AWQ)
* [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/Swallow-70B-instruct-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/Swallow-70B-instruct-GGUF)
* [tokyotech-llm's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/tokyotech-llm/Swallow-70b-instruct-hf)
<!-- repositories-available end -->
<!-- prompt-template start -->
## Prompt template: Swallow-Instruct
```
以下に、あるタスクを説明する指示があります。リクエストを適切に完了するための回答を記述してください。\n\n### 指示:\n{prompt}\n\n### 応答:
```
<!-- prompt-template end -->
<!-- README_GPTQ.md-compatible clients start -->
## Known compatible clients / servers
GPTQ models are currently supported on Linux (NVidia/AMD) and Windows (NVidia only). macOS users: please use GGUF models.
These GPTQ models are known to work in the following inference servers/webuis.
- [text-generation-webui](https://github.com/oobabooga/text-generation-webui)
- [KoboldAI United](https://github.com/henk717/koboldai)
- [LoLLMS Web UI](https://github.com/ParisNeo/lollms-webui)
- [Hugging Face Text Generation Inference (TGI)](https://github.com/huggingface/text-generation-inference)
This may not be a complete list; if you know of others, please let me know!
<!-- README_GPTQ.md-compatible clients end -->
<!-- README_GPTQ.md-provided-files start -->
## Provided files, and GPTQ parameters
Multiple quantisation parameters are provided, to allow you to choose the best one for your hardware and requirements.
Each separate quant is in a different branch. See below for instructions on fetching from different branches.
Most GPTQ files are made with AutoGPTQ. Mistral models are currently made with Transformers.
<details>
<summary>Explanation of GPTQ parameters</summary>
- Bits: The bit size of the quantised model.
- GS: GPTQ group size. Higher numbers use less VRAM, but have lower quantisation accuracy. "None" is the lowest possible value.
- Act Order: True or False. Also known as `desc_act`. True results in better quantisation accuracy. Some GPTQ clients have had issues with models that use Act Order plus Group Size, but this is generally resolved now.
- Damp %: A GPTQ parameter that affects how samples are processed for quantisation. 0.01 is default, but 0.1 results in slightly better accuracy.
- GPTQ dataset: The calibration dataset used during quantisation. Using a dataset more appropriate to the model's training can improve quantisation accuracy. Note that the GPTQ calibration dataset is not the same as the dataset used to train the model - please refer to the original model repo for details of the training dataset(s).
- Sequence Length: The length of the dataset sequences used for quantisation. Ideally this is the same as the model sequence length. For some very long sequence models (16+K), a lower sequence length may have to be used. Note that a lower sequence length does not limit the sequence length of the quantised model. It only impacts the quantisation accuracy on longer inference sequences.
- ExLlama Compatibility: Whether this file can be loaded with ExLlama, which currently only supports Llama and Mistral models in 4-bit.
</details>
| Branch | Bits | GS | Act Order | Damp % | GPTQ Dataset | Seq Len | Size | ExLlama | Desc |
| ------ | ---- | -- | --------- | ------ | ------------ | ------- | ---- | ------- | ---- |
| [main](https://huggingface.co/TheBloke/Swallow-70B-instruct-GPTQ/tree/main) | 4 | None | Yes | 0.1 | [Alpaca Japanese](https://huggingface.co/datasets/fujiki/japanese_alpaca_data/viewer/) | 4096 | 35.70 GB | Yes | 4-bit, with Act Order. No group size, to lower VRAM requirements. |
| [gptq-4bit-128g-actorder_True](https://huggingface.co/TheBloke/Swallow-70B-instruct-GPTQ/tree/gptq-4bit-128g-actorder_True) | 4 | 128 | Yes | 0.1 | [Alpaca Japanese](https://huggingface.co/datasets/fujiki/japanese_alpaca_data/viewer/) | 4096 | 37.02 GB | Yes | 4-bit, with Act Order and group size 128g. Uses even less VRAM than 64g, but with slightly lower accuracy. |
| [gptq-4bit-32g-actorder_True](https://huggingface.co/TheBloke/Swallow-70B-instruct-GPTQ/tree/gptq-4bit-32g-actorder_True) | 4 | 32 | Yes | 0.1 | [Alpaca Japanese](https://huggingface.co/datasets/fujiki/japanese_alpaca_data/viewer/) | 4096 | 41.03 GB | Yes | 4-bit, with Act Order and group size 32g. Gives highest possible inference quality, with maximum VRAM usage. |
| [gptq-3bit--1g-actorder_True](https://huggingface.co/TheBloke/Swallow-70B-instruct-GPTQ/tree/gptq-3bit--1g-actorder_True) | 3 | None | Yes | 0.1 | [Alpaca Japanese](https://huggingface.co/datasets/fujiki/japanese_alpaca_data/viewer/) | 4096 | 27.14 GB | No | 3-bit, with Act Order and no group size. Lowest possible VRAM requirements. May be lower quality than 3-bit 128g. |
| [gptq-3bit-128g-actorder_True](https://huggingface.co/TheBloke/Swallow-70B-instruct-GPTQ/tree/gptq-3bit-128g-actorder_True) | 3 | 128 | Yes | 0.1 | [Alpaca Japanese](https://huggingface.co/datasets/fujiki/japanese_alpaca_data/viewer/) | 4096 | 28.40 GB | No | 3-bit, with group size 128g and act-order. Higher quality than 128g-False. |
| [gptq-3bit-32g-actorder_True](https://huggingface.co/TheBloke/Swallow-70B-instruct-GPTQ/tree/gptq-3bit-32g-actorder_True) | 3 | 32 | Yes | 0.1 | [Alpaca Japanese](https://huggingface.co/datasets/fujiki/japanese_alpaca_data/viewer/) | 4096 | 32.21 GB | No | 3-bit, with group size 64g and act-order. Highest quality 3-bit option. |
<!-- README_GPTQ.md-provided-files end -->
<!-- README_GPTQ.md-download-from-branches start -->
## How to download, including from branches
### In text-generation-webui
To download from the `main` branch, enter `TheBloke/Swallow-70B-instruct-GPTQ` in the "Download model" box.
To download from another branch, add `:branchname` to the end of the download name, eg `TheBloke/Swallow-70B-instruct-GPTQ:gptq-4bit-128g-actorder_True`
### From the command line
I recommend using the `huggingface-hub` Python library:
```shell
pip3 install huggingface-hub
```
To download the `main` branch to a folder called `Swallow-70B-instruct-GPTQ`:
```shell
mkdir Swallow-70B-instruct-GPTQ
huggingface-cli download TheBloke/Swallow-70B-instruct-GPTQ --local-dir Swallow-70B-instruct-GPTQ --local-dir-use-symlinks False
```
To download from a different branch, add the `--revision` parameter:
```shell
mkdir Swallow-70B-instruct-GPTQ
huggingface-cli download TheBloke/Swallow-70B-instruct-GPTQ --revision gptq-4bit-128g-actorder_True --local-dir Swallow-70B-instruct-GPTQ --local-dir-use-symlinks False
```
<details>
<summary>More advanced huggingface-cli download usage</summary>
If you remove the `--local-dir-use-symlinks False` parameter, the files will instead be stored in the central Hugging Face cache directory (default location on Linux is: `~/.cache/huggingface`), and symlinks will be added to the specified `--local-dir`, pointing to their real location in the cache. This allows for interrupted downloads to be resumed, and allows you to quickly clone the repo to multiple places on disk without triggering a download again. The downside, and the reason why I don't list that as the default option, is that the files are then hidden away in a cache folder and it's harder to know where your disk space is being used, and to clear it up if/when you want to remove a download model.
The cache location can be changed with the `HF_HOME` environment variable, and/or the `--cache-dir` parameter to `huggingface-cli`.
For more documentation on downloading with `huggingface-cli`, please see: [HF -> Hub Python Library -> Download files -> Download from the CLI](https://huggingface.co/docs/huggingface_hub/guides/download#download-from-the-cli).
To accelerate downloads on fast connections (1Gbit/s or higher), install `hf_transfer`:
```shell
pip3 install hf_transfer
```
And set environment variable `HF_HUB_ENABLE_HF_TRANSFER` to `1`:
```shell
mkdir Swallow-70B-instruct-GPTQ
HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download TheBloke/Swallow-70B-instruct-GPTQ --local-dir Swallow-70B-instruct-GPTQ --local-dir-use-symlinks False
```
Windows Command Line users: You can set the environment variable by running `set HF_HUB_ENABLE_HF_TRANSFER=1` before the download command.
</details>
### With `git` (**not** recommended)
To clone a specific branch with `git`, use a command like this:
```shell
git clone --single-branch --branch gptq-4bit-128g-actorder_True https://huggingface.co/TheBloke/Swallow-70B-instruct-GPTQ
```
Note that using Git with HF repos is strongly discouraged. It will be much slower than using `huggingface-hub`, and will use twice as much disk space as it has to store the model files twice (it stores every byte both in the intended target folder, and again in the `.git` folder as a blob.)
<!-- README_GPTQ.md-download-from-branches end -->
<!-- README_GPTQ.md-text-generation-webui start -->
## How to easily download and use this model in [text-generation-webui](https://github.com/oobabooga/text-generation-webui)
Please make sure you're using the latest version of [text-generation-webui](https://github.com/oobabooga/text-generation-webui).
It is strongly recommended to use the text-generation-webui one-click-installers unless you're sure you know how to make a manual install.
1. Click the **Model tab**.
2. Under **Download custom model or LoRA**, enter `TheBloke/Swallow-70B-instruct-GPTQ`.
- To download from a specific branch, enter for example `TheBloke/Swallow-70B-instruct-GPTQ:gptq-4bit-128g-actorder_True`
- see Provided Files above for the list of branches for each option.
3. Click **Download**.
4. The model will start downloading. Once it's finished it will say "Done".
5. In the top left, click the refresh icon next to **Model**.
6. In the **Model** dropdown, choose the model you just downloaded: `Swallow-70B-instruct-GPTQ`
7. The model will automatically load, and is now ready for use!
8. If you want any custom settings, set them and then click **Save settings for this model** followed by **Reload the Model** in the top right.
- Note that you do not need to and should not set manual GPTQ parameters any more. These are set automatically from the file `quantize_config.json`.
9. Once you're ready, click the **Text Generation** tab and enter a prompt to get started!
<!-- README_GPTQ.md-text-generation-webui end -->
<!-- README_GPTQ.md-use-from-tgi start -->
## Serving this model from Text Generation Inference (TGI)
It's recommended to use TGI version 1.1.0 or later. The official Docker container is: `ghcr.io/huggingface/text-generation-inference:1.1.0`
Example Docker parameters:
```shell
--model-id TheBloke/Swallow-70B-instruct-GPTQ --port 3000 --quantize gptq --max-input-length 3696 --max-total-tokens 4096 --max-batch-prefill-tokens 4096
```
Example Python code for interfacing with TGI (requires huggingface-hub 0.17.0 or later):
```shell
pip3 install huggingface-hub
```
```python
from huggingface_hub import InferenceClient
endpoint_url = "https://your-endpoint-url-here"
prompt = "Tell me about AI"
prompt_template=f'''以下に、あるタスクを説明する指示があります。リクエストを適切に完了するための回答を記述してください。\n\n### 指示:\n{prompt}\n\n### 応答:
'''
client = InferenceClient(endpoint_url)
response = client.text_generation(prompt,
max_new_tokens=128,
do_sample=True,
temperature=0.7,
top_p=0.95,
top_k=40,
repetition_penalty=1.1)
print(f"Model output: {response}")
```
<!-- README_GPTQ.md-use-from-tgi end -->
<!-- README_GPTQ.md-use-from-python start -->
## Python code example: inference from this GPTQ model
### Install the necessary packages
Requires: Transformers 4.33.0 or later, Optimum 1.12.0 or later, and AutoGPTQ 0.4.2 or later.
```shell
pip3 install --upgrade transformers optimum
# If using PyTorch 2.1 + CUDA 12.x:
pip3 install --upgrade auto-gptq
# or, if using PyTorch 2.1 + CUDA 11.x:
pip3 install --upgrade auto-gptq --extra-index-url https://huggingface.github.io/autogptq-index/whl/cu118/
```
If you are using PyTorch 2.0, you will need to install AutoGPTQ from source. Likewise if you have problems with the pre-built wheels, you should try building from source:
```shell
pip3 uninstall -y auto-gptq
git clone https://github.com/PanQiWei/AutoGPTQ
cd AutoGPTQ
git checkout v0.5.1
pip3 install .
```
### Example Python code
```python
from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
model_name_or_path = "TheBloke/Swallow-70B-instruct-GPTQ"
# To use a different branch, change revision
# For example: revision="gptq-4bit-128g-actorder_True"
model = AutoModelForCausalLM.from_pretrained(model_name_or_path,
device_map="auto",
trust_remote_code=False,
revision="main")
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
prompt = "Write a story about llamas"
system_message = "You are a story writing assistant"
prompt_template=f'''以下に、あるタスクを説明する指示があります。リクエストを適切に完了するための回答を記述してください。\n\n### 指示:\n{prompt}\n\n### 応答:
'''
print("\n\n*** Generate:")
input_ids = tokenizer(prompt_template, return_tensors='pt').input_ids.cuda()
output = model.generate(inputs=input_ids, temperature=0.7, do_sample=True, top_p=0.95, top_k=40, max_new_tokens=512)
print(tokenizer.decode(output[0]))
# Inference can also be done using transformers' pipeline
print("*** Pipeline:")
pipe = pipeline(
"text-generation",
model=model,
tokenizer=tokenizer,
max_new_tokens=512,
do_sample=True,
temperature=0.7,
top_p=0.95,
top_k=40,
repetition_penalty=1.1
)
print(pipe(prompt_template)[0]['generated_text'])
```
<!-- README_GPTQ.md-use-from-python end -->
<!-- README_GPTQ.md-compatibility start -->
## Compatibility
The files provided are tested to work with Transformers. For non-Mistral models, AutoGPTQ can also be used directly.
[ExLlama](https://github.com/turboderp/exllama) is compatible with Llama architecture models (including Mistral, Yi, DeepSeek, SOLAR, etc) in 4-bit. Please see the Provided Files table above for per-file compatibility.
For a list of clients/servers, please see "Known compatible clients / servers", above.
<!-- README_GPTQ.md-compatibility end -->
<!-- footer start -->
<!-- 200823 -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute
Thanks to the [chirper.ai](https://chirper.ai) team!
Thanks to Clay from [gpus.llm-utils.org](llm-utils)!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Aemon Algiz.
**Patreon special mentions**: Michael Levine, 阿明, Trailburnt, Nikolai Manek, John Detwiler, Randy H, Will Dee, Sebastain Graf, NimbleBox.ai, Eugene Pentland, Emad Mostaque, Ai Maven, Jim Angel, Jeff Scroggin, Michael Davis, Manuel Alberto Morcote, Stephen Murray, Robert, Justin Joy, Luke @flexchar, Brandon Frisco, Elijah Stavena, S_X, Dan Guido, Undi ., Komninos Chatzipapas, Shadi, theTransient, Lone Striker, Raven Klaugh, jjj, Cap'n Zoog, Michel-Marie MAUDET (LINAGORA), Matthew Berman, David, Fen Risland, Omer Bin Jawed, Luke Pendergrass, Kalila, OG, Erik Bjäreholt, Rooh Singh, Joseph William Delisle, Dan Lewis, TL, John Villwock, AzureBlack, Brad, Pedro Madruga, Caitlyn Gatomon, K, jinyuan sun, Mano Prime, Alex, Jeffrey Morgan, Alicia Loh, Illia Dulskyi, Chadd, transmissions 11, fincy, Rainer Wilmers, ReadyPlayerEmma, knownsqashed, Mandus, biorpg, Deo Leter, Brandon Phillips, SuperWojo, Sean Connelly, Iucharbius, Jack West, Harry Royden McLaughlin, Nicholas, terasurfer, Vitor Caleffi, Duane Dunston, Johann-Peter Hartmann, David Ziegler, Olakabola, Ken Nordquist, Trenton Dambrowitz, Tom X Nguyen, Vadim, Ajan Kanaga, Leonard Tan, Clay Pascal, Alexandros Triantafyllidis, JM33133, Xule, vamX, ya boyyy, subjectnull, Talal Aujan, Alps Aficionado, wassieverse, Ari Malik, James Bentley, Woland, Spencer Kim, Michael Dempsey, Fred von Graf, Elle, zynix, William Richards, Stanislav Ovsiannikov, Edmond Seymore, Jonathan Leane, Martin Kemka, usrbinkat, Enrico Ros
Thank you to all my generous patrons and donaters!
And thank you again to a16z for their generous grant.
<!-- footer end -->
# Original model card: tokyotech-llm's Swallow 70B Instruct
# Swallow
Our Swallow model has undergone continuous pre-training from the Llama 2 family, primarily with the addition of Japanese language data. The tuned versions use supervised fine-tuning (SFT).
Links to other models can be found in the index.
## Swallow Model Index
|Model|Swallow-hf|Swallow-instruct-hf|
|---|---|---|
|7B| [Link](https://huggingface.co/tokyotech-llm/Swallow-7b-hf) | [Link](https://huggingface.co/tokyotech-llm/Swallow-7b-instruct-hf)|
|13B| [Link](https://huggingface.co/tokyotech-llm/Swallow-13b-hf) | [Link](https://huggingface.co/tokyotech-llm/Swallow-13b-instruct-hf)|
|70B| [Link](https://huggingface.co/tokyotech-llm/Swallow-70b-hf) | [Link](https://huggingface.co/tokyotech-llm/Swallow-70b-instruct-hf)|
This repository provides large language models developed by [TokyoTech-LLM](https://tokyotech-llm.github.io/).
Read our [blog post](https://zenn.dev/tokyotech_lm/articles/d6cb3a8fdfc907) or our paper (preprint coming soon) for more details!
## Model Details
* **Model type**: Please refer to LLaMA-2 technical report for details on the model architecture.
* **Language(s)**: Japanese English
* **Library**: [Megatron-LM](https://github.com/rioyokotalab/Megatron-Llama2)
* **Tokenizer**: This model employs a tokenizer that features a broadened vocabulary based on Japanese data. This allows for a more efficient representation of text using fewer tokens, leading to a notably faster inference process.
* **Contact**: swallow[at]nlp.c.titech.ac.jp
## Base Model Performance
### Japanese version
|Model|Size|JCommonsenseQA|JEMHopQA|NIILC|JSQuAD|XL-Sum|MGSM|WMT20-en-ja|WMT20-ja-en|
|---|---|---|---|---|---|---|---|---|---|
| | |4-shot|4-shot|4-shot|4-shot|1-shot|4-shot|4-shot|4-shot|
|Llama 2|7B|0.3852|0.4240|0.3410|0.7917|0.1905|0.0760|0.1783|0.1738|
|Swallow|7B|0.4808|0.5078|0.5968|0.8573|0.1830|0.1240|0.2510|0.1511|
|Llama 2|13B|0.6997|0.4415|0.4170|0.8533|0.2139|0.1320|0.2146|0.1982|
|Swallow|13B|0.7837|0.5063|0.6398|0.9005|0.2168|0.2040|0.2720|0.1771|
|Llama 2|70B|0.8686|0.4656|0.5256|0.9080|**0.2361**|0.3560|0.2643|**0.2398**|
|Swallow|70B|**0.9348**|**0.6290**|**0.6960**|**0.9176**|0.2266|**0.4840**|**0.3043**|0.2298|
## Usage
First install additional dependencies in [requirements.txt](./requirements.txt):
```sh
pip install -r requirements.txt
```
### Use the instruct model
```python
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
model_name = "tokyotech-llm/Swallow-7b-instruct-hf"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.bfloat16, low_cpu_mem_usage=True, device_map="auto")
PROMPT_DICT = {
"prompt_input": (
"以下に、あるタスクを説明する指示があり、それに付随する入力が更なる文脈を提供しています。"
"リクエストを適切に完了するための回答を記述してください。\n\n"
"### 指示:\n{instruction}\n\n### 入力:\n{input}\n\n### 応答:"
),
"prompt_no_input": (
"以下に、あるタスクを説明する指示があります。"
"リクエストを適切に完了するための回答を記述してください。\n\n"
"### 指示:\n{instruction}\n\n### 応答:"
),
}
def create_prompt(instruction, input=None):
"""
Generates a prompt based on the given instruction and an optional input.
If input is provided, it uses the 'prompt_input' template from PROMPT_DICT.
If no input is provided, it uses the 'prompt_no_input' template.
Args:
instruction (str): The instruction describing the task.
input (str, optional): Additional input providing context for the task. Default is None.
Returns:
str: The generated prompt.
"""
if input:
# Use the 'prompt_input' template when additional input is provided
return PROMPT_DICT["prompt_input"].format(instruction=instruction, input=input)
else:
# Use the 'prompt_no_input' template when no additional input is provided
return PROMPT_DICT["prompt_no_input"].format(instruction=instruction)
# Example usage
instruction_example = "以下のトピックに関する詳細な情報を提供してください。"
input_example = "東京工業大学の主なキャンパスについて教えてください"
prompt = create_prompt(instruction_example, input_example)
input_ids = tokenizer.encode(
prompt,
add_special_tokens=False,
return_tensors="pt"
)
tokens = model.generate(
input_ids.to(device=model.device),
max_new_tokens=128,
temperature=0.99,
top_p=0.95,
do_sample=True,
)
out = tokenizer.decode(tokens[0], skip_special_tokens=True)
print(out)
```
### Use the base model
```python
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
model_name = "tokyotech-llm/Swallow-7b-hf"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.bfloat16, device_map="auto")
prompt = "東京工業大学の主なキャンパスは、"
input_ids = tokenizer.encode(
prompt,
add_special_tokens=False,
return_tensors="pt"
)
tokens = model.generate(
input_ids.to(device=model.device),
max_new_tokens=128,
temperature=0.99,
top_p=0.95,
do_sample=True,
)
out = tokenizer.decode(tokens[0], skip_special_tokens=True)
print(out)
```
## Training Datasets
### Continual Pre-Training
The following datasets were used for continual pre-training.
- [Japanese Wikipedia](https://dumps.wikimedia.org/other/cirrussearch)
- [RefinedWeb](https://huggingface.co/datasets/tiiuae/falcon-refinedweb)
- Swallow Corpus
- [The Pile](https://huggingface.co/datasets/EleutherAI/pile)
### Instruction Tuning
The following datasets were used for the instruction tuning.
- [Anthropic HH-RLHF](https://huggingface.co/datasets/kunishou/hh-rlhf-49k-ja)
- [Databricks Dolly 15-k](https://huggingface.co/datasets/kunishou/databricks-dolly-15k-ja)
- [OpenAssistant Conversations Dataset](https://huggingface.co/datasets/kunishou/oasst1-89k-ja)
## Risks and Limitations
The models released here are still in the early stages of our research and development and have not been tuned to ensure outputs align with human intent and safety considerations.
## Acknowledgements
We thank Meta Research for releasing Llama 2 under an open license for others to build on.
Our project is supported by the [ABCI Large-scale Language Model Building Support Program](https://abci.ai/en/link/llm_support_program.html) of the National Institute of Advanced Industrial Science and Technology.
## License
Llama 2 is licensed under the LLAMA 2 Community License, Copyright © Meta Platforms, Inc. All Rights Reserved.
## Authors
Here are the team members:
- From [Okazaki Laboratory](https://www.nlp.c.titech.ac.jp/index.en.html), the following members:
- [Naoaki Okazaki](https://www.chokkan.org/index.ja.html)
- [Sakae Mizuki](https://s-mizuki-nlp.github.io/)
- [Hiroki Iida](https://meshidenn.github.io/)
- [Mengsay Loem](https://loem-ms.github.io/)
- [Shota Hirai](https://huggingface.co/Kotemo428)
- [Kakeru Hattori](https://aya-se.vercel.app/)
- [Masanari Ohi](https://twitter.com/stjohn2007)
- From [YOKOTA Laboratory](https://www.rio.gsic.titech.ac.jp/en/index.html), the following members:
- [Rio Yokota](https://twitter.com/rioyokota)
- [Kazuki Fujii](https://twitter.com/okoge_kaz)
- [Taishi Nakamura](https://twitter.com/Setuna7777_2)
|
obrytskyy/Lab3
|
obrytskyy
| 2023-12-19T21:46:01Z | 0 | 0 |
stable-baselines3
|
[
"stable-baselines3",
"LunarLander-v2",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-12-19T21:45:29Z |
---
library_name: stable-baselines3
tags:
- LunarLander-v2
- deep-reinforcement-learning
- reinforcement-learning
- stable-baselines3
model-index:
- name: PPO
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: LunarLander-v2
type: LunarLander-v2
metrics:
- type: mean_reward
value: 259.05 +/- 18.24
name: mean_reward
verified: false
---
# **PPO** Agent playing **LunarLander-v2**
This is a trained model of a **PPO** agent playing **LunarLander-v2**
using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3).
## Usage (with Stable-baselines3)
TODO: Add your code
```python
from stable_baselines3 import ...
from huggingface_sb3 import load_from_hub
...
```
|
TheBloke/openchat-3.5-1210-AWQ
|
TheBloke
| 2023-12-19T21:40:08Z | 66 | 2 |
transformers
|
[
"transformers",
"safetensors",
"mistral",
"text-generation",
"openchat",
"C-RLFT",
"conversational",
"dataset:openchat/openchat_sharegpt4_dataset",
"dataset:kaist-ai/Feedback-Collection",
"dataset:imone/OpenOrca_FLAN",
"dataset:LDJnr/LessWrong-Amplify-Instruct",
"dataset:LDJnr/Pure-Dove",
"dataset:LDJnr/Verified-Camel",
"dataset:tiedong/goat",
"dataset:glaiveai/glaive-code-assistant",
"dataset:meta-math/MetaMathQA",
"dataset:OpenAssistant/oasst_top1_2023-08-25",
"dataset:TIGER-Lab/MathInstruct",
"arxiv:2309.11235",
"arxiv:2303.08774",
"arxiv:2212.10560",
"base_model:openchat/openchat-3.5-1210",
"base_model:quantized:openchat/openchat-3.5-1210",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"4-bit",
"awq",
"region:us"
] |
text-generation
| 2023-12-14T14:35:33Z |
---
base_model: openchat/openchat-3.5-1210
datasets:
- openchat/openchat_sharegpt4_dataset
- kaist-ai/Feedback-Collection
- imone/OpenOrca_FLAN
- LDJnr/LessWrong-Amplify-Instruct
- LDJnr/Pure-Dove
- LDJnr/Verified-Camel
- tiedong/goat
- glaiveai/glaive-code-assistant
- meta-math/MetaMathQA
- OpenAssistant/oasst_top1_2023-08-25
- TIGER-Lab/MathInstruct
inference: false
library_name: transformers
license: apache-2.0
model_creator: OpenChat
model_name: Openchat 3.5 1210
model_type: mistral
pipeline_tag: text-generation
prompt_template: 'GPT4 Correct User: {prompt}<|end_of_turn|>GPT4 Correct Assistant:
'
quantized_by: TheBloke
tags:
- openchat
- mistral
- C-RLFT
---
<!-- markdownlint-disable MD041 -->
<!-- header start -->
<!-- 200823 -->
<div style="width: auto; margin-left: auto; margin-right: auto">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<div style="text-align:center; margin-top: 0em; margin-bottom: 0em"><p style="margin-top: 0.25em; margin-bottom: 0em;">TheBloke's LLM work is generously supported by a grant from <a href="https://a16z.com">andreessen horowitz (a16z)</a></p></div>
<hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
<!-- header end -->
# Openchat 3.5 1210 - AWQ
- Model creator: [OpenChat](https://huggingface.co/openchat)
- Original model: [Openchat 3.5 1210](https://huggingface.co/openchat/openchat-3.5-1210)
<!-- description start -->
## Description
This repo contains AWQ model files for [OpenChat's Openchat 3.5 1210](https://huggingface.co/openchat/openchat-3.5-1210).
These files were quantised using hardware kindly provided by [Massed Compute](https://massedcompute.com/).
### About AWQ
AWQ is an efficient, accurate and blazing-fast low-bit weight quantization method, currently supporting 4-bit quantization. Compared to GPTQ, it offers faster Transformers-based inference with equivalent or better quality compared to the most commonly used GPTQ settings.
AWQ models are currently supported on Linux and Windows, with NVidia GPUs only. macOS users: please use GGUF models instead.
It is supported by:
- [Text Generation Webui](https://github.com/oobabooga/text-generation-webui) - using Loader: AutoAWQ
- [vLLM](https://github.com/vllm-project/vllm) - version 0.2.2 or later for support for all model types.
- [Hugging Face Text Generation Inference (TGI)](https://github.com/huggingface/text-generation-inference)
- [Transformers](https://huggingface.co/docs/transformers) version 4.35.0 and later, from any code or client that supports Transformers
- [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) - for use from Python code
<!-- description end -->
<!-- repositories-available start -->
## Repositories available
* [AWQ model(s) for GPU inference.](https://huggingface.co/TheBloke/openchat-3.5-1210-AWQ)
* [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/openchat-3.5-1210-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/openchat-3.5-1210-GGUF)
* [OpenChat's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/openchat/openchat-3.5-1210)
<!-- repositories-available end -->
<!-- prompt-template start -->
## Prompt template: OpenChat-Correct
```
GPT4 Correct User: {prompt}<|end_of_turn|>GPT4 Correct Assistant:
```
<!-- prompt-template end -->
<!-- README_AWQ.md-provided-files start -->
## Provided files, and AWQ parameters
I currently release 128g GEMM models only. The addition of group_size 32 models, and GEMV kernel models, is being actively considered.
Models are released as sharded safetensors files.
| Branch | Bits | GS | AWQ Dataset | Seq Len | Size |
| ------ | ---- | -- | ----------- | ------- | ---- |
| [main](https://huggingface.co/TheBloke/openchat-3.5-1210-AWQ/tree/main) | 4 | 128 | [VMware Open Instruct](https://huggingface.co/datasets/VMware/open-instruct/viewer/) | 4096 | 4.15 GB
<!-- README_AWQ.md-provided-files end -->
<!-- README_AWQ.md-text-generation-webui start -->
## How to easily download and use this model in [text-generation-webui](https://github.com/oobabooga/text-generation-webui)
Please make sure you're using the latest version of [text-generation-webui](https://github.com/oobabooga/text-generation-webui).
It is strongly recommended to use the text-generation-webui one-click-installers unless you're sure you know how to make a manual install.
1. Click the **Model tab**.
2. Under **Download custom model or LoRA**, enter `TheBloke/openchat-3.5-1210-AWQ`.
3. Click **Download**.
4. The model will start downloading. Once it's finished it will say "Done".
5. In the top left, click the refresh icon next to **Model**.
6. In the **Model** dropdown, choose the model you just downloaded: `openchat-3.5-1210-AWQ`
7. Select **Loader: AutoAWQ**.
8. Click Load, and the model will load and is now ready for use.
9. If you want any custom settings, set them and then click **Save settings for this model** followed by **Reload the Model** in the top right.
10. Once you're ready, click the **Text Generation** tab and enter a prompt to get started!
<!-- README_AWQ.md-text-generation-webui end -->
<!-- README_AWQ.md-use-from-vllm start -->
## Multi-user inference server: vLLM
Documentation on installing and using vLLM [can be found here](https://vllm.readthedocs.io/en/latest/).
- Please ensure you are using vLLM version 0.2 or later.
- When using vLLM as a server, pass the `--quantization awq` parameter.
For example:
```shell
python3 -m vllm.entrypoints.api_server --model TheBloke/openchat-3.5-1210-AWQ --quantization awq --dtype auto
```
- When using vLLM from Python code, again set `quantization=awq`.
For example:
```python
from vllm import LLM, SamplingParams
prompts = [
"Tell me about AI",
"Write a story about llamas",
"What is 291 - 150?",
"How much wood would a woodchuck chuck if a woodchuck could chuck wood?",
]
prompt_template=f'''GPT4 Correct User: {prompt}<|end_of_turn|>GPT4 Correct Assistant:
'''
prompts = [prompt_template.format(prompt=prompt) for prompt in prompts]
sampling_params = SamplingParams(temperature=0.8, top_p=0.95)
llm = LLM(model="TheBloke/openchat-3.5-1210-AWQ", quantization="awq", dtype="auto")
outputs = llm.generate(prompts, sampling_params)
# Print the outputs.
for output in outputs:
prompt = output.prompt
generated_text = output.outputs[0].text
print(f"Prompt: {prompt!r}, Generated text: {generated_text!r}")
```
<!-- README_AWQ.md-use-from-vllm start -->
<!-- README_AWQ.md-use-from-tgi start -->
## Multi-user inference server: Hugging Face Text Generation Inference (TGI)
Use TGI version 1.1.0 or later. The official Docker container is: `ghcr.io/huggingface/text-generation-inference:1.1.0`
Example Docker parameters:
```shell
--model-id TheBloke/openchat-3.5-1210-AWQ --port 3000 --quantize awq --max-input-length 3696 --max-total-tokens 4096 --max-batch-prefill-tokens 4096
```
Example Python code for interfacing with TGI (requires [huggingface-hub](https://github.com/huggingface/huggingface_hub) 0.17.0 or later):
```shell
pip3 install huggingface-hub
```
```python
from huggingface_hub import InferenceClient
endpoint_url = "https://your-endpoint-url-here"
prompt = "Tell me about AI"
prompt_template=f'''GPT4 Correct User: {prompt}<|end_of_turn|>GPT4 Correct Assistant:
'''
client = InferenceClient(endpoint_url)
response = client.text_generation(prompt,
max_new_tokens=128,
do_sample=True,
temperature=0.7,
top_p=0.95,
top_k=40,
repetition_penalty=1.1)
print(f"Model output: ", response)
```
<!-- README_AWQ.md-use-from-tgi end -->
<!-- README_AWQ.md-use-from-python start -->
## Inference from Python code using Transformers
### Install the necessary packages
- Requires: [Transformers](https://huggingface.co/docs/transformers) 4.35.0 or later.
- Requires: [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) 0.1.6 or later.
```shell
pip3 install --upgrade "autoawq>=0.1.6" "transformers>=4.35.0"
```
Note that if you are using PyTorch 2.0.1, the above AutoAWQ command will automatically upgrade you to PyTorch 2.1.0.
If you are using CUDA 11.8 and wish to continue using PyTorch 2.0.1, instead run this command:
```shell
pip3 install https://github.com/casper-hansen/AutoAWQ/releases/download/v0.1.6/autoawq-0.1.6+cu118-cp310-cp310-linux_x86_64.whl
```
If you have problems installing [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) using the pre-built wheels, install it from source instead:
```shell
pip3 uninstall -y autoawq
git clone https://github.com/casper-hansen/AutoAWQ
cd AutoAWQ
pip3 install .
```
### Transformers example code (requires Transformers 4.35.0 and later)
```python
from transformers import AutoModelForCausalLM, AutoTokenizer, TextStreamer
model_name_or_path = "TheBloke/openchat-3.5-1210-AWQ"
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path)
model = AutoModelForCausalLM.from_pretrained(
model_name_or_path,
low_cpu_mem_usage=True,
device_map="cuda:0"
)
# Using the text streamer to stream output one token at a time
streamer = TextStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True)
prompt = "Tell me about AI"
prompt_template=f'''GPT4 Correct User: {prompt}<|end_of_turn|>GPT4 Correct Assistant:
'''
# Convert prompt to tokens
tokens = tokenizer(
prompt_template,
return_tensors='pt'
).input_ids.cuda()
generation_params = {
"do_sample": True,
"temperature": 0.7,
"top_p": 0.95,
"top_k": 40,
"max_new_tokens": 512,
"repetition_penalty": 1.1
}
# Generate streamed output, visible one token at a time
generation_output = model.generate(
tokens,
streamer=streamer,
**generation_params
)
# Generation without a streamer, which will include the prompt in the output
generation_output = model.generate(
tokens,
**generation_params
)
# Get the tokens from the output, decode them, print them
token_output = generation_output[0]
text_output = tokenizer.decode(token_output)
print("model.generate output: ", text_output)
# Inference is also possible via Transformers' pipeline
from transformers import pipeline
pipe = pipeline(
"text-generation",
model=model,
tokenizer=tokenizer,
**generation_params
)
pipe_output = pipe(prompt_template)[0]['generated_text']
print("pipeline output: ", pipe_output)
```
<!-- README_AWQ.md-use-from-python end -->
<!-- README_AWQ.md-compatibility start -->
## Compatibility
The files provided are tested to work with:
- [text-generation-webui](https://github.com/oobabooga/text-generation-webui) using `Loader: AutoAWQ`.
- [vLLM](https://github.com/vllm-project/vllm) version 0.2.0 and later.
- [Hugging Face Text Generation Inference (TGI)](https://github.com/huggingface/text-generation-inference) version 1.1.0 and later.
- [Transformers](https://huggingface.co/docs/transformers) version 4.35.0 and later.
- [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) version 0.1.1 and later.
<!-- README_AWQ.md-compatibility end -->
<!-- footer start -->
<!-- 200823 -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute
Thanks to the [chirper.ai](https://chirper.ai) team!
Thanks to Clay from [gpus.llm-utils.org](llm-utils)!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Aemon Algiz.
**Patreon special mentions**: Michael Levine, 阿明, Trailburnt, Nikolai Manek, John Detwiler, Randy H, Will Dee, Sebastain Graf, NimbleBox.ai, Eugene Pentland, Emad Mostaque, Ai Maven, Jim Angel, Jeff Scroggin, Michael Davis, Manuel Alberto Morcote, Stephen Murray, Robert, Justin Joy, Luke @flexchar, Brandon Frisco, Elijah Stavena, S_X, Dan Guido, Undi ., Komninos Chatzipapas, Shadi, theTransient, Lone Striker, Raven Klaugh, jjj, Cap'n Zoog, Michel-Marie MAUDET (LINAGORA), Matthew Berman, David, Fen Risland, Omer Bin Jawed, Luke Pendergrass, Kalila, OG, Erik Bjäreholt, Rooh Singh, Joseph William Delisle, Dan Lewis, TL, John Villwock, AzureBlack, Brad, Pedro Madruga, Caitlyn Gatomon, K, jinyuan sun, Mano Prime, Alex, Jeffrey Morgan, Alicia Loh, Illia Dulskyi, Chadd, transmissions 11, fincy, Rainer Wilmers, ReadyPlayerEmma, knownsqashed, Mandus, biorpg, Deo Leter, Brandon Phillips, SuperWojo, Sean Connelly, Iucharbius, Jack West, Harry Royden McLaughlin, Nicholas, terasurfer, Vitor Caleffi, Duane Dunston, Johann-Peter Hartmann, David Ziegler, Olakabola, Ken Nordquist, Trenton Dambrowitz, Tom X Nguyen, Vadim, Ajan Kanaga, Leonard Tan, Clay Pascal, Alexandros Triantafyllidis, JM33133, Xule, vamX, ya boyyy, subjectnull, Talal Aujan, Alps Aficionado, wassieverse, Ari Malik, James Bentley, Woland, Spencer Kim, Michael Dempsey, Fred von Graf, Elle, zynix, William Richards, Stanislav Ovsiannikov, Edmond Seymore, Jonathan Leane, Martin Kemka, usrbinkat, Enrico Ros
Thank you to all my generous patrons and donaters!
And thank you again to a16z for their generous grant.
<!-- footer end -->
# Original model card: OpenChat's Openchat 3.5 1210
<div align="center">
<img src="https://raw.githubusercontent.com/imoneoi/openchat/master/assets/logo_new.png" style="width: 65%">
<h1>Advancing Open-source Language Models with Mixed-Quality Data</h1>
</div>
<p align="center" style="margin-top: 0px;">
<a href="https://openchat.team">
<img src="https://github.com/alpayariyak/openchat/blob/master/assets/logo_nobg.png?raw=true" alt="OpenChat Logo" style="width:20px; vertical-align: middle; display: inline-block; margin-right: 5px; margin-left: 10px; margin-top: 0px; margin-bottom: 0px;"/>
<span class="link-text" style=" margin-right: 5px;">Online Demo</span>
</a> |
<a href="https://github.com/imoneoi/openchat">
<img src="https://camo.githubusercontent.com/4133dc1cd4511d4a292b84ce10e52e4ed92569fb2a8165381c9c47be5edc2796/68747470733a2f2f6564656e742e6769746875622e696f2f537570657254696e7949636f6e732f696d616765732f706e672f6769746875622e706e67" alt="GitHub Logo" style="width:20px; vertical-align: middle; display: inline-block; margin-right: 5px; margin-left: 5px; margin-top: 0px; margin-bottom: 0px;"/>
<span class="link-text" style=" margin-right: 5px;">GitHub</span>
</a> |
<a href="https://arxiv.org/pdf/2309.11235.pdf">
<img src="https://github.com/alpayariyak/openchat/blob/master/assets/arxiv-logomark-small-square-border.png?raw=true" alt="ArXiv Logo" style="width:20px; vertical-align: middle; display: inline-block; margin-right: 5px; margin-left: 5px; margin-top: 0px; margin-bottom: 0px;"/>
<span class="link-text" style="margin-right: 5px;">Paper</span>
</a> |
<a href="https://discord.gg/pQjnXvNKHY">
<img src="https://cloud.githubusercontent.com/assets/6291467/26705903/96c2d66e-477c-11e7-9f4e-f3c0efe96c9a.png" alt="Discord Logo" style="width:20px; vertical-align: middle; display: inline-block; margin-right: 5px; margin-left: 5px; margin-top: 0px; margin-bottom: 0px;"/>
<span class="link-text">Discord</span>
</a>
</p>
<hr>
<div style="background-color: white; padding: 0.7em; border-radius: 0.5em; color: black; display: flex; flex-direction: column; justify-content: center; text-align: center; ont-size: 0.5em;">
<a href="https://huggingface.co/openchat/openchat_3.5" style="text-decoration: none; color: black;">
<span style="font-size: 1.7em; font-family: 'Helvetica'; letter-spacing: 0.1em; font-weight: bold; color: black;">OPENCHAT</span><span style="font-size: 1.8em; font-family: 'Helvetica'; color: #3c72db; ">3.5</span>
<span style="font-size: 0.7em; font-family: 'Helvetica'; color: white; vertical-align: top; background-color:red; border-radius: 6em; padding: 0.066em 0.4em; letter-spacing: 0.1em; font-weight: bold;">1210</span>
<span style="font-size: 0.85em; font-family: 'Helvetica'; color: black;">
<br> 🏆 The Overall Best Performing Open Source 7B Model 🏆
<br> 🤖 Outperforms <span style="font-weight: bold;">ChatGPT</span> (March) and <span style="font-weight: bold;">Grok-1</span> 🤖
<br> 🚀<span style="font-size: 1em; font-family: 'Helvetica'; color: black; font-weight: bold;">15</span>-point improvement in Coding over <span style="font-size: 0.9em;
font-family: 'Helvetica'; color: black; font-weight: bold;">OpenChat-3.5🚀</span>
<br><br><span style="font-size: 1em; font-family: 'Helvetica'; color: #3c72db; font-weight: bold;">New Features</span>
<br> 💡 2 Modes: Coding + Generalist, Mathematical Reasoning 💡
<br> 🧑⚖️ Experimental support for Evaluator and Feedback capabilities 🧑⚖️
</span>
</a>
</div>
<div style="display: flex; justify-content: center; align-items: center">
<img src="https://github.com/alpayariyak/openchat/blob/master/assets/1210bench.png?raw=true" style="width: 100%; border-radius: 1em">
</div>
<div>
<h3> Table of Contents</h3>
</div>
1. [Usage](#usage)
2. [Benchmarks](#benchmarks)
3. [Limitations](#limitations)
4. [License](#license)
5. [Dataset Details](#dataset-details)
6. [Citation](#citation)
7. [Acknowledgements](#acknowledgements)
<div align="center">
<h2> Usage </h2>
</div>
To use this model, we highly recommend installing the OpenChat package by following the [installation guide](https://github.com/imoneoi/openchat#installation) in our repository and using the OpenChat OpenAI-compatible API server by running the serving command from the table below. The server is optimized for high-throughput deployment using [vLLM](https://github.com/vllm-project/vllm) and can run on a consumer GPU with 24GB RAM. To enable tensor parallelism, append `--tensor-parallel-size N` to the serving command.
Once started, the server listens at `localhost:18888` for requests and is compatible with the [OpenAI ChatCompletion API specifications](https://platform.openai.com/docs/api-reference/chat). Please refer to the example request below for reference. Additionally, you can use the [OpenChat Web UI](https://github.com/imoneoi/openchat#web-ui) for a user-friendly experience.
If you want to deploy the server as an online service, you can use `--api-keys sk-KEY1 sk-KEY2 ...` to specify allowed API keys and `--disable-log-requests --disable-log-stats --log-file openchat.log` for logging only to a file. For security purposes, we recommend using an [HTTPS gateway](https://fastapi.tiangolo.com/es/deployment/concepts/#security-https) in front of the server.
| Model | Size | Context | Weights | Serving |
|-------------------|------|---------|------------------------------------------------------------------|------------------------------------------------------------------------------------------------------------------|
| OpenChat 3.5 1210 | 7B | 8192 | [Huggingface](https://huggingface.co/openchat/openchat_3.5_1210) | `python -m ochat.serving.openai_api_server --model openchat/openchat_3.5_1210 --engine-use-ray --worker-use-ray` |
<details>
<summary>Example request (click to expand)</summary>
💡 **Default Mode (GPT4 Correct)**: Best for coding, chat and general tasks
```bash
curl http://localhost:18888/v1/chat/completions \
-H "Content-Type: application/json" \
-d '{
"model": "openchat_3.5",
"messages": [{"role": "user", "content": "You are a large language model named OpenChat. Write a poem to describe yourself"}]
}'
```
🧮 **Mathematical Reasoning Mode**: Tailored for solving math problems
```bash
curl http://localhost:18888/v1/chat/completions \
-H "Content-Type: application/json" \
-d '{
"model": "openchat_3.5",
"condition": "Math Correct",
"messages": [{"role": "user", "content": "10.3 − 7988.8133 = "}]
}'
```
</details>
### Conversation templates
💡 **Default Mode (GPT4 Correct)**: Best for coding, chat and general tasks
```
GPT4 Correct User: Hello<|end_of_turn|>GPT4 Correct Assistant: Hi<|end_of_turn|>GPT4 Correct User: How are you today?<|end_of_turn|>GPT4 Correct Assistant:
```
🧮 **Mathematical Reasoning Mode**: Tailored for solving math problems
```
Math Correct User: 10.3 − 7988.8133=<|end_of_turn|>Math Correct Assistant:
```
⚠️ **Notice:** Remember to set `<|end_of_turn|>` as end of generation token.
The default (GPT4 Correct) template is also available as the integrated `tokenizer.chat_template`,
which can be used instead of manually specifying the template:
```python
messages = [
{"role": "user", "content": "Hello"},
{"role": "assistant", "content": "Hi"},
{"role": "user", "content": "How are you today?"}
]
tokens = tokenizer.apply_chat_template(messages, add_generation_prompt=True)
assert tokens == [1, 420, 6316, 28781, 3198, 3123, 1247, 28747, 22557, 32000, 420, 6316, 28781, 3198, 3123, 21631, 28747, 15359, 32000, 420, 6316, 28781, 3198, 3123, 1247, 28747, 1602, 460, 368, 3154, 28804, 32000, 420, 6316, 28781, 3198, 3123, 21631, 28747]
```
<div align="center">
<h2> (Experimental) Evaluator / Feedback Capabilities </h2>
</div>
We've included evaluator capabilities in this release to advance open-source models as evaluators. You can use `Default Mode (GPT4 Correct)` with the following prompt (same as [Prometheus](https://huggingface.co/datasets/kaist-ai/Feedback-Collection)) to evaluate a response.
```
###Task Description:
An instruction (might include an Input inside it), a response to evaluate, a reference answer that gets a score of 5, and a score rubric representing a evaluation criteria are given.
1. Write a detailed feedback that assess the quality of the response strictly based on the given score rubric, not evaluating in general.
2. After writing a feedback, write a score that is an integer between 1 and 5. You should refer to the score rubric.
3. The output format should look as follows: "Feedback: (write a feedback for criteria) [RESULT] (an integer number between 1 and 5)"
4. Please do not generate any other opening, closing, and explanations.
###The instruction to evaluate:
{orig_instruction}
###Response to evaluate:
{orig_response}
###Reference Answer (Score 5):
{orig_reference_answer}
###Score Rubrics:
[{orig_criteria}]
Score 1: {orig_score1_description}
Score 2: {orig_score2_description}
Score 3: {orig_score3_description}
Score 4: {orig_score4_description}
Score 5: {orig_score5_description}
###Feedback:
```
<div align="center">
<h2> Benchmarks </h2>
</div>
| Model | # Params | Average | MT-Bench | HumanEval | BBH MC | AGIEval | TruthfulQA | MMLU | GSM8K | BBH CoT |
|--------------------|----------|----------|--------------|-----------------|----------|----------|---------------|--------------|--------------|-------------|
| OpenChat-3.5-1210 | **7B** | **63.8** | 7.76 | **68.9** | **49.5** | **48.0** | **61.8** | 65.3 | **77.3** | 61.8 |
| OpenChat-3.5 | **7B** | 61.6 | 7.81 | 55.5 | 47.6 | 47.4 | 59.1 | 64.3 | **77.3** | 63.5 |
| ChatGPT (March)* | ? | 61.5 | **7.94** | 48.1 | 47.6 | 47.1 | 57.7 | **67.3** | 74.9 | **70.1** |
| | | | | | | | | | | |
| OpenHermes 2.5 | 7B | 59.3 | 7.54 | 48.2 | 49.4 | 46.5 | 57.5 | 63.8 | 73.5 | 59.9 |
| OpenOrca Mistral | 7B | 52.7 | 6.86 | 38.4 | 49.4 | 42.9 | 45.9 | 59.3 | 59.1 | 58.1 |
| Zephyr-β^ | 7B | 34.6 | 7.34 | 22.0 | 40.6 | 39.0 | 40.8 | 39.8 | 5.1 | 16.0 |
| Mistral | 7B | - | 6.84 | 30.5 | 39.0 | 38.0 | - | 60.1 | 52.2 | - |
<details>
<summary>Evaluation Details(click to expand)</summary>
*: ChatGPT (March) results are from [GPT-4 Technical Report](https://arxiv.org/abs/2303.08774), [Chain-of-Thought Hub](https://github.com/FranxYao/chain-of-thought-hub), and our evaluation. Please note that ChatGPT is not a fixed baseline and evolves rapidly over time.
^: Zephyr-β often fails to follow few-shot CoT instructions, likely because it was aligned with only chat data but not trained on few-shot data.
**: Mistral and Open-source SOTA results are taken from reported results in instruction-tuned model papers and official repositories.
All models are evaluated in chat mode (e.g. with the respective conversation template applied). All zero-shot benchmarks follow the same setting as in the AGIEval paper and Orca paper. CoT tasks use the same configuration as Chain-of-Thought Hub, HumanEval is evaluated with EvalPlus, and MT-bench is run using FastChat. To reproduce our results, follow the instructions in [our repository](https://github.com/imoneoi/openchat/#benchmarks).
</details>
<div>
<h3>HumanEval+</h3>
</div>
| Model | Size | HumanEval+ pass@1 |
|-----------------------------|----------|------------|
| ChatGPT (December 12, 2023) | - | 64.6 |
| WizardCoder-Python-34B-V1.0 | 34B | 64.6 |
| **OpenChat 3.5 (Dec 10)** | **7B** | **63.4** |
| OpenHermes 2.5 | 7B | 41.5 |
<div>
<h3>OpenChat-3.5-1210 vs. Grok</h3>
</div>
| | License | # Param | Average | MMLU | HumanEval | MATH | GSM8k |
|-------------------|-------------|---------|----------|------|-----------|----------|----------|
| OpenChat 3.5 1210 | Apache-2.0 | **7B** | **60.1** | 65.3 | **68.9** | **28.9** | **77.3** |
| OpenChat 3.5 | Apache-2.0 | **7B** | 56.4 | 64.3 | 55.5 | 28.6 | **77.3** |
| Grok-0 | Proprietary | 33B | 44.5 | 65.7 | 39.7 | 15.7 | 56.8 |
| Grok-1 | Proprietary | ???B | 55.8 | 73 | 63.2 | 23.9 | 62.9 |
*: Grok results are reported by [X.AI](https://x.ai/).
<div align="center">
<h2> 中文评估结果 / Chinese Evaluations </h2>
</div>
⚠️ Note that this model was not explicitly trained in Chinese (only < 0.1% of the data is in Chinese). 请注意本模型没有针对性训练中文（中文数据占比小于0.1%）。
<div>
<h3>Multi-Level Multi-Discipline Chinese Evaluation Suite (CEVAL)</h3>
<div>
| Model | Avg | STEM | Social Science | Humanities | Others |
|----------|-------|-------|----------------|------------|--------|
| ChatGPT | 54.4 | 52.9 | 61.8 | 50.9 | 53.6 |
| OpenChat | 47.29 | 45.22 | 52.49 | 48.52 | 45.08 |
<div>
<h3>Massive Multitask Language Understanding in Chinese (CMMLU, 5-shot)</h3>
</div>
| Models | STEM | Humanities | SocialSciences | Other | ChinaSpecific | Avg |
|----------|-------|------------|----------------|-------|---------------|-------|
| ChatGPT | 47.81 | 55.68 | 56.5 | 62.66 | 50.69 | 55.51 |
| OpenChat | 38.7 | 45.99 | 48.32 | 50.23 | 43.27 | 45.85 |
<div align="center">
<h2> Limitations </h2>
</div>
**Foundation Model Limitations**
Despite its advanced capabilities, OpenChat is still bound by the limitations inherent in its foundation models. These limitations may impact the model's performance in areas such as:
- Complex reasoning
- Mathematical and arithmetic tasks
- Programming and coding challenges
**Hallucination of Non-existent Information**
OpenChat may sometimes generate information that does not exist or is not accurate, also known as "hallucination". Users should be aware of this possibility and verify any critical information obtained from the model.
**Safety**
OpenChat may sometimes generate harmful, hate speech, biased responses, or answer unsafe questions. It's crucial to apply additional AI safety measures in use cases that require safe and moderated responses.
<div align="center">
<h2> License </h2>
</div>
Our OpenChat 3.5 code and models are distributed under the Apache License 2.0.
<div align="center">
<h2> Dataset Details </h2>
</div>
OpenChat 3.5 was trained with C-RLFT on a collection of publicly available high-quality instruction data, with a custom processing pipeline. We detail some notable subsets included here:
- [OpenChat ShareGPT](https://huggingface.co/datasets/openchat/openchat_sharegpt4_dataset)
- [Open-Orca with FLAN answers](https://huggingface.co/datasets/imone/OpenOrca_FLAN)
- [Feedback-Collection](https://huggingface.co/datasets/kaist-ai/Feedback-Collection)
- Capybara [1](https://huggingface.co/datasets/LDJnr/Pure-Dove) [2](https://huggingface.co/datasets/LDJnr/Verified-Camel) [3](https://huggingface.co/datasets/LDJnr/LessWrong-Amplify-Instruct)
- [GOAT](https://huggingface.co/datasets/tiedong/goat)
- [Glaive](https://huggingface.co/datasets/glaiveai/glaive-code-assistant)
- [MetaMathQA](https://huggingface.co/datasets/meta-math/MetaMathQA)
- [MathInstruct](https://huggingface.co/datasets/TIGER-Lab/MathInstruct)
- [OpenAssistant](https://huggingface.co/datasets/OpenAssistant/oasst_top1_2023-08-25)
<div align="center">
<h2> Citation </h2>
</div>
```
@article{wang2023openchat,
title={OpenChat: Advancing Open-source Language Models with Mixed-Quality Data},
author={Wang, Guan and Cheng, Sijie and Zhan, Xianyuan and Li, Xiangang and Song, Sen and Liu, Yang},
journal={arXiv preprint arXiv:2309.11235},
year={2023}
}
```
<div align="center">
<h2> Acknowledgments </h2>
</div>
We extend our heartfelt gratitude to AutoMeta and caesus from Alignment Lab AI, LDJ and Teknium from Nous Research, alpin and TearGosling from Pygmalion AI for their substantial contributions to data collection and model training.
Special thanks go to Changling Liu from GPT Desk Pte. Ltd., Qiying Yu at Tsinghua University, Baochang Ma, and Hao Wan from 01.AI company for their generous provision of resources. We are also deeply grateful to Jianxiong Li and Peng Li at Tsinghua University for their insightful discussions.
Furthermore, we appreciate the developers behind the following projects for their significant contributions to our research: [Mistral](https://mistral.ai/), [Chain-of-Thought Hub](https://github.com/FranxYao/chain-of-thought-hub), [Llama 2](https://ai.meta.com/llama/), [Self-Instruct](https://arxiv.org/abs/2212.10560), [FastChat (Vicuna)](https://github.com/lm-sys/FastChat), [Alpaca](https://github.com/tatsu-lab/stanford_alpaca.git), and [StarCoder](https://github.com/bigcode-project/starcoder). Their work has been instrumental in driving our research forward.
|
mnoukhov/pythia410m-tldr-sft
|
mnoukhov
| 2023-12-19T21:39:38Z | 16 | 0 |
transformers
|
[
"transformers",
"pytorch",
"safetensors",
"gpt_neox",
"text-generation",
"base_model:mnoukhov/pythia410m-tldr-sft",
"base_model:finetune:mnoukhov/pythia410m-tldr-sft",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2023-10-11T23:20:53Z |
---
base_model: mnoukhov/pythia410m-tldr-sft
---
pythia410m finetuned on the openai_summarize_tldr dataset
|
TheBloke/Swallow-13B-GGUF
|
TheBloke
| 2023-12-19T21:39:27Z | 221 | 4 |
transformers
|
[
"transformers",
"gguf",
"llama",
"text-generation",
"en",
"ja",
"base_model:tokyotech-llm/Swallow-13b-hf",
"base_model:quantized:tokyotech-llm/Swallow-13b-hf",
"license:llama2",
"region:us"
] |
text-generation
| 2023-12-19T21:32:03Z |
---
base_model: tokyotech-llm/Swallow-13b-hf
inference: false
language:
- en
- ja
library_name: transformers
license: llama2
model_creator: tokyotech-llm
model_name: Swallow 13B
model_type: llama
pipeline_tag: text-generation
prompt_template: '{prompt}
'
quantized_by: TheBloke
---
<!-- markdownlint-disable MD041 -->
<!-- header start -->
<!-- 200823 -->
<div style="width: auto; margin-left: auto; margin-right: auto">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<div style="text-align:center; margin-top: 0em; margin-bottom: 0em"><p style="margin-top: 0.25em; margin-bottom: 0em;">TheBloke's LLM work is generously supported by a grant from <a href="https://a16z.com">andreessen horowitz (a16z)</a></p></div>
<hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
<!-- header end -->
# Swallow 13B - GGUF
- Model creator: [tokyotech-llm](https://huggingface.co/tokyotech-llm)
- Original model: [Swallow 13B](https://huggingface.co/tokyotech-llm/Swallow-13b-hf)
<!-- description start -->
## Description
This repo contains GGUF format model files for [tokyotech-llm's Swallow 13B](https://huggingface.co/tokyotech-llm/Swallow-13b-hf).
These files were quantised using hardware kindly provided by [Massed Compute](https://massedcompute.com/).
<!-- description end -->
<!-- README_GGUF.md-about-gguf start -->
### About GGUF
GGUF is a new format introduced by the llama.cpp team on August 21st 2023. It is a replacement for GGML, which is no longer supported by llama.cpp.
Here is an incomplete list of clients and libraries that are known to support GGUF:
* [llama.cpp](https://github.com/ggerganov/llama.cpp). The source project for GGUF. Offers a CLI and a server option.
* [text-generation-webui](https://github.com/oobabooga/text-generation-webui), the most widely used web UI, with many features and powerful extensions. Supports GPU acceleration.
* [KoboldCpp](https://github.com/LostRuins/koboldcpp), a fully featured web UI, with GPU accel across all platforms and GPU architectures. Especially good for story telling.
* [GPT4All](https://gpt4all.io/index.html), a free and open source local running GUI, supporting Windows, Linux and macOS with full GPU accel.
* [LM Studio](https://lmstudio.ai/), an easy-to-use and powerful local GUI for Windows and macOS (Silicon), with GPU acceleration. Linux available, in beta as of 27/11/2023.
* [LoLLMS Web UI](https://github.com/ParisNeo/lollms-webui), a great web UI with many interesting and unique features, including a full model library for easy model selection.
* [Faraday.dev](https://faraday.dev/), an attractive and easy to use character-based chat GUI for Windows and macOS (both Silicon and Intel), with GPU acceleration.
* [llama-cpp-python](https://github.com/abetlen/llama-cpp-python), a Python library with GPU accel, LangChain support, and OpenAI-compatible API server.
* [candle](https://github.com/huggingface/candle), a Rust ML framework with a focus on performance, including GPU support, and ease of use.
* [ctransformers](https://github.com/marella/ctransformers), a Python library with GPU accel, LangChain support, and OpenAI-compatible AI server. Note, as of time of writing (November 27th 2023), ctransformers has not been updated in a long time and does not support many recent models.
<!-- README_GGUF.md-about-gguf end -->
<!-- repositories-available start -->
## Repositories available
* [AWQ model(s) for GPU inference.](https://huggingface.co/TheBloke/Swallow-13B-AWQ)
* [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/Swallow-13B-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/Swallow-13B-GGUF)
* [tokyotech-llm's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/tokyotech-llm/Swallow-13b-hf)
<!-- repositories-available end -->
<!-- prompt-template start -->
## Prompt template: None
```
{prompt}
```
<!-- prompt-template end -->
<!-- compatibility_gguf start -->
## Compatibility
These quantised GGUFv2 files are compatible with llama.cpp from August 27th onwards, as of commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221)
They are also compatible with many third party UIs and libraries - please see the list at the top of this README.
## Explanation of quantisation methods
<details>
<summary>Click to see details</summary>
The new methods available are:
* GGML_TYPE_Q2_K - "type-1" 2-bit quantization in super-blocks containing 16 blocks, each block having 16 weight. Block scales and mins are quantized with 4 bits. This ends up effectively using 2.5625 bits per weight (bpw)
* GGML_TYPE_Q3_K - "type-0" 3-bit quantization in super-blocks containing 16 blocks, each block having 16 weights. Scales are quantized with 6 bits. This end up using 3.4375 bpw.
* GGML_TYPE_Q4_K - "type-1" 4-bit quantization in super-blocks containing 8 blocks, each block having 32 weights. Scales and mins are quantized with 6 bits. This ends up using 4.5 bpw.
* GGML_TYPE_Q5_K - "type-1" 5-bit quantization. Same super-block structure as GGML_TYPE_Q4_K resulting in 5.5 bpw
* GGML_TYPE_Q6_K - "type-0" 6-bit quantization. Super-blocks with 16 blocks, each block having 16 weights. Scales are quantized with 8 bits. This ends up using 6.5625 bpw
Refer to the Provided Files table below to see what files use which methods, and how.
</details>
<!-- compatibility_gguf end -->
<!-- README_GGUF.md-provided-files start -->
## Provided files
| Name | Quant method | Bits | Size | Max RAM required | Use case |
| ---- | ---- | ---- | ---- | ---- | ----- |
| [swallow-13b.Q2_K.gguf](https://huggingface.co/TheBloke/Swallow-13B-GGUF/blob/main/swallow-13b.Q2_K.gguf) | Q2_K | 2 | 5.50 GB| 8.00 GB | smallest, significant quality loss - not recommended for most purposes |
| [swallow-13b.Q3_K_S.gguf](https://huggingface.co/TheBloke/Swallow-13B-GGUF/blob/main/swallow-13b.Q3_K_S.gguf) | Q3_K_S | 3 | 5.73 GB| 8.23 GB | very small, high quality loss |
| [swallow-13b.Q3_K_M.gguf](https://huggingface.co/TheBloke/Swallow-13B-GGUF/blob/main/swallow-13b.Q3_K_M.gguf) | Q3_K_M | 3 | 6.41 GB| 8.91 GB | very small, high quality loss |
| [swallow-13b.Q3_K_L.gguf](https://huggingface.co/TheBloke/Swallow-13B-GGUF/blob/main/swallow-13b.Q3_K_L.gguf) | Q3_K_L | 3 | 7.00 GB| 9.50 GB | small, substantial quality loss |
| [swallow-13b.Q4_0.gguf](https://huggingface.co/TheBloke/Swallow-13B-GGUF/blob/main/swallow-13b.Q4_0.gguf) | Q4_0 | 4 | 7.45 GB| 9.95 GB | legacy; small, very high quality loss - prefer using Q3_K_M |
| [swallow-13b.Q4_K_S.gguf](https://huggingface.co/TheBloke/Swallow-13B-GGUF/blob/main/swallow-13b.Q4_K_S.gguf) | Q4_K_S | 4 | 7.49 GB| 9.99 GB | small, greater quality loss |
| [swallow-13b.Q4_K_M.gguf](https://huggingface.co/TheBloke/Swallow-13B-GGUF/blob/main/swallow-13b.Q4_K_M.gguf) | Q4_K_M | 4 | 7.95 GB| 10.45 GB | medium, balanced quality - recommended |
| [swallow-13b.Q5_0.gguf](https://huggingface.co/TheBloke/Swallow-13B-GGUF/blob/main/swallow-13b.Q5_0.gguf) | Q5_0 | 5 | 9.06 GB| 11.56 GB | legacy; medium, balanced quality - prefer using Q4_K_M |
| [swallow-13b.Q5_K_S.gguf](https://huggingface.co/TheBloke/Swallow-13B-GGUF/blob/main/swallow-13b.Q5_K_S.gguf) | Q5_K_S | 5 | 9.06 GB| 11.56 GB | large, low quality loss - recommended |
| [swallow-13b.Q5_K_M.gguf](https://huggingface.co/TheBloke/Swallow-13B-GGUF/blob/main/swallow-13b.Q5_K_M.gguf) | Q5_K_M | 5 | 9.32 GB| 11.82 GB | large, very low quality loss - recommended |
| [swallow-13b.Q6_K.gguf](https://huggingface.co/TheBloke/Swallow-13B-GGUF/blob/main/swallow-13b.Q6_K.gguf) | Q6_K | 6 | 10.77 GB| 13.27 GB | very large, extremely low quality loss |
| [swallow-13b.Q8_0.gguf](https://huggingface.co/TheBloke/Swallow-13B-GGUF/blob/main/swallow-13b.Q8_0.gguf) | Q8_0 | 8 | 13.95 GB| 16.45 GB | very large, extremely low quality loss - not recommended |
**Note**: the above RAM figures assume no GPU offloading. If layers are offloaded to the GPU, this will reduce RAM usage and use VRAM instead.
<!-- README_GGUF.md-provided-files end -->
<!-- README_GGUF.md-how-to-download start -->
## How to download GGUF files
**Note for manual downloaders:** You almost never want to clone the entire repo! Multiple different quantisation formats are provided, and most users only want to pick and download a single file.
The following clients/libraries will automatically download models for you, providing a list of available models to choose from:
* LM Studio
* LoLLMS Web UI
* Faraday.dev
### In `text-generation-webui`
Under Download Model, you can enter the model repo: TheBloke/Swallow-13B-GGUF and below it, a specific filename to download, such as: swallow-13b.Q4_K_M.gguf.
Then click Download.
### On the command line, including multiple files at once
I recommend using the `huggingface-hub` Python library:
```shell
pip3 install huggingface-hub
```
Then you can download any individual model file to the current directory, at high speed, with a command like this:
```shell
huggingface-cli download TheBloke/Swallow-13B-GGUF swallow-13b.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False
```
<details>
<summary>More advanced huggingface-cli download usage (click to read)</summary>
You can also download multiple files at once with a pattern:
```shell
huggingface-cli download TheBloke/Swallow-13B-GGUF --local-dir . --local-dir-use-symlinks False --include='*Q4_K*gguf'
```
For more documentation on downloading with `huggingface-cli`, please see: [HF -> Hub Python Library -> Download files -> Download from the CLI](https://huggingface.co/docs/huggingface_hub/guides/download#download-from-the-cli).
To accelerate downloads on fast connections (1Gbit/s or higher), install `hf_transfer`:
```shell
pip3 install hf_transfer
```
And set environment variable `HF_HUB_ENABLE_HF_TRANSFER` to `1`:
```shell
HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download TheBloke/Swallow-13B-GGUF swallow-13b.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False
```
Windows Command Line users: You can set the environment variable by running `set HF_HUB_ENABLE_HF_TRANSFER=1` before the download command.
</details>
<!-- README_GGUF.md-how-to-download end -->
<!-- README_GGUF.md-how-to-run start -->
## Example `llama.cpp` command
Make sure you are using `llama.cpp` from commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221) or later.
```shell
./main -ngl 35 -m swallow-13b.Q4_K_M.gguf --color -c 4096 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "{prompt}"
```
Change `-ngl 32` to the number of layers to offload to GPU. Remove it if you don't have GPU acceleration.
Change `-c 4096` to the desired sequence length. For extended sequence models - eg 8K, 16K, 32K - the necessary RoPE scaling parameters are read from the GGUF file and set by llama.cpp automatically. Note that longer sequence lengths require much more resources, so you may need to reduce this value.
If you want to have a chat-style conversation, replace the `-p <PROMPT>` argument with `-i -ins`
For other parameters and how to use them, please refer to [the llama.cpp documentation](https://github.com/ggerganov/llama.cpp/blob/master/examples/main/README.md)
## How to run in `text-generation-webui`
Further instructions can be found in the text-generation-webui documentation, here: [text-generation-webui/docs/04 ‐ Model Tab.md](https://github.com/oobabooga/text-generation-webui/blob/main/docs/04%20%E2%80%90%20Model%20Tab.md#llamacpp).
## How to run from Python code
You can use GGUF models from Python using the [llama-cpp-python](https://github.com/abetlen/llama-cpp-python) or [ctransformers](https://github.com/marella/ctransformers) libraries. Note that at the time of writing (Nov 27th 2023), ctransformers has not been updated for some time and is not compatible with some recent models. Therefore I recommend you use llama-cpp-python.
### How to load this model in Python code, using llama-cpp-python
For full documentation, please see: [llama-cpp-python docs](https://abetlen.github.io/llama-cpp-python/).
#### First install the package
Run one of the following commands, according to your system:
```shell
# Base ctransformers with no GPU acceleration
pip install llama-cpp-python
# With NVidia CUDA acceleration
CMAKE_ARGS="-DLLAMA_CUBLAS=on" pip install llama-cpp-python
# Or with OpenBLAS acceleration
CMAKE_ARGS="-DLLAMA_BLAS=ON -DLLAMA_BLAS_VENDOR=OpenBLAS" pip install llama-cpp-python
# Or with CLBLast acceleration
CMAKE_ARGS="-DLLAMA_CLBLAST=on" pip install llama-cpp-python
# Or with AMD ROCm GPU acceleration (Linux only)
CMAKE_ARGS="-DLLAMA_HIPBLAS=on" pip install llama-cpp-python
# Or with Metal GPU acceleration for macOS systems only
CMAKE_ARGS="-DLLAMA_METAL=on" pip install llama-cpp-python
# In windows, to set the variables CMAKE_ARGS in PowerShell, follow this format; eg for NVidia CUDA:
$env:CMAKE_ARGS = "-DLLAMA_OPENBLAS=on"
pip install llama-cpp-python
```
#### Simple llama-cpp-python example code
```python
from llama_cpp import Llama
# Set gpu_layers to the number of layers to offload to GPU. Set to 0 if no GPU acceleration is available on your system.
llm = Llama(
model_path="./swallow-13b.Q4_K_M.gguf", # Download the model file first
n_ctx=4096, # The max sequence length to use - note that longer sequence lengths require much more resources
n_threads=8, # The number of CPU threads to use, tailor to your system and the resulting performance
n_gpu_layers=35 # The number of layers to offload to GPU, if you have GPU acceleration available
)
# Simple inference example
output = llm(
"{prompt}", # Prompt
max_tokens=512, # Generate up to 512 tokens
stop=["</s>"], # Example stop token - not necessarily correct for this specific model! Please check before using.
echo=True # Whether to echo the prompt
)
# Chat Completion API
llm = Llama(model_path="./swallow-13b.Q4_K_M.gguf", chat_format="llama-2") # Set chat_format according to the model you are using
llm.create_chat_completion(
messages = [
{"role": "system", "content": "You are a story writing assistant."},
{
"role": "user",
"content": "Write a story about llamas."
}
]
)
```
## How to use with LangChain
Here are guides on using llama-cpp-python and ctransformers with LangChain:
* [LangChain + llama-cpp-python](https://python.langchain.com/docs/integrations/llms/llamacpp)
* [LangChain + ctransformers](https://python.langchain.com/docs/integrations/providers/ctransformers)
<!-- README_GGUF.md-how-to-run end -->
<!-- footer start -->
<!-- 200823 -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute
Thanks to the [chirper.ai](https://chirper.ai) team!
Thanks to Clay from [gpus.llm-utils.org](llm-utils)!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Aemon Algiz.
**Patreon special mentions**: Michael Levine, 阿明, Trailburnt, Nikolai Manek, John Detwiler, Randy H, Will Dee, Sebastain Graf, NimbleBox.ai, Eugene Pentland, Emad Mostaque, Ai Maven, Jim Angel, Jeff Scroggin, Michael Davis, Manuel Alberto Morcote, Stephen Murray, Robert, Justin Joy, Luke @flexchar, Brandon Frisco, Elijah Stavena, S_X, Dan Guido, Undi ., Komninos Chatzipapas, Shadi, theTransient, Lone Striker, Raven Klaugh, jjj, Cap'n Zoog, Michel-Marie MAUDET (LINAGORA), Matthew Berman, David, Fen Risland, Omer Bin Jawed, Luke Pendergrass, Kalila, OG, Erik Bjäreholt, Rooh Singh, Joseph William Delisle, Dan Lewis, TL, John Villwock, AzureBlack, Brad, Pedro Madruga, Caitlyn Gatomon, K, jinyuan sun, Mano Prime, Alex, Jeffrey Morgan, Alicia Loh, Illia Dulskyi, Chadd, transmissions 11, fincy, Rainer Wilmers, ReadyPlayerEmma, knownsqashed, Mandus, biorpg, Deo Leter, Brandon Phillips, SuperWojo, Sean Connelly, Iucharbius, Jack West, Harry Royden McLaughlin, Nicholas, terasurfer, Vitor Caleffi, Duane Dunston, Johann-Peter Hartmann, David Ziegler, Olakabola, Ken Nordquist, Trenton Dambrowitz, Tom X Nguyen, Vadim, Ajan Kanaga, Leonard Tan, Clay Pascal, Alexandros Triantafyllidis, JM33133, Xule, vamX, ya boyyy, subjectnull, Talal Aujan, Alps Aficionado, wassieverse, Ari Malik, James Bentley, Woland, Spencer Kim, Michael Dempsey, Fred von Graf, Elle, zynix, William Richards, Stanislav Ovsiannikov, Edmond Seymore, Jonathan Leane, Martin Kemka, usrbinkat, Enrico Ros
Thank you to all my generous patrons and donaters!
And thank you again to a16z for their generous grant.
<!-- footer end -->
<!-- original-model-card start -->
# Original model card: tokyotech-llm's Swallow 13B
# Swallow
Our Swallow model has undergone continuous pre-training from the Llama 2 family, primarily with the addition of Japanese language data. The tuned versions use supervised fine-tuning (SFT).
Links to other models can be found in the index.
## Swallow Model Index
|Model|Swallow-hf|Swallow-instruct-hf|
|---|---|---|
|7B| [Link](https://huggingface.co/tokyotech-llm/Swallow-7b-hf) | [Link](https://huggingface.co/tokyotech-llm/Swallow-7b-instruct-hf)|
|13B| [Link](https://huggingface.co/tokyotech-llm/Swallow-13b-hf) | [Link](https://huggingface.co/tokyotech-llm/Swallow-13b-instruct-hf)|
|70B| [Link](https://huggingface.co/tokyotech-llm/Swallow-70b-hf) | [Link](https://huggingface.co/tokyotech-llm/Swallow-70b-instruct-hf)|
This repository provides large language models developed by [TokyoTech-LLM](https://tokyotech-llm.github.io/).
Read our [blog post](https://zenn.dev/tokyotech_lm/articles/d6cb3a8fdfc907) or our paper (preprint coming soon) for more details!
## Model Details
* **Model type**: Please refer to LLaMA-2 technical report for details on the model architecture.
* **Language(s)**: Japanese English
* **Library**: [Megatron-LM](https://github.com/rioyokotalab/Megatron-Llama2)
* **Tokenizer**: This model employs a tokenizer that features a broadened vocabulary based on Japanese data. This allows for a more efficient representation of text using fewer tokens, leading to a notably faster inference process.
* **Contact**: swallow[at]nlp.c.titech.ac.jp
## Base Model Performance
### Japanese version
|Model|Size|JCommonsenseQA|JEMHopQA|NIILC|JSQuAD|XL-Sum|MGSM|WMT20-en-ja|WMT20-ja-en|
|---|---|---|---|---|---|---|---|---|---|
| | |4-shot|4-shot|4-shot|4-shot|1-shot|4-shot|4-shot|4-shot|
|Llama 2|7B|0.3852|0.4240|0.3410|0.7917|0.1905|0.0760|0.1783|0.1738|
|Swallow|7B|0.4808|0.5078|0.5968|0.8573|0.1830|0.1240|0.2510|0.1511|
|Llama 2|13B|0.6997|0.4415|0.4170|0.8533|0.2139|0.1320|0.2146|0.1982|
|Swallow|13B|0.7837|0.5063|0.6398|0.9005|0.2168|0.2040|0.2720|0.1771|
|Llama 2|70B|0.8686|0.4656|0.5256|0.9080|**0.2361**|0.3560|0.2643|**0.2398**|
|Swallow|70B|**0.9348**|**0.6290**|**0.6960**|**0.9176**|0.2266|**0.4840**|**0.3043**|0.2298|
## Usage
First install additional dependencies in [requirements.txt](./requirements.txt):
```sh
pip install -r requirements.txt
```
### Use the instruct model
```python
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
model_name = "tokyotech-llm/Swallow-7b-instruct-hf"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.bfloat16, low_cpu_mem_usage=True, device_map="auto")
PROMPT_DICT = {
"prompt_input": (
"以下に、あるタスクを説明する指示があり、それに付随する入力が更なる文脈を提供しています。"
"リクエストを適切に完了するための回答を記述してください。\n\n"
"### 指示:\n{instruction}\n\n### 入力:\n{input}\n\n### 応答:"
),
"prompt_no_input": (
"以下に、あるタスクを説明する指示があります。"
"リクエストを適切に完了するための回答を記述してください。\n\n"
"### 指示:\n{instruction}\n\n### 応答:"
),
}
def create_prompt(instruction, input=None):
"""
Generates a prompt based on the given instruction and an optional input.
If input is provided, it uses the 'prompt_input' template from PROMPT_DICT.
If no input is provided, it uses the 'prompt_no_input' template.
Args:
instruction (str): The instruction describing the task.
input (str, optional): Additional input providing context for the task. Default is None.
Returns:
str: The generated prompt.
"""
if input:
# Use the 'prompt_input' template when additional input is provided
return PROMPT_DICT["prompt_input"].format(instruction=instruction, input=input)
else:
# Use the 'prompt_no_input' template when no additional input is provided
return PROMPT_DICT["prompt_no_input"].format(instruction=instruction)
# Example usage
instruction_example = "以下のトピックに関する詳細な情報を提供してください。"
input_example = "東京工業大学の主なキャンパスについて教えてください"
prompt = create_prompt(instruction_example, input_example)
input_ids = tokenizer.encode(
prompt,
add_special_tokens=False,
return_tensors="pt"
)
tokens = model.generate(
input_ids.to(device=model.device),
max_new_tokens=128,
temperature=0.99,
top_p=0.95,
do_sample=True,
)
out = tokenizer.decode(tokens[0], skip_special_tokens=True)
print(out)
```
### Use the base model
```python
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
model_name = "tokyotech-llm/Swallow-7b-hf"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.bfloat16, device_map="auto")
prompt = "東京工業大学の主なキャンパスは、"
input_ids = tokenizer.encode(
prompt,
add_special_tokens=False,
return_tensors="pt"
)
tokens = model.generate(
input_ids.to(device=model.device),
max_new_tokens=128,
temperature=0.99,
top_p=0.95,
do_sample=True,
)
out = tokenizer.decode(tokens[0], skip_special_tokens=True)
print(out)
```
## Training Datasets
### Continual Pre-Training
The following datasets were used for continual pre-training.
- [Japanese Wikipedia](https://dumps.wikimedia.org/other/cirrussearch)
- [RefinedWeb](https://huggingface.co/datasets/tiiuae/falcon-refinedweb)
- Swallow Corpus
- [The Pile](https://huggingface.co/datasets/EleutherAI/pile)
### Instruction Tuning
The following datasets were used for the instruction tuning.
- [Anthropic HH-RLHF](https://huggingface.co/datasets/kunishou/hh-rlhf-49k-ja)
- [Databricks Dolly 15-k](https://huggingface.co/datasets/kunishou/databricks-dolly-15k-ja)
- [OpenAssistant Conversations Dataset](https://huggingface.co/datasets/kunishou/oasst1-89k-ja)
## Risks and Limitations
The models released here are still in the early stages of our research and development and have not been tuned to ensure outputs align with human intent and safety considerations.
## Acknowledgements
We thank Meta Research for releasing Llama 2 under an open license for others to build on.
Our project is supported by the [ABCI Large-scale Language Model Building Support Program](https://abci.ai/en/link/llm_support_program.html) of the National Institute of Advanced Industrial Science and Technology.
## License
Llama 2 is licensed under the LLAMA 2 Community License, Copyright © Meta Platforms, Inc. All Rights Reserved.
## Authors
Here are the team members:
- From [Okazaki Laboratory](https://www.nlp.c.titech.ac.jp/index.en.html), the following members:
- [Naoaki Okazaki](https://www.chokkan.org/index.ja.html)
- [Sakae Mizuki](https://s-mizuki-nlp.github.io/)
- [Hiroki Iida](https://meshidenn.github.io/)
- [Mengsay Loem](https://loem-ms.github.io/)
- [Shota Hirai](https://huggingface.co/Kotemo428)
- [Kakeru Hattori](https://aya-se.vercel.app/)
- [Masanari Ohi](https://twitter.com/stjohn2007)
- From [YOKOTA Laboratory](https://www.rio.gsic.titech.ac.jp/en/index.html), the following members:
- [Rio Yokota](https://twitter.com/rioyokota)
- [Kazuki Fujii](https://twitter.com/okoge_kaz)
- [Taishi Nakamura](https://twitter.com/Setuna7777_2)
<!-- original-model-card end -->
|
intermezzo672/NHS-pubmedbert
|
intermezzo672
| 2023-12-19T21:37:11Z | 6 | 1 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"bert",
"text-classification",
"generated_from_trainer",
"base_model:microsoft/BiomedNLP-BiomedBERT-base-uncased-abstract",
"base_model:finetune:microsoft/BiomedNLP-BiomedBERT-base-uncased-abstract",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2023-12-19T20:32:25Z |
---
license: mit
base_model: microsoft/BiomedNLP-PubMedBERT-base-uncased-abstract
tags:
- generated_from_trainer
metrics:
- accuracy
- precision
- recall
- f1
model-index:
- name: NHS-pubmedbert
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# NHS-pubmedbert
This model is a fine-tuned version of [microsoft/BiomedNLP-PubMedBERT-base-uncased-abstract](https://huggingface.co/microsoft/BiomedNLP-PubMedBERT-base-uncased-abstract) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.6667
- Accuracy: 0.8177
- Precision: 0.8190
- Recall: 0.8177
- F1: 0.8143
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 3e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 6
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:---------:|:------:|:------:|
| 0.0827 | 1.0 | 397 | 0.4385 | 0.7994 | 0.8128 | 0.7994 | 0.8011 |
| 0.0149 | 2.0 | 794 | 0.4484 | 0.8227 | 0.8232 | 0.8227 | 0.8229 |
| 0.0027 | 3.0 | 1191 | 0.6667 | 0.8177 | 0.8190 | 0.8177 | 0.8143 |
### Framework versions
- Transformers 4.35.2
- Pytorch 2.1.0+cu121
- Datasets 2.15.0
- Tokenizers 0.15.0
|
jbeomlee93/output_lr3e-6_datav3_modify_size1024_step1000_zoomin_textprompt_v2
|
jbeomlee93
| 2023-12-19T21:36:26Z | 1 | 0 |
diffusers
|
[
"diffusers",
"tensorboard",
"safetensors",
"stable-diffusion-xl",
"stable-diffusion-xl-diffusers",
"text-to-image",
"controlnet",
"base_model:diffusers/stable-diffusion-xl-1.0-inpainting-0.1",
"base_model:adapter:diffusers/stable-diffusion-xl-1.0-inpainting-0.1",
"license:openrail++",
"region:us"
] |
text-to-image
| 2023-12-19T19:43:14Z |
---
license: openrail++
base_model: diffusers/stable-diffusion-xl-1.0-inpainting-0.1
tags:
- stable-diffusion-xl
- stable-diffusion-xl-diffusers
- text-to-image
- diffusers
- controlnet
inference: true
---
# controlnet-jbeomlee93/output_lr3e-6_datav3_modify_size1024_step1000_zoomin_textprompt_v2
These are controlnet weights trained on diffusers/stable-diffusion-xl-1.0-inpainting-0.1 with new type of conditioning.
|
jbeomlee93/output_lr1e-6_datav3_modify_size1024_step1000_zoomin_textprompt_v2
|
jbeomlee93
| 2023-12-19T21:36:03Z | 1 | 0 |
diffusers
|
[
"diffusers",
"tensorboard",
"safetensors",
"stable-diffusion-xl",
"stable-diffusion-xl-diffusers",
"text-to-image",
"controlnet",
"base_model:diffusers/stable-diffusion-xl-1.0-inpainting-0.1",
"base_model:adapter:diffusers/stable-diffusion-xl-1.0-inpainting-0.1",
"license:openrail++",
"region:us"
] |
text-to-image
| 2023-12-19T19:42:25Z |
---
license: openrail++
base_model: diffusers/stable-diffusion-xl-1.0-inpainting-0.1
tags:
- stable-diffusion-xl
- stable-diffusion-xl-diffusers
- text-to-image
- diffusers
- controlnet
inference: true
---
# controlnet-jbeomlee93/output_lr1e-6_datav3_modify_size1024_step1000_zoomin_textprompt_v2
These are controlnet weights trained on diffusers/stable-diffusion-xl-1.0-inpainting-0.1 with new type of conditioning.
|
jbeomlee93/output_lr1e-5_datav3_modify_size1024_step1000_zoomin_textprompt_v2
|
jbeomlee93
| 2023-12-19T21:33:24Z | 1 | 0 |
diffusers
|
[
"diffusers",
"tensorboard",
"safetensors",
"stable-diffusion-xl",
"stable-diffusion-xl-diffusers",
"text-to-image",
"controlnet",
"base_model:diffusers/stable-diffusion-xl-1.0-inpainting-0.1",
"base_model:adapter:diffusers/stable-diffusion-xl-1.0-inpainting-0.1",
"license:openrail++",
"region:us"
] |
text-to-image
| 2023-12-19T19:39:06Z |
---
license: openrail++
base_model: diffusers/stable-diffusion-xl-1.0-inpainting-0.1
tags:
- stable-diffusion-xl
- stable-diffusion-xl-diffusers
- text-to-image
- diffusers
- controlnet
inference: true
---
# controlnet-jbeomlee93/output_lr1e-5_datav3_modify_size1024_step1000_zoomin_textprompt_v2
These are controlnet weights trained on diffusers/stable-diffusion-xl-1.0-inpainting-0.1 with new type of conditioning.
|
Ja-le/whisper-tiny-hi
|
Ja-le
| 2023-12-19T21:28:28Z | 15 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"whisper",
"automatic-speech-recognition",
"generated_from_trainer",
"base_model:openai/whisper-tiny",
"base_model:finetune:openai/whisper-tiny",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] |
automatic-speech-recognition
| 2023-12-19T19:21:53Z |
---
license: apache-2.0
base_model: openai/whisper-tiny
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: whisper-tiny-hi
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# whisper-tiny-hi
This model is a fine-tuned version of [openai/whisper-tiny](https://huggingface.co/openai/whisper-tiny) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 2.2091
- Wer: 86.2832
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0001
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 50
- training_steps: 500
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| 3.5332 | 3.57 | 25 | 1.7565 | 89.3805 |
| 0.3504 | 7.14 | 50 | 1.6160 | 90.2655 |
| 0.0551 | 10.71 | 75 | 1.8524 | 93.3628 |
| 0.0285 | 14.29 | 100 | 1.9261 | 123.0088 |
| 0.022 | 17.86 | 125 | 2.0688 | 92.9204 |
| 0.0075 | 21.43 | 150 | 2.0535 | 89.3805 |
| 0.0054 | 25.0 | 175 | 2.1533 | 86.7257 |
| 0.0016 | 28.57 | 200 | 2.1682 | 91.1504 |
| 0.001 | 32.14 | 225 | 2.2014 | 87.6106 |
| 0.001 | 35.71 | 250 | 2.1406 | 87.1681 |
| 0.0037 | 39.29 | 275 | 2.1968 | 88.0531 |
| 0.0012 | 42.86 | 300 | 2.1761 | 107.0796 |
| 0.0004 | 46.43 | 325 | 2.1874 | 88.0531 |
| 0.0003 | 50.0 | 350 | 2.2005 | 87.1681 |
| 0.0003 | 53.57 | 375 | 2.2018 | 87.1681 |
| 0.0002 | 57.14 | 400 | 2.2041 | 87.1681 |
| 0.0002 | 60.71 | 425 | 2.2055 | 86.7257 |
| 0.0002 | 64.29 | 450 | 2.2072 | 86.2832 |
| 0.0002 | 67.86 | 475 | 2.2089 | 87.1681 |
| 0.0002 | 71.43 | 500 | 2.2091 | 86.2832 |
### Framework versions
- Transformers 4.37.0.dev0
- Pytorch 2.1.0+cu121
- Datasets 2.15.0
- Tokenizers 0.15.0
|
afrideva/palmer-x-002-GGUF
|
afrideva
| 2023-12-19T21:27:03Z | 11 | 1 | null |
[
"gguf",
"ggml",
"quantized",
"q2_k",
"q3_k_m",
"q4_k_m",
"q5_k_m",
"q6_k",
"q8_0",
"text-generation",
"en",
"dataset:appvoid/no-prompt-15k",
"base_model:appvoid/palmer-x-002",
"base_model:quantized:appvoid/palmer-x-002",
"license:apache-2.0",
"region:us"
] |
text-generation
| 2023-12-19T21:23:33Z |
---
base_model: appvoid/palmer-x-002
datasets:
- appvoid/no-prompt-15k
inference: false
language:
- en
license: apache-2.0
model_creator: appvoid
model_name: palmer-x-002
pipeline_tag: text-generation
quantized_by: afrideva
tags:
- gguf
- ggml
- quantized
- q2_k
- q3_k_m
- q4_k_m
- q5_k_m
- q6_k
- q8_0
---
# appvoid/palmer-x-002-GGUF
Quantized GGUF model files for [palmer-x-002](https://huggingface.co/appvoid/palmer-x-002) from [appvoid](https://huggingface.co/appvoid)
| Name | Quant method | Size |
| ---- | ---- | ---- |
| [palmer-x-002.fp16.gguf](https://huggingface.co/afrideva/palmer-x-002-GGUF/resolve/main/palmer-x-002.fp16.gguf) | fp16 | 2.20 GB |
| [palmer-x-002.q2_k.gguf](https://huggingface.co/afrideva/palmer-x-002-GGUF/resolve/main/palmer-x-002.q2_k.gguf) | q2_k | 483.12 MB |
| [palmer-x-002.q3_k_m.gguf](https://huggingface.co/afrideva/palmer-x-002-GGUF/resolve/main/palmer-x-002.q3_k_m.gguf) | q3_k_m | 550.82 MB |
| [palmer-x-002.q4_k_m.gguf](https://huggingface.co/afrideva/palmer-x-002-GGUF/resolve/main/palmer-x-002.q4_k_m.gguf) | q4_k_m | 668.79 MB |
| [palmer-x-002.q5_k_m.gguf](https://huggingface.co/afrideva/palmer-x-002-GGUF/resolve/main/palmer-x-002.q5_k_m.gguf) | q5_k_m | 783.02 MB |
| [palmer-x-002.q6_k.gguf](https://huggingface.co/afrideva/palmer-x-002-GGUF/resolve/main/palmer-x-002.q6_k.gguf) | q6_k | 904.39 MB |
| [palmer-x-002.q8_0.gguf](https://huggingface.co/afrideva/palmer-x-002-GGUF/resolve/main/palmer-x-002.q8_0.gguf) | q8_0 | 1.17 GB |
## Original Model Card:
# x-002
This is an incremental model update on `palmer-002` using dpo technique. X means dpo+sft spinoff.
### evaluation
|Model| ARC_C| HellaSwag| PIQA| Winogrande|
|------|-----|-----------|------|-------------|
|tinyllama-2t| 0.2807| 0.5463| 0.7067| 0.5683|
|palmer-001| 0.2807| 0.5524| 0.7106| 0.5896|
|tinyllama-2.5t|0.3191|0.5896| 0.7307| 0.5872|
|palmer-002|**0.3242**|**0.5956**|0.7345|0.5888|
|palmer-x-002|0.3224|0.5941|**0.7383**|**0.5912**|
### training
~500 dpo samples as experimental data to check on improvements. It seems like data is making it better on some benchmarks while also degrading quality on others.
### prompt
```
no prompt
```
As you can notice, the model actually completes by default questions that are the most-likely to be asked, which is good because most people will use it to answer as a chatbot.
<a href="https://ko-fi.com/appvoid" target="_blank"><img src="https://cdn.buymeacoffee.com/buttons/v2/default-yellow.png" alt="Buy Me A Coffee" style="height: 48px !important;width: 180px !important; filter: invert(70%);" ></a>
|
afrideva/minima-3b-layla-v2-GGUF
|
afrideva
| 2023-12-19T20:50:07Z | 35 | 4 | null |
[
"gguf",
"ggml",
"quantized",
"q2_k",
"q3_k_m",
"q4_k_m",
"q5_k_m",
"q6_k",
"q8_0",
"text-generation",
"en",
"base_model:l3utterfly/minima-3b-layla-v2",
"base_model:quantized:l3utterfly/minima-3b-layla-v2",
"license:llama2",
"region:us"
] |
text-generation
| 2023-12-19T20:41:45Z |
---
base_model: l3utterfly/minima-3b-layla-v2
inference: false
language:
- en
license: llama2
model_creator: l3utterfly
model_name: minima-3b-layla-v2
pipeline_tag: text-generation
quantized_by: afrideva
tags:
- gguf
- ggml
- quantized
- q2_k
- q3_k_m
- q4_k_m
- q5_k_m
- q6_k
- q8_0
---
# l3utterfly/minima-3b-layla-v2-GGUF
Quantized GGUF model files for [minima-3b-layla-v2](https://huggingface.co/l3utterfly/minima-3b-layla-v2) from [l3utterfly](https://huggingface.co/l3utterfly)
| Name | Quant method | Size |
| ---- | ---- | ---- |
| [minima-3b-layla-v2.fp16.gguf](https://huggingface.co/afrideva/minima-3b-layla-v2-GGUF/resolve/main/minima-3b-layla-v2.fp16.gguf) | fp16 | 6.04 GB |
| [minima-3b-layla-v2.q2_k.gguf](https://huggingface.co/afrideva/minima-3b-layla-v2-GGUF/resolve/main/minima-3b-layla-v2.q2_k.gguf) | q2_k | 1.30 GB |
| [minima-3b-layla-v2.q3_k_m.gguf](https://huggingface.co/afrideva/minima-3b-layla-v2-GGUF/resolve/main/minima-3b-layla-v2.q3_k_m.gguf) | q3_k_m | 1.51 GB |
| [minima-3b-layla-v2.q4_k_m.gguf](https://huggingface.co/afrideva/minima-3b-layla-v2-GGUF/resolve/main/minima-3b-layla-v2.q4_k_m.gguf) | q4_k_m | 1.85 GB |
| [minima-3b-layla-v2.q5_k_m.gguf](https://huggingface.co/afrideva/minima-3b-layla-v2-GGUF/resolve/main/minima-3b-layla-v2.q5_k_m.gguf) | q5_k_m | 2.15 GB |
| [minima-3b-layla-v2.q6_k.gguf](https://huggingface.co/afrideva/minima-3b-layla-v2-GGUF/resolve/main/minima-3b-layla-v2.q6_k.gguf) | q6_k | 2.48 GB |
| [minima-3b-layla-v2.q8_0.gguf](https://huggingface.co/afrideva/minima-3b-layla-v2-GGUF/resolve/main/minima-3b-layla-v2.q8_0.gguf) | q8_0 | 3.21 GB |
## Original Model Card:
# Model Card
### Model Description
[MiniMA-3B](https://huggingface.co/GeneZC/MiniMA-3B) (from GeneZC) fine-tuned by:
1. Teatime Roleplay dataset for text completion
2. ShareGPT datasets for multi-turn conversations.
- **Developed by:** l3utterfly
- **Funded by:** Layla Network
- **Model type:** Llama2
- **Language(s) (NLP):** English
- **License:** Llama2
- **Finetuned from model:** MiniMA-3B
## Uses
Base model used by Layla - the offline personal assistant: https://www.layla-network.ai
Help & support: https://discord.gg/x546YJ6nYC
Prompt:
```
USER:
ASSISTANT:
```
[<img src="https://raw.githubusercontent.com/OpenAccess-AI-Collective/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/OpenAccess-AI-Collective/axolotl)
|
ChameleonAI/ChameleonAILoras
|
ChameleonAI
| 2023-12-19T20:49:50Z | 0 | 11 | null |
[
"region:us"
] | null | 2023-04-16T15:29:20Z |
---
# For reference on model card metadata, see the spec: https://github.com/huggingface/hub-docs/blob/main/modelcard.md?plain=1
# Doc / guide: https://huggingface.co/docs/hub/model-cards
{}
---
# Chameleon AI Loras
<!-- Provide a quick summary of what the model is/does. -->
You can find all my Loras uploaded to civitai here. Feels like the website is mostly down at the moment, so this is mostly a safety net.
## Model List
- [Judgement (Helltaker)](https://huggingface.co/ChameleonAI/ChameleonAILoras#1-judgement-helltaker)
- [Pascal (Tales of Grace)](https://huggingface.co/ChameleonAI/ChameleonAILoras#2-pascal-tales-of-grace)
- [Shishiro Botan (Hololive)](https://huggingface.co/ChameleonAI/ChameleonAILoras#3-shishiro-botan-hololive)
- [Sophia (Granblue Fantasy)](https://huggingface.co/ChameleonAI/ChameleonAILoras#4-sophia-granblue-fantasy)
- [Juliet Persia (Boarding School Juliet)](https://huggingface.co/ChameleonAI/ChameleonAILoras#5-juliet-persia-boarding-school-juliet)
- [Martha (Fate/Grand Order)](https://huggingface.co/ChameleonAI/ChameleonAILoras#6-martha-fategrand-order)
- [Tsunomaki Watame (Hololive)](https://huggingface.co/ChameleonAI/ChameleonAILoras#7-tsunomaki-watame-hololive)
- [Shana (Shakugan no Shana)](https://huggingface.co/ChameleonAI/ChameleonAILoras#8-shana-shakugan-no-shana)
- [Nonna (Girls und Panzer)](https://huggingface.co/ChameleonAI/ChameleonAILoras#9-nonna-girls-und-panzer)
- [Reimu Hakurei (Touhou)](https://huggingface.co/ChameleonAI/ChameleonAILoras#10-reimu-hakurei-touhou)
- [Ayase Fuuka (Yotsuba to!)](https://huggingface.co/ChameleonAI/ChameleonAILoras#11-ayase-fuuka-yotsuba-to)
- [Herja (Granblue Fantasy)](https://huggingface.co/ChameleonAI/ChameleonAILoras#12-herja-granblue-fantasy)
- [Sailor Jupiter (Sailor Moon)](https://huggingface.co/ChameleonAI/ChameleonAILoras#13-sailor-jupiter-sailor-moon)
- [Ouro Kronii (Hololive)](https://huggingface.co/ChameleonAI/ChameleonAILoras#14-ouro-kronii-hololive)
- [Back Tattoo Concept](https://huggingface.co/ChameleonAI/ChameleonAILoras#15-back-tattoo-concept)
- [Perona (One Piece)](https://huggingface.co/ChameleonAI/ChameleonAILoras#16-perona-one-piece)
- [Himari Azuma (Mato Seihei no Slave)](https://huggingface.co/ChameleonAI/ChameleonAILoras#17-himari-azuma-mato-seihei-no-slave)
- [Silva (Granblue Fantasy)](https://huggingface.co/ChameleonAI/ChameleonAILoras#18-silva-granblue-fantasy)
- [Sui-Feng (Bleach)](https://huggingface.co/ChameleonAI/ChameleonAILoras#19-sui-feng-bleach)
- [Sakura Matou (Fate)](https://huggingface.co/ChameleonAI/ChameleonAILoras#20-sakura-matou-fate)
- [Tweyen (Granblue Fantasy)](https://huggingface.co/ChameleonAI/ChameleonAILoras#21-tweyen-granblue-fantasy)
- [Magilou (Tales of Berseria)](https://huggingface.co/ChameleonAI/ChameleonAILoras#22-magilou-tales-of-berseria)
- [Yoruichi (Bleach)](https://huggingface.co/ChameleonAI/ChameleonAILoras#23-yoruichi-bleach)
- [Darjeeling (Girls und Panzer)](https://huggingface.co/ChameleonAI/ChameleonAILoras#24-darjeeling-girls-und-panzer)
- [Female Matsuri Kazamaki (Ayakashi Triangle)](https://huggingface.co/ChameleonAI/ChameleonAILoras#25-female-matsuri-kazamaki-ayakashi-triangle)
- [Mira Kamiunten (Mato Seihei no Slave)](https://huggingface.co/ChameleonAI/ChameleonAILoras#26-mira-kamiunten-mato-seihei-no-slave)
- [Dunkerque (Azur Lane)](https://huggingface.co/ChameleonAI/ChameleonAILoras#27-dunkerque-azur-lane)
- [Tae Takemi (Persona 5)](https://huggingface.co/ChameleonAI/ChameleonAILoras#28-tae-takemi-persona-5)
- [Satonaka Chie (Persona 4)](https://huggingface.co/ChameleonAI/ChameleonAILoras#29-satonaka-chie-persona-4)
- [Female Robin (Fire Emblem)](https://huggingface.co/ChameleonAI/ChameleonAILoras#30-female-robin-fire-emblem)
- [Kyouka Uzen (Mato Seihei no Slave)](https://huggingface.co/ChameleonAI/ChameleonAILoras#31-kyouka-uzen-mato-seihei-no-slave)
- [Izumo Tenka (Mato Seihei no Slave)](https://huggingface.co/ChameleonAI/ChameleonAILoras#32-izumo-tenka-mato-seihei-no-slave)
- [Nui Sociere (NIJISANJI)](https://huggingface.co/ChameleonAI/ChameleonAILoras#33-nui-sociere-nijisanji)
- [Erina Makina (Phase-Connect)](https://huggingface.co/ChameleonAI/ChameleonAILoras#34-erina-makina-phase-connect)
- [Tadokoro Megumi (Food Wars)](https://huggingface.co/ChameleonAI/ChameleonAILoras#35-tadokoro-megumi-food-wars)
- [Alice Zuberg (Sword Art Online)](https://huggingface.co/ChameleonAI/ChameleonAILoras#36-alice-zuberg-sword-art-online)
- [Yukihana Lamy (Hololive)](https://huggingface.co/ChameleonAI/ChameleonAILoras#37-yukihana-lamy-hololive)
- [Shirogane Noel (Hololive)](https://huggingface.co/ChameleonAI/ChameleonAILoras#38-shirogane-noel-hololive)
- [Essex (Azur Lane)](https://huggingface.co/ChameleonAI/ChameleonAILoras#39-essex-azur-lane)
- [Ereshkigal (Fate/Grand Order)](https://huggingface.co/ChameleonAI/ChameleonAILoras#40-ereshkigal-fategrand-order)
- [Inugami Korone (Hololive)](https://huggingface.co/ChameleonAI/ChameleonAILoras#41-inugami-korone-hololive)
- [Nakiri Ayamae (Hololive)](https://huggingface.co/ChameleonAI/ChameleonAILoras#42-nakiri-ayamae-hololive)
- [Shirakami Fubuki (Hololive)](https://huggingface.co/ChameleonAI/ChameleonAILoras#43-shirakami-fubuki-hololive)
- [Laplus Darknesss (Hololive)](https://huggingface.co/ChameleonAI/ChameleonAILoras#44-laplus-darknesss-hololive)
- [Houshou Marine (Hololive)](https://huggingface.co/ChameleonAI/ChameleonAILoras#45-houshou-marine-hololive)
- [Hoshimachi Suisei (Hololive)](https://huggingface.co/ChameleonAI/ChameleonAILoras#46-hoshimachi-suisei-hololive)
- [Pavolia Reine (Hololive)](https://huggingface.co/ChameleonAI/ChameleonAILoras#47-pavolia-reine-hololive)
- [Sailor Mars (Sailor Moon)](https://huggingface.co/ChameleonAI/ChameleonAILoras#48-sailor-mars-sailor-moon)
- [Akagi Towa/Twilight (Go! Princess Pretty Cure)](https://huggingface.co/ChameleonAI/ChameleonAILoras#49akagi-towatwilight-go-princess-pretty-cure)
- [Itsumi Erika (Girls und Panzer)](https://huggingface.co/ChameleonAI/ChameleonAILoras#50-itsumi-erika-girls-und-panzer)
- [Otonashi Kotori (Idolmaster)](https://huggingface.co/ChameleonAI/ChameleonAILoras#51-otonashi-kotori-idolmaster)
- [Tio Plato (Kiseki Games)](https://huggingface.co/ChameleonAI/ChameleonAILoras#52-tio-plato-kiseki-games)
- [Jessica (Granblue Fantasy)](https://huggingface.co/ChameleonAI/ChameleonAILoras#53-jessica-granblue-fantasy)
- [Kagamine Rin (Vocaloid)](https://huggingface.co/ChameleonAI/ChameleonAILoras#54-kagamine-rin-vocaloid)
- [Miku-390 (Darling in the Franxx)](https://huggingface.co/ChameleonAI/ChameleonAILoras#55-miku-390-darling-in-the-franxx)
- [Shigure (KanColle)](https://huggingface.co/ChameleonAI/ChameleonAILoras#56-shigure-kancolle)
- [Sena Kashiwazaki (Boku wa Tomodachi ga Sukunai)](https://huggingface.co/ChameleonAI/ChameleonAILoras#57-sena-kashiwazaki-boku-wa-tomodachi-ga-sukunai)
- [Secelia Dote (Mobile Suit Gundam The Witch from Mercury)](https://huggingface.co/ChameleonAI/ChameleonAILoras#58-secelia-dote-mobile-suit-gundam-the-witch-from-mercury)
- [Sailor Pluto (Sailor Moon)](https://huggingface.co/ChameleonAI/ChameleonAILoras#59-sailor-pluto-sailor-moon)
- [Feldt Grace (Mobile Suit Gundam 00)](https://huggingface.co/ChameleonAI/ChameleonAILoras#60-feldt-grace-mobile-suit-gundam-00)
- [Cagliostro (Granblue Fantasy)](https://huggingface.co/ChameleonAI/ChameleonAILoras#61-cagliostro-granblue-fantasy)
- [Ashelia (Final Fantasy XII)](https://huggingface.co/ChameleonAI/ChameleonAILoras#62-ashelia-final-fantasy-xii)
- [Tewi Inaba (Touhou)](https://huggingface.co/ChameleonAI/ChameleonAILoras#63-tewi-inaba-touhou)
- [Ferry (Granblue Fantasy)](https://huggingface.co/ChameleonAI/ChameleonAILoras#64-ferry-granblue-fantasy)
- [Ronye Arabel (Sword Art Online)](https://huggingface.co/ChameleonAI/ChameleonAILoras#65-ronye-arabel-sword-art-online)
- [Shrug Top (Concept LoRA)](https://huggingface.co/ChameleonAI/ChameleonAILoras#66-shrug-top-concept-lora)
- [Lum Outfit Cosplay (Concept LoRA)](https://huggingface.co/ChameleonAI/ChameleonAILoras#67-lum-outfit-cosplay-concept-lora)
- [Love Espada (Maken ki)](https://huggingface.co/ChameleonAI/ChameleonAILoras#68-love-espada-maken-ki)
- [Heles (Granblue Fantasy)](https://huggingface.co/ChameleonAI/ChameleonAILoras#69-heles-granblue-fantasy)
- [Io (Phantasy Star Online 2)](https://huggingface.co/ChameleonAI/ChameleonAILoras#70-io-phantasy-star-online-2)
- [Irisviel von Einzbern (Fate)](https://huggingface.co/ChameleonAI/ChameleonAILoras#71-irisviel-von-einzbern-fate)
- [Kjera (Arknights)](https://huggingface.co/ChameleonAI/ChameleonAILoras#72-kjera-arknights)
- [Rinwell (Tales of Arise)](https://huggingface.co/ChameleonAI/ChameleonAILoras#73-rinwell-tales-of-arise)
- [Zooey (Granblue Fantasy)](https://huggingface.co/ChameleonAI/ChameleonAILoras#74-zooey-granblue-fantasy)
- [Trick or Treatment Cosplay (Fate)](https://huggingface.co/ChameleonAI/ChameleonAILoras#75-trick-or-treatment-cosplay-fate)
- [Angel Mort Uniform (Higurashi)](https://huggingface.co/ChameleonAI/ChameleonAILoras#76-angel-mort-uniform-higurashi)
- [True off shoulder Bikini](https://huggingface.co/ChameleonAI/ChameleonAILoras#77-true-off-shoulder-bikini)
- [Honolulu (Azur Lane)](https://huggingface.co/ChameleonAI/ChameleonAILoras#78-honolulu-azur-lane)
### Model Details
#### 1. Judgement (Helltaker)
Weight: 0.9-1.0
Prompts: "Judgement"
Sub-prompts: "colored skin, horns, tail, long hair, ponytail"
Outfit: "Judgement, colored skin, horns, tail, long hair, ponytail, gauntlets, jacket, navel, belt, chain, armband, thighhighs"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-Judgement.safetensors)
#### 2. Pascal (Tales of Grace)
Weight: 1.0
Prompts: "PascalTales"
Sub-prompts: "short hair"
Outfit: "PascalTales, PascalOutfit, short hair"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-Pascal.safetensors)
#### 3. Shishiro Botan (Hololive)
Weight: 0.8-1.0
Prompts: "Botan"
Sub-prompts: -
Outfits:
Normal Outfit: "Botan, BotanOutfit, fur-trimmed, jacket, long hair"
Casual Outfit: "Botan, BotanCasual, standing, long hair"
Suit:
1. with corset: "Botan, BotanSuit, long hair, long sleeves, pantyhose, collared shirt, corset"
2. with vest: "Botan, BotanSuit, long hair, long sleeves, pantyhose, collared shirt, vest"
3. Sports bra only: "Botan, BotanSuit, grey hair, long hair, black sports bra, navel, midriff, cleavage, bare shoulders, pantyhose, dog tags"
New Year's Outfit: "Botan, BotanKimono, hair ornament, hair flower, double bun"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-Botan.safetensors)
#### 4. Sophia (Granblue Fantasy)
Weight: 0.7
Prompts: "Sophia"
Sub-prompts: "blue hair"
Outfits:
Pious Pilgrim: "Sophia, twintails, gloves, hat, SophiaDress"
Enlightened Priestess: "Sophia, twintails, white dress, white gloves, elbow gloves, hair flower, circlet"
Pilgrim on a Short Break: "Sophia, long hair, beret, coat, open coat, sweater dress, belt"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-Sophia.safetensors)
#### 5. Juliet Persia (Boarding School Juliet)
Weight: 1.0
Prompts: "JulietPersia"
Sub-prompts: "hair ribbon"
Outfit: "JulietPersia, JulietSchoolUniform, hair ribbon"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-JulietPersia.safetensors)
#### 6. Martha (Fate/Grand Order)
Weight: 0.7-1.0
Prompts: "Martha"
Sub-prompts: "-"
Outfits:
Base Outfit: "Martha, MarthaOutfit, red thighhighs"
Santa: "Martha, MarthaSanta, santa hat, christmas, fur-trimmed, apron"
Swimsuit Ruler: "Martha, MarthaBikini, frilled bikini, collarbone, navel, choker, thigh strap"
Aerial Drive: "Martha, MarthaMecha, mecha musume"
Holy Maiden's Teaching: "Martha, MarthaWarrior"
Heroic Spirit Traveling Outfit: "Martha, MarthaTravel, double bun, chinese dress"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-Martha.safetensors)
#### 7. Tsunomaki Watame (Hololive)
Weight: 1.0
Prompts: "Tsunomaki Watame"
Sub-prompts: "sheep girl, sheep ears, sheep horns"
Outfits:
Normal Outfit: Tsunomaki Watame, long hair, WatameBase, fur-trimmed dress, white dress, bare shoulders, fur-trimmed sleeves, hairclip, cape, belt pouch, brooch, fur-trimmed boots
Casual Outfit: Tsunomaki Watame, side ponytail, WatameCasual, striped dress, sailor collar, hairclip, red choker, see-through sleeves, white bow, cardigan, open cardigan
Sleepwear: Tsunomaki Watame, hair bun, WatameSleep, blue jacket, open jacket, camisole, hairclip, thigh strap, hair flower, short shorts, pink shorts, barefoot
Watame Night Fever: Tsunomaki Watame, very long hair, WatameIdol, black dress, halterneck, detached sleeves, hairclip, red gloves, single glove, thigh strap, overskirt, hair ribbon, bare shoulders
New Year's: Tsunomaki Watame, twin braids, long hair, WatameKimono, floral print, red kimono, sash, hair flower, bow, fur scarf, hat
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/WatameV4.safetensors)
#### 8. Shana (Shakugan no Shana)
Weight: 1.0
Prompts: "Shana"
Sub-prompts: "-"
Outfits:
red hair version: "Shana, red hair, red eyes"
black hair version: "Shana, black hair, brown eyes"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-Shana.safetensors)
#### 9. Nonna (Girls und Panzer)
Weight: 1.0
Prompts: "NonnaGuP"
Sub-prompts: "-"
Outfit: "NonnaGuP, pravda school uniform, green jacket, insignia, red shirt, black skirt, pleated skirt"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-NonnaGuP.safetensors)
#### 10. Reimu Hakurei (Touhou)
Weight: 1.0
Prompts: "Hakurei Reimu"
Sub-prompts: "hair tubes, hair bow"
Outfits: "Hakurei Reimu, red skirt, red shirt, long sleeves, navel, bare shoulders, hair bow, frills, wide sleeves, detached sleeves, hair tubes, yellow ascot"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-Reimu.safetensors)
#### 11. Ayase Fuuka (Yotsuba to!)
Weight: 1.0
Prompts: "Ayase Fuuka"
Sub-prompts: "-"
Outfit: "-"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-AyaseFuukaLORA.safetensors)
#### 12. Herja (Granblue Fantasy)
Weight: 0.8
Prompts: "Herja"
Sub-prompts: "ponytail"
Outfit: "Herja, ponytail, thighhighs, bare shoulders, ribbed sweater, sweater dress, belt, scarf, cape, thigh boots"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-Herja.safetensors)
#### 13. Sailor Jupiter (Sailor Moon)
Weight: 0.9
Prompts: "SMJupiter, SMJupiterOutfit"
Sub-prompts: "ponytail"
Outfit: "SMJupiter, SMJupiterOutfit, green sailor collar, green skirt, sailor senshi uniform, ponytail"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-SailorJupiter.safetensors)
#### 14. Ouro Kronii (Hololive)
Weight: 1.0
Prompts: "Ouro Kronii"
Sub-prompts: "-"
Outfits:
Default: "Ouro Kronii, 1girl, solo, breasts, short hair, skirt, shirt, thighhighs, gloves, long sleeves, bow, cleavage, bare shoulders, jewelry, white shirt, earrings, detached sleeves, sleeveless, black gloves, striped, puffy sleeves, black thighhighs, miniskirt, black skirt, bowtie, clothing cutout, sleeveless shirt, chain, blue bow, cleavage cutout, vertical stripes, zipper, asymmetrical legwear, striped skirt, head chain, bow earrings"
Casual: "Ouro Kronii, 1girl, solo, long hair, multicolored hair, breasts, pants, bag, sweater, coat, turtleneck, denim, jeans, handbag, turtleneck sweater, high-waist pants, bare shoulders"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-OuroKronii.safetensors)
#### 15. Back Tattoo Concept
Weight: 0.6-2.0
Prompts: "back tattoo"
Sub-prompts: "-"
Outfit: "-"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CONCEPT-BackTattoo.safetensors)
#### 16. Perona (One Piece)
Weight: 1.0
Prompts: "Perona"
Sub-prompts: "twintails, twin drills, circle-shaped eyes, black eyes"
Outfit: "-"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-Perona.safetensors)
#### 17. Himari Azuma (Mato Seihei no Slave)
Weight: 0.8
Prompts: "Himari Azuma"
Sub-prompts: "hairrings, long hair"
Outfit: "epaulettes, uniform, short sleeves, skirt, thighhighs"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-HimariAzuma.safetensors)
#### 18. Silva (Granblue Fantasy)
Weight: 0.55
Prompts: "Silva"
Sub-prompts: "silver hair"
Outfits:
Silva (Water):
long hair, skirt, long sleeves, navel, cleavage, braid, ahoge, midriff, belt, crop top, knee boots, coat
Silva (Light):
white shirt, navel, cleavage, collarbone, ponytail, black jacket, belt, black pants, crop top, open jacket
Silva (Summer):
beach, cleavage, navel, bare shoulders, blue bikini, blue thigh strap, sarong
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-Silva.safetensors)
#### 19. Sui-Feng (Bleach)
Weight: 0.8
Prompts: "Sui-Feng"
Sub-prompts: "-"
Outfit: "sui-feng, japanese clothes, black hakama, hip vent, sleeveless"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-SuiFeng.safetensors)
#### 20. Sakura Matou (Fate)
Weight: 0.8
Prompts: "Matou Sakura"
Sub-prompts: "-"
Outfit: "-"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-HimariAzuma.safetensors)
#### 21. Tweyen (Granblue Fantasy)
Weight: 1.0
Prompts: "Tweyen"
Sub-prompts: "head wings"
Outfits:
Armor: "Tweyen, TweyenArmor, smile, head wings, black shorts, elbow gloves, detached leggings, midriff"
Bikini: "Tweyen, TweyenBikini"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-Tweyen.safetensors)
#### 22. Magilou (Tales of Berseria)
Weight: 1.0
Prompts: "Magilou"
Sub-prompts: "very long hair, hair inbetween eyes"
Outfit: "Magilou, thighhighs, witch hat, bare shoulders, detached sleeves, book, strapless, garter straps, asymmetrical legwear, fur collar, asymmetrical sleeves, mismatched sleeves"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-Magilou.safetensors)
#### 23. Yoruichi (Bleach)
Weight: 1.0
Prompts: "Shihouin Yoruichi"
Sub-prompts: "purple hair, yellow eyes"
Outfit: "Shihouin Yoruichi, black leotard, elbow gloves, thighhighs"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-Yoruichi.safetensors)
#### 24. Darjeeling (Girls und Panzer)
Weight: 1.0
Prompts: "Darjeeling"
Sub-prompts: "twin braids"
Outfits:
St. Gloriana's Military Uniform:
"black skirt, uniform, red jacket, st. gloriana's military uniform"
St. Gloriana's School Uniform:
"st. gloriana's school uniform, school uniform, blue sweater, white shirt, pantyhose, pleated skirt, blue skirt, black necktie"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-Darjeeling.safetensors)
#### 25. Female Matsuri Kazamaki (Ayakashi Triangle)
Weight: 1.0
Prompts: "Matsuri Kazamaki"
Sub-prompts: "-"
Outfit: "-"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-MatsuriKazamaki.safetensors)
#### 26. Mira Kamiunten (Mato Seihei no Slave)
Weight: 1.0
Prompts: "Mira Kamiunten"
Sub-prompts: "-"
Outfit: "Mira Kamiunten, large breasts, sarashi, bandage, navel, collar, open jacket, black jacket, black baggy pants, cap"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-MiraKamiunten.safetensors)
#### 27. Dunkerque (Azur Lane)
Weight: 1.0
Prompts: "DunkerqueAL"
Sub-prompts: "grey hair"
Outfits:
Default:
"DunkerqueAL, DunkerqueUniform"
Summer Sucré:
"DunkerqueAL, DunkerqueBikini, ponytail"
Afternoon Venus:
"DunkerqueAL, DunkerqueCasual, white dress, sun hat, eyewear on head, tinted eyewear, long sleeves"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-Dunkerque.safetensors)
#### 28. Tae Takemi (Persona 5)
Weight: 0.7
Prompts: "Tae Takemi"
Sub-prompts: "-"
Outfit: "Tae Takemi, necklace, choker, dress, labcoat"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-TaeTakemi18MB.safetensors)
#### 29. Satonaka Chie (Persona 4)
Weight: 0.7
Prompts: "Satonaka Chie"
Sub-prompts: "-"
Outfit: "Satonaka Chie, green jacket, black skirt, track jacket"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-SatonakaChieLORA.safetensors)
#### 30. Female Robin (Fire Emblem)
Weight: 0.7
Prompts: "RobinFE"
Sub-prompts: "twintails, white hair, brown eyes"
Outfit: "-"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-FemaleRobin18MB.safetensors)
#### 31. Kyouka Uzen (Mato Seihei no Slave)
Weight: 0.7
Prompts: "Kyouka Uzen"
Sub-prompts: "silver hair, horns"
Outfit: "Kyouka Uzen, military, military uniform, black skirt, pleated skirt, thigh boots, white gloves, belt, epaulettes"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-KyoukaUzenLORA.safetensors)
#### 32. Izumo Tenka (Mato Seihei no Slave)
Weight: 0.7
Prompts: "Izumo Tenka"
Sub-prompts: "short hair"
Outfit: "Izumo Tenka, short hair, belt, black shorts, buttons, cape, earrings, epaulettes, gloves, military, military uniform, thigh boots, uniform, white gloves"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-IzumoTenkaLORASmallSize.safetensors)
#### 33. Nui Sociere (NIJISANJI)
Weight: 0.7
Prompts: "Nui Sociere"
Sub-prompts: "blonde hair"
Outfit: "Nui Sociere, black dress, black gloves, black thighighs, bra peek, cape, cleavage, collarbone, clothing cutout, elbow gloves, witch hat, navel cutout, ribbon, side slit"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-NuiSociere.safetensors)
#### 34. Erina Makina (Phase-Connect)
Weight: 0.7
Prompts: "Erina Makina"
Sub-prompts: "-"
Outfit: "-"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-ErinaMakinaReducedv2.safetensors)
#### 35. Tadokoro Megumi (Food Wars)
Weight: 0.7
Prompts: "Tadokoro Megumi"
Sub-prompts: "twin braids"
Outfits:
School Uniform: "Tadokoro Megumi, MegumiUniform"
Cooking Outfit: "Tadokoro Megumi, MegumiChef, white shirt, white pants"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-TadokoroMegumiReducedv3.safetensors)
#### 36. Alice Zuberg (Sword Art Online)
Weight: 0.7
Prompts: "Alice"
Sub-prompts: "long hair, braid"
Outfit: "Alice, AliceArmor, armor, armored dress"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-AliceZuberg.safetensors)
#### 37. Yukihana Lamy (Hololive)
Weight: 0.7
Prompts: "Lamy"
Sub-prompts: "-"
Outfits:
Base:
"Lamy, Lamyoutfit"
Second Outfit (Casual):
"Lamy, LamyCasual, two side up"
Third Outfit (Sleepwear):
"Lamy, LamySleepwear, messy hair"
Fourth Outfit (New Year's):
"Lamy, LamyKimono, pink haori" Only add pink haori, if you want her to wear it.
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-LamyChameleonAI_v1.0.safetensors)
#### 38. Shirogane Noel (Hololive)
Weight: 0.7
Prompts: "Noel"
Sub-prompts: "-"
Outfits:
Base:
"Noel, NoelOutfit"
Second Outfit (Casual):
"Noel, NoelCasual, brown skirt, plaid skirt, off-shoulder sweater, white sweater, collarbone, bra straps"
Third Outfit (Bavarian Beer Girl):
"Noel, NoelBeerGirl, german clothes, dirndl, black bowtie, puffy sleeves, detached sleeves, white waist apron, flower, hair ornament, twin braids""
Fourth Outfit:
"Noel, NoelNoArmor, sleeveless black undershirt, white off-shoulder shirt, see-through clevage, white detached sleeves, wide sleeves, black shorts, cross-laced shorts, white overskirt, thigh strap"
Fifth Outfit (School Uniform):
"Noel, NoelUniform"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-NoelChameleonAI_v.1.0.safetensors)
#### 39. Essex (Azur Lane)
Weight: 0.7
Prompts: "Essex"
Sub-prompts: "-"
Outfits:
Base:
Essex, EssexUniform (Weight ~0.7)
A Trip Down Route 66:
Essex, EssexBiker, blue hair, yellow eyes, ponytail, sports bra, pants, open jacket (Weight ~0.7)
Brush and Ink:
Essex, EssexChinaDress, smile, blush, hair bun, double bun, blue hair (Weight ~0.7)
Craft Fairytail:
Essex, EssexPartyDress (Weight ~0.7)
Detective Essex:
Essex, EssexDetective, smile, happy, white shirt, fingerless gloves, midriff, necktie, skirt (Weight ~0.85)
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-EssexChameleonAI_v1.0.safetensors)
#### 40. Ereshkigal (Fate/Grand Order)
Weight: 0.6
Prompts: "Ereshkigal"
Sub-prompts: "-"
Outfit: "Ereshkigal, EreshkigalDress"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-EreshkigalChameleonAI_v1.0.safetensors)
#### 41. Inugami Korone (Hololive)
Weight: 0.7
Prompts: "Korone"
Sub-prompts: "dog girl, dog ears, dog tail, twin braids, sidelocks"
Outfit: "Korone, hair ornament, ((white dress)), yellow jacket, bow, collar, collarbone, dress, jacket, open clothes, open jacket, red bow, short dress, sleeveless dress"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-Korone_v.1.0.safetensors)
#### 42. Nakiri Ayamae (Hololive)
Weight: 1.0
Prompts: "Nakiri Ayame"
Sub-prompts: "-"
Outfits:
Base Outfit:
"Nakiri Ayame, long hair, double bun, AyameBase, black kimono, hair bell, obi, white thighhighs, bare shoulders, long sleeves, oni mask, mask on head"
Bikini:
"Nakiri Ayame, long hair, twintails, AyameBikini, blue bikini, frilled bikini, bikini skirt, hair bow, black bow, frilled choker"
Lolita Fashion:
"Nakiri Ayame, long hair, AyameFrills, white shirt, collared shirt, long sleeves, shoulder cutouts, black skirt, high-waist skirt, blue bow, frilled hairband"
Casual:
"Nakiri Ayame, long hair, twintails, AyameCasual, white shirt, sleeveless shirt, black necktie, black skirt, pleated skirt, hair ribbon, x hair ornament, black thighhighs, thigh strap, black jacket, open jacket, off shoulder, black choker"
New Year's:
"Nakiri Ayame, long hair, side ponytail, AyameNewYears, red kimono, floral print, hair flower, sash, wide sleeves"
Shrine Maiden:
"Nakiri Ayame, long hair, braids, AyameShrine, japanese clothes, hakama skirt, white thighhighs"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-NakiriAyame_v2.safetensors)
#### 43. Shirakami Fubuki (Hololive)
Weight: 0.6-0.8
Prompts: "Fubuki"
Sub-prompts: "fox ears, fox girl, white hair, aqua eyes"
Outfit: "Fubuki, strapless top, front slit, blue neckerchief, white detached wide sleeves, black shorts, single thighhigh, thigh strap, navel"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-ShirakamiFubuki_v.1.0.safetensors)
#### 44. Laplus Darknesss (Hololive)
Weight: 0.6-0.9
Prompts: "Laplus"
Sub-prompts: "grey hair, purple streak, yellow eyes, horns"
Outfit: "Laplus, long sleeves, belt, collar, sleeves past wrists, ascot, single thighhigh, sleeves past fingers, yellow ascot, ankle cuffs"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-LaplusDarknesss_v.1.0.safetensors)
#### 45. Houshou Marine (Hololive)
Weight: 1.0
Prompts: "Houshou Marine"
Sub-prompts: "long hair, twintails, hair ribbon"
Outfits:
Base Outfit:
"houshouBase, heterochromia, red eyes, yellow eyes, twintails, long hair, hair ribbon, large breasts, white gloves, frilled choker, red ascot, leotard, leotard under clothes, red jacket, cropped jacket, sleeveless jacket, black coat, off shoulder, bicorne, red skirt, miniskirt, leather belt, black thighhighs"
Gothic:
"houshouGothic, heterochromia, red eyes, yellow eyes, twintails, black ribbon, large breasts, mini top hat, hat flower, gothic lolita, short dress, red dress, frilled dress, detached sleeves, frilled sleeves, corset, bowtie, black gloves, pocket watch, white thighhighs"
Bikini:
"houshouBikini, heterochromia, red eyes, yellow eyes, ponytail, long hair, jewelry, baseball cap, sunglasses, eyewear on headwear, black jacket, open jacket, white shorts, short shorts, red bikini, string bikini, o-ring thigh strap"
Band:
"houshouBand, heterochromia, red eyes, yellow eyes, bangs, long hair, streaked hair, shako cap, two-sided jacket, off shoulder, elbow gloves, lace-trimmed leotard, bodystocking, high-waist skirt, showgirl skirt, thigh boots, large breasts"
Officer:
"houshouOfficer, heterochromia, red eyes, yellow eyes, large breasts, bangs, short hair, sleeveless shirt, collared shirt, black skirt, side slit, pantyhose, id card, fingerless gloves"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-HoushouMarine_v2.safetensors)
#### 46. Hoshimachi Suisei (Hololive)
Weight: 0.6-0.9
Prompts: "Suisei"
Sub-prompts: "blue hair, blue eyes, side ponytail, bangs"
Outfit: "Suisei, plaid, plaid dress, grey dress, blue belt, blue ribbon, star brooch"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-Suisei.safetensors)
#### 47. Pavolia Reine (Hololive)
Weight: 0.7
Prompts: "Reine"
Sub-prompts: "aqua eyes, bright pupils"
Outfits:
Base:
"Reine, long hair, hair ornament, thighighs, dress, jewelry, braid, earrings, detached sleeves, side ponytail, blue dress, feather hair ornament, braided bang, ((navel cutout)), long skirt, navel, one-piece"
Casual:
"Reine, short hair, crop top, see-through shirt, covered cleavage, pants"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-PavoliaReine_v.1.0.safetensors)
#### 48. Sailor Mars (Sailor Moon)
Weight: 1.0
Prompts: "SMMars"
Sub-prompts: "very long hair, parted bangs, 1990s \(style\)"
Outfit: "SMMars, sailor senshi uniform, red sailor collar, red skirt, elbow gloves"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-SailorMarsLORA.safetensors)
#### 49. Akagi Towa/Twilight (Go! Princess Pretty Cure)
Weight: 1.0
Prompts: "Akagi Towa, Twilight"
Sub-prompts: "Akagi Towa, long hair, pink hair, parted bangs"
"Twilight, white hair, long hair, quad tails"
Outfits:
Akagi Towa:
"Akagi Towa, long hair, pink hair, parted bangs, red dress, magical girl, tiara, red sleeves, detached sleeves, skirt, earrings"
Twilight:
"Twilight, white hair, long hair, quad tails, black dress, red pantyhose, TwilightBelt, skirt, bow"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-Twilight.safetensors)
#### 50. Itsumi Erika (Girls und Panzer)
Weight: 1.0
Prompts: "Itsumi Erika"
Sub-prompts: "medium hair, bangs"
Outfit: "Itsumi Erika, bangs, kuromorimine military uniform, black jacket, red shirt, red skirt, hat"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-ItsumiErika.safetensors)
#### 51. Otonashi Kotori (Idolmaster)
Weight: 1.0
Prompts: "Otonashi Kotori"
Sub-prompts: "short hair"
Outfit: "Otonashi Kotori, short hair, hairband, thighhighs, black skirt, pencil skirt, yellow bowtie, white shirt, green vest"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-OtonashiKotori.safetensors)
#### 52. Tio Plato (Kiseki Games)
Weight: 0.8
Prompts: "Tio Plato"
Sub-prompts: "long hair, cat ears"
Outfits:
CS3:
"Tio Plato, long hair, thighhighs, long sleeves, pleated skirt, necktie, shirt, jacket, vest, open jacket, cat ears"
Azure/Zero:
"Tio Plato, long hair, skirt, armor, cape, thighhighs, cat ears"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-TioPlato.safetensors)
#### 53. Jessica (Granblue Fantasy)
Weight: 1.0
Prompts: "Jessica"
Sub-prompts: "long hair, bangs, goggles"
Outfits:
Base:
"jessicaOutfit, long hair, bangs, black thighhighs, gloves, goggles, goggles on head, black dress, shrug \(clothing\), cleavage"
Bikini:
"jessicaSummer, long hair, bangs, cat ears, frilled bikini"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-JessicaGB.safetensors)
#### 54. Kagamine Rin (Vocaloid)
Weight: 1.0
Prompts: "Kagamine Rin"
Sub-prompts: "short hair, bow, number tattoo"
Outfit: "Kagamine Rin, short hair, number tattoo, bow, white shirt, detached sleeves, belt, sailor collar, headphones, shorts, leg warmers"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-KagamineRinLORA.safetensors)
#### 55. Miku-390 (Darling in the Franxx)
Weight: 1.0
Prompts: "Miku390"
Sub-prompts: "long hair, twintails"
Outfits:
Uniform: "mikuUniform, twintails, long hair, military uniform"
Pilot Suit: "mikuPilot, twintails, long hair, bodysuit, white bodysuit, skin tight"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-Miku390.safetensors)
#### 56. Shigure (KanColle)
Weight: 1.0
Prompts: "ShigureKancolle"
Sub-prompts: "long hair, single braid, hair flaps"
Outfits:
School Uniform: "ShigureBase, black serafuku, school uniform, pleated skirt, fingerless gloves"
Bikini: "ShigureBikini, sarong, sailor bikini" add "sarong" if you want.
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-ShigureKancolle.safetensors)
#### 57. Sena Kashiwazaki (Boku wa Tomodachi ga Sukunai)
Weight: 1.0
Prompts: "Sena Kashiwazaki"
Sub-prompts: "long hair, butterfly hair ornament"
Outfits:
School Uniform with jacket:
"Sena Kashiwazaki, long hair, butterfly hair ornament, st. chronica academy school uniform, green jacket, plaid skirt"
School Uniform without jacket:
"Sena Kashiwazaki, long hair, butterfly hair ornament, st. chronica academy school uniform, dress shirt, plaid skirt"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-SenaKashiwazakiLORA.safetensors)
#### 58. Secelia Dote (Mobile Suit Gundam The Witch from Mercury)
Weight: 1.0
Prompts: "Secelia Dote"
Sub-prompts: "short hair"
Outfit: "SeceliaUniform, asticassia school uniform, partially unzipped, green shorts, white thighhighs"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-SeceliaDote.safetensors)
#### 59. Sailor Pluto (Sailor Moon)
Weight: 1.0
Prompts: "SMPluto"
Sub-prompts: "long hair, single hair bun"
Outfit: "sailor senshi uniform, knee boots, bow, white gloves, elbow gloves, pleated skirt"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-SailorPluto.safetensors)
#### 60. Feldt Grace (Mobile Suit Gundam 00)
Weight: 1.0
Prompts: "Feldt Grace"
Sub-prompts: "long hair, twintails"
Outfits:
Jacket:
"Feldt Grace, FeldtUniform, long hair, cropped jacket, pink jacket, white pants, belt, white gloves"
Bodysuit:
"Feldt Grace, FeldtSuit, long hair, twintails, elbow gloves, thigh boots, belt, skin thigh, yellow bodysuit"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-FeldtGrace.safetensors)
#### 61. Cagliostro (Granblue Fantasy)
Weight: 1.0
Prompts: "Cagliostro"
Sub-prompts: "long hair"
Outfits:
Base:
"Cagliostro, long hair, CagliostroBase, hairband, crown, black thighhighs, red bow, red skirt, cape"
Halloween:
"Cagliostro, long hair, CagliostroHalloween, thighhighs, orange bowtie, orange skirt, halloween costume, hood, black cape"
Dark:
"Cagliostro, long hair, blue skirt, shirt, gloves, bow,black thighhighs, cape"
Swimsuit:
"Cagliostro, long hair, CagliostroSwimsuit, ponytail, hair flower, sailor collar, one-piece swimsuit, heart-shaped eyewear, eyewear on head"
Grand:
"Cagliostro, long hair, CagliostroGrand, white dress, long sleeves, detached sleeves, frills, frilled dress, hairband"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-Cagliostro.safetensors)
#### 62. Ashelia (Final Fantasy XII)
Weight: 1.0
Prompts: "Ashelia"
Sub-prompts: "short hair, blonde hair or silver hair"
Outfit: "Ashelia, short hair, AsheliaOutfit, miniskirt, thighhighs, jewelry"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-Ashelia.safetensors)
#### 63. Tewi Inaba (Touhou)
Weight: 1.0
Prompts: "Tewi Inaba"
Sub-prompts: "short hair, black hair, red eyes"
Outfit: "Inaba Tewi, short hair, black hair, red eyes, TewiBase, carrot necklace, pink dress, short sleeves, puffy sleeves"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-InabaTewi.safetensors)
#### 64. Ferry (Granblue Fantasy)
Weight: 1.0
Prompts: "Ferry"
Sub-prompts: "long hair"
Outfits:
Base:
"Ferry, long hair, FerryBase, thighhighs, bare shoulders, jewelry, sleeveless, white dress, blue skirt, gloves"
Bikini:
"Ferry, long hair, FerryBikini, bikini, earrings, hair flower, thigh strap, bikini skirt"
Halloween:
"Ferry, long hair, FerryHalloween, orange dress, white shirt, striped skirt, gloves, top hat, earrings, cape"
Grand:
"Ferry, long hair, FerryGrand, thighhighs, elbow gloves, black dress, earrings"
Santa:
"Ferry, long hair, FerrySanta, red dress, fur trim, thighhighs, gloves, long sleeves, detached sleeves, earrings"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-Ferry.safetensors)
#### 65. Ronye Arabel (Sword Art Online)
Weight: 1.0
Prompts: "Ronye Arabel"
Sub-prompts: "short hair, blue eyes"
Outfit: "Ronye Arabel, short hair, RonyeUniform, long sleeves, school uniform, grey skirt, black thighhighs"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-RonyeArabel.safetensors)
#### 66. Shrug Top (Concept LoRA)
Weight: 1.0
Prompts: "shrug \(clothing\), long sleeves" or "shrug \(clothing\), short sleeves"
Remember the to put a backslash infront of the brackets.
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CONCEPT-ShrugTopLORA.safetensors)
#### 67. Lum Outfit Cosplay (Concept LoRA)
Weight: 1.0
Prompts: "LumCosplay, bikini, strapless"
Sub-prompts: "horns, boots"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CONCEPT-LumCosplay.safetensors)
#### 68. Love Espada (Maken ki)
Weight: 1.0
Prompts: "Love Espada"
Sub-prompts: "long hair, ponytail"
Outfit: "Love Espada, long hair, ponytail, breasts, school uniform, serafuku, sleeveless, white gloves, belt, blue skirt, white thighhighs, hair ribbon"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-LoveEspada.safetensors)
#### 69. Heles (Granblue Fantasy)
Weight: 1.0
Prompts: "Heles"
Sub-prompts: "long hair"
Outfits:
Base:
"Heles, long hair, HelesBase, blue dress, overskirt, cleavage, shoulder armor, pauldrons, thighhighs, armor, gloves, hairband"
Summer:
"Heles, long hair, HelesSummer, jewelry, collarbone, white one-piece swimsuit, covered navel, sarong, sun hat, hat flower, ears through headwear"
Both sarong and sun hat are removable.
Irestill Evening Dress:
"Heles, long hair, HelesEvening, black thighhighs, bare shoulders, elbow gloves, white gloves, black dress, backless outfit"
Wind:
"Heles, long hair, HelesWind, red dress, breastplate, bare shoulders, armor, tiara, gauntlets, armored boots"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-HelesLORA.safetensors)
#### 70. Io (Phantasy Star Online 2)
Weight: 1.0
Prompts: "IoPSO2"
Sub-prompts: "short hair, heterochromia, blue eyes, red eyes"
Outfit: "IoPSO2, heterochromia, blue eyes, red eyes, short hair, tattoo, horns, IoBase, uniform, black shirt, shorts, thighhighs"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-IoPSO2.safetensors)
#### 71. Irisviel von Einzbern (Fate)
Weight: 1.0
Prompts: "Irisviel von Einzbern"
Sub-prompts: "long hair"
Outfits:
Casual (Red shirt and white skirt):
"Irisviel von Einzbern, long hair, red shirt, dress shirt, white skirt, pantyhose"
Winter (White coat):
"Irisviel von Einzbern, long hair, pantyhose, thighboots, coat, fur hat"
Dress of Heaven:
"Irisviel von Einzbern, long hair, IrisvielCaster, white dress, navel, bare shoulders, detached sleeves, crown, stomach tattoo"
The Black Grail:
"Irisviel von Einzbern, long hair, IrisvielBlackGrail, black dress, navel, bare shoulders, detached sleeves, crown, stomach tattoo"
Halloween Princess:
"Irisviel von Einzbern, long hair, IrisvielHalloween, thighhighs, cleavage, tail, detached sleeves, wings, horns, demon girl"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-Irisviel.safetensors)
#### 72. Kjera (Arknights)
Weight: 1.0
Prompts: "Kjera"
Sub-prompts: "short hair, hair flaps"
Outfits:
Base:
Kjera, short hair, hair flaps, tail, KjeraBase, white dress, fur trim, hair ornament, long sleeves, jewelry, pantyhose, wide sleeves, cape
Maid:
Kjera, short hair, tail, KjeraMaid, maid, bare shoulders, cleavage, necklace, pantyhose, detached sleeves, maid headdress, strapless
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-KjeraLora.safetensors)
#### 73. Rinwell (Tales of Arise)
Weight: 1.0
Prompts: "Rinwell"
Sub-prompts: "short hair"
Outfit: "Rinwell, short hair, hair ornament, skirt, thighhighs, fingerless gloves, detached sleeves, hood down, sleeveless coat"
Hootle: "owl"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-Rinwell.safetensors)
#### 74. Zooey (Granblue Fantasy)
Weight: 1.0
Prompts: "Zooey, dark skin"
Sub-prompts: "long hair"
Outfits:
Base (Promo):
Zooey, dark skin, ZooeyBase, long hair, thighhighs, black gloves, bare shoulders, blue dress, armored dress, breastplate, armor
Grand (Bikini):
Zooey, dark skin, ZooeyGrand, long hair, hair flower, white bikini, front-tie top
Event (Shrine Maiden):
Zooey, dark skin, ZooeyEvent, long hair, red skirt, white thighhighs, hair bow, detached sleeves, wide sleeves
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-Zooey.safetensors)
#### 75. Trick or Treatment Cosplay (Fate)
Weight: 1.0
Prompts: "Trick or Treatment Cosplay, revealing clothes, shrug \(clothing\), short sleeves, layered bikini, purple bikini, green bikini, belt, nurse cap, microskirt, green gloves, thighhighs, garter straps"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CONCEPT-TrickOrTreatmentCosplayLORA.safetensors)
#### 76. Angel Mort Uniform (Higurashi)
Weight: 1.0
Prompts: "Angel Mort Uniform, thighhighs, bare shoulders, detached sleeves, leotard, waitress"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CONCEPT-AngelMortUniform.safetensors)
#### 77. True off shoulder Bikini
Weight: 1.0
Prompts: "off-shoulder bikini, insert color bikini"
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CONCEPT-TrueOffShoulderBikini.safetensors)
#### 78. Honolulu (Azur Lane)
Weight: 1.0
Prompts: "Honolulu"
Sub-prompts: "long hair, twintails"
Outfits:
Default:
Honolulu, HonoluluBase, long hair, twintails, black ribbon, peaked cap, chain, jacket, jacket on shoulder, short dress, white gloves, elbow gloves, black thighhighs, garter straps
Umbrella Girl (School Uniform):
Honolulu, HonoluluSchool, long hair, twintails, serafuku, short sleeves, hair ribbon, black ribbon, black pantyhose, black skirt, beret, choker
Summer Accident?! (Bikini):
Honolulu, HonoluluBikini, very long hair, twintails, black bikini, hair ribbon, eyewear on head, sunglasses, star hair ornament
Among the Stalls (Kimono):
Honolulu, HonoluluKimono, very long hair, twintails, blue kimono, floral print, sash, hair flower, white thighhighs
Manjuu Mischief (Christmas):
Honolulu, HonoluluGift, long hair, twintails, green ribbon, ribbon bondage
Manjuu Mischief basically only works for the green wrapping. You might need to increase the weight on ribbon bondage like (ribbon bondage:1.2).
[Download](https://huggingface.co/ChameleonAI/ChameleonAILoras/blob/main/CHAR-Honolulu.safetensors)
|
hr-wesbeaver/test_model_1
|
hr-wesbeaver
| 2023-12-19T20:40:20Z | 7 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"distilbert",
"text-classification",
"generated_from_trainer",
"base_model:distilbert/distilbert-base-uncased",
"base_model:finetune:distilbert/distilbert-base-uncased",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2023-12-19T20:40:03Z |
---
license: apache-2.0
base_model: distilbert-base-uncased
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: test_model_1
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# test_model_1
This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4457
- Accuracy: 0.8989
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-06
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 15
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| No log | 1.0 | 125 | 0.3754 | 0.8949 |
| No log | 2.0 | 250 | 0.3674 | 0.8969 |
| No log | 3.0 | 375 | 0.3574 | 0.8999 |
| 0.1345 | 4.0 | 500 | 0.3786 | 0.9029 |
| 0.1345 | 5.0 | 625 | 0.3912 | 0.8979 |
| 0.1345 | 6.0 | 750 | 0.3946 | 0.8969 |
| 0.1345 | 7.0 | 875 | 0.4166 | 0.8959 |
| 0.0837 | 8.0 | 1000 | 0.4201 | 0.8969 |
| 0.0837 | 9.0 | 1125 | 0.4270 | 0.9029 |
| 0.0837 | 10.0 | 1250 | 0.4283 | 0.8969 |
| 0.0837 | 11.0 | 1375 | 0.4416 | 0.9039 |
| 0.0554 | 12.0 | 1500 | 0.4422 | 0.9019 |
| 0.0554 | 13.0 | 1625 | 0.4409 | 0.8989 |
| 0.0554 | 14.0 | 1750 | 0.4419 | 0.8989 |
| 0.0554 | 15.0 | 1875 | 0.4457 | 0.8989 |
### Framework versions
- Transformers 4.36.1
- Pytorch 2.1.2
- Datasets 2.15.0
- Tokenizers 0.15.0
|
r3dhummingbird/DialoGPT-small-neku
|
r3dhummingbird
| 2023-12-19T20:28:58Z | 11 | 0 |
transformers
|
[
"transformers",
"pytorch",
"safetensors",
"gpt2",
"text-generation",
"conversational",
"license:mit",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2022-03-02T23:29:05Z |
---
thumbnail: https://huggingface.co/front/thumbnails/dialogpt.png
tags:
- conversational
license: mit
---
# DialoGPT Trained on the Speech of a Game Character
This is an instance of [microsoft/DialoGPT-small](https://huggingface.co/microsoft/DialoGPT-small) trained on a game character, Neku Sakuraba from [The World Ends With You](https://en.wikipedia.org/wiki/The_World_Ends_with_You). The data comes from [a Kaggle game script dataset](https://www.kaggle.com/ruolinzheng/twewy-game-script).
Chat with the model:
```python
from transformers import AutoTokenizer, AutoModelWithLMHead
tokenizer = AutoTokenizer.from_pretrained("r3dhummingbird/DialoGPT-small-neku")
model = AutoModelWithLMHead.from_pretrained("r3dhummingbird/DialoGPT-small-neku")
# Let's chat for 4 lines
for step in range(4):
# encode the new user input, add the eos_token and return a tensor in Pytorch
new_user_input_ids = tokenizer.encode(input(">> User:") + tokenizer.eos_token, return_tensors='pt')
# print(new_user_input_ids)
# append the new user input tokens to the chat history
bot_input_ids = torch.cat([chat_history_ids, new_user_input_ids], dim=-1) if step > 0 else new_user_input_ids
# generated a response while limiting the total chat history to 1000 tokens,
chat_history_ids = model.generate(
bot_input_ids, max_length=200,
pad_token_id=tokenizer.eos_token_id,
no_repeat_ngram_size=3,
do_sample=True,
top_k=100,
top_p=0.7,
temperature=0.8
)
# pretty print last ouput tokens from bot
print("NekuBot: {}".format(tokenizer.decode(chat_history_ids[:, bot_input_ids.shape[-1]:][0], skip_special_tokens=True)))
```
|
ElenaaaKazakovaaa111111776654/bert-base-banking77-pt2
|
ElenaaaKazakovaaa111111776654
| 2023-12-19T20:28:13Z | 3 | 0 |
transformers
|
[
"transformers",
"pytorch",
"bert",
"text-classification",
"generated_from_trainer",
"dataset:banking77",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2023-12-19T20:23:32Z |
---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- banking77
model-index:
- name: bert-base-banking77-pt2
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# bert-base-banking77-pt2
This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the banking77 dataset.
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
### Framework versions
- Transformers 4.27.1
- Pytorch 2.1.2+cu121
- Datasets 2.9.0
- Tokenizers 0.13.3
|
livingbox/minimalistic-model-weights
|
livingbox
| 2023-12-19T20:25:13Z | 0 | 0 |
diffusers
|
[
"diffusers",
"text-to-image",
"stable-diffusion",
"license:creativeml-openrail-m",
"autotrain_compatible",
"endpoints_compatible",
"diffusers:StableDiffusionPipeline",
"region:us"
] |
text-to-image
| 2023-12-19T20:21:24Z |
---
license: creativeml-openrail-m
tags:
- text-to-image
- stable-diffusion
---
### minimalistic-model-weights Dreambooth model trained by livingbox with [TheLastBen's fast-DreamBooth](https://colab.research.google.com/github/TheLastBen/fast-stable-diffusion/blob/main/fast-DreamBooth.ipynb) notebook
Test the concept via A1111 Colab [fast-Colab-A1111](https://colab.research.google.com/github/TheLastBen/fast-stable-diffusion/blob/main/fast_stable_diffusion_AUTOMATIC1111.ipynb)
Sample pictures of this concept:
|
TheBloke/Swallow-70B-instruct-AWQ
|
TheBloke
| 2023-12-19T20:01:18Z | 10 | 1 |
transformers
|
[
"transformers",
"safetensors",
"llama",
"text-generation",
"en",
"ja",
"base_model:tokyotech-llm/Swallow-70b-instruct-hf",
"base_model:quantized:tokyotech-llm/Swallow-70b-instruct-hf",
"license:llama2",
"autotrain_compatible",
"text-generation-inference",
"4-bit",
"awq",
"region:us"
] |
text-generation
| 2023-12-19T18:37:03Z |
---
base_model: tokyotech-llm/Swallow-70b-instruct-hf
inference: false
language:
- en
- ja
library_name: transformers
license: llama2
model_creator: tokyotech-llm
model_name: Swallow 70B Instruct
model_type: llama
pipeline_tag: text-generation
prompt_template: "\u4EE5\u4E0B\u306B\u3001\u3042\u308B\u30BF\u30B9\u30AF\u3092\u8AAC\
\u660E\u3059\u308B\u6307\u793A\u304C\u3042\u308A\u307E\u3059\u3002\u30EA\u30AF\u30A8\
\u30B9\u30C8\u3092\u9069\u5207\u306B\u5B8C\u4E86\u3059\u308B\u305F\u3081\u306E\u56DE\
\u7B54\u3092\u8A18\u8FF0\u3057\u3066\u304F\u3060\u3055\u3044\u3002\\n\\n### \u6307\
\u793A:\\n{prompt}\\n\\n### \u5FDC\u7B54:\n"
quantized_by: TheBloke
---
<!-- markdownlint-disable MD041 -->
<!-- header start -->
<!-- 200823 -->
<div style="width: auto; margin-left: auto; margin-right: auto">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<div style="text-align:center; margin-top: 0em; margin-bottom: 0em"><p style="margin-top: 0.25em; margin-bottom: 0em;">TheBloke's LLM work is generously supported by a grant from <a href="https://a16z.com">andreessen horowitz (a16z)</a></p></div>
<hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
<!-- header end -->
# Swallow 70B Instruct - AWQ
- Model creator: [tokyotech-llm](https://huggingface.co/tokyotech-llm)
- Original model: [Swallow 70B Instruct](https://huggingface.co/tokyotech-llm/Swallow-70b-instruct-hf)
<!-- description start -->
## Description
This repo contains AWQ model files for [tokyotech-llm's Swallow 70B Instruct](https://huggingface.co/tokyotech-llm/Swallow-70b-instruct-hf).
### About AWQ
AWQ is an efficient, accurate and blazing-fast low-bit weight quantization method, currently supporting 4-bit quantization. Compared to GPTQ, it offers faster Transformers-based inference with equivalent or better quality compared to the most commonly used GPTQ settings.
AWQ models are currently supported on Linux and Windows, with NVidia GPUs only. macOS users: please use GGUF models instead.
It is supported by:
- [Text Generation Webui](https://github.com/oobabooga/text-generation-webui) - using Loader: AutoAWQ
- [vLLM](https://github.com/vllm-project/vllm) - version 0.2.2 or later for support for all model types.
- [Hugging Face Text Generation Inference (TGI)](https://github.com/huggingface/text-generation-inference)
- [Transformers](https://huggingface.co/docs/transformers) version 4.35.0 and later, from any code or client that supports Transformers
- [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) - for use from Python code
<!-- description end -->
<!-- repositories-available start -->
## Repositories available
* [AWQ model(s) for GPU inference.](https://huggingface.co/TheBloke/Swallow-70B-instruct-AWQ)
* [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/Swallow-70B-instruct-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/Swallow-70B-instruct-GGUF)
* [tokyotech-llm's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/tokyotech-llm/Swallow-70b-instruct-hf)
<!-- repositories-available end -->
<!-- prompt-template start -->
## Prompt template: Swallow-Instruct
```
以下に、あるタスクを説明する指示があります。リクエストを適切に完了するための回答を記述してください。\n\n### 指示:\n{prompt}\n\n### 応答:
```
<!-- prompt-template end -->
<!-- README_AWQ.md-provided-files start -->
## Provided files, and AWQ parameters
I currently release 128g GEMM models only. The addition of group_size 32 models, and GEMV kernel models, is being actively considered.
Models are released as sharded safetensors files.
| Branch | Bits | GS | AWQ Dataset | Seq Len | Size |
| ------ | ---- | -- | ----------- | ------- | ---- |
| [main](https://huggingface.co/TheBloke/Swallow-70B-instruct-AWQ/tree/main) | 4 | 128 | [Alpaca Japanese](https://huggingface.co/datasets/fujiki/japanese_alpaca_data/viewer/) | 4096 | 36.98 GB
<!-- README_AWQ.md-provided-files end -->
<!-- README_AWQ.md-text-generation-webui start -->
## How to easily download and use this model in [text-generation-webui](https://github.com/oobabooga/text-generation-webui)
Please make sure you're using the latest version of [text-generation-webui](https://github.com/oobabooga/text-generation-webui).
It is strongly recommended to use the text-generation-webui one-click-installers unless you're sure you know how to make a manual install.
1. Click the **Model tab**.
2. Under **Download custom model or LoRA**, enter `TheBloke/Swallow-70B-instruct-AWQ`.
3. Click **Download**.
4. The model will start downloading. Once it's finished it will say "Done".
5. In the top left, click the refresh icon next to **Model**.
6. In the **Model** dropdown, choose the model you just downloaded: `Swallow-70B-instruct-AWQ`
7. Select **Loader: AutoAWQ**.
8. Click Load, and the model will load and is now ready for use.
9. If you want any custom settings, set them and then click **Save settings for this model** followed by **Reload the Model** in the top right.
10. Once you're ready, click the **Text Generation** tab and enter a prompt to get started!
<!-- README_AWQ.md-text-generation-webui end -->
<!-- README_AWQ.md-use-from-vllm start -->
## Multi-user inference server: vLLM
Documentation on installing and using vLLM [can be found here](https://vllm.readthedocs.io/en/latest/).
- Please ensure you are using vLLM version 0.2 or later.
- When using vLLM as a server, pass the `--quantization awq` parameter.
For example:
```shell
python3 -m vllm.entrypoints.api_server --model TheBloke/Swallow-70B-instruct-AWQ --quantization awq --dtype auto
```
- When using vLLM from Python code, again set `quantization=awq`.
For example:
```python
from vllm import LLM, SamplingParams
prompts = [
"Tell me about AI",
"Write a story about llamas",
"What is 291 - 150?",
"How much wood would a woodchuck chuck if a woodchuck could chuck wood?",
]
prompt_template=f'''以下に、あるタスクを説明する指示があります。リクエストを適切に完了するための回答を記述してください。\n\n### 指示:\n{prompt}\n\n### 応答:
'''
prompts = [prompt_template.format(prompt=prompt) for prompt in prompts]
sampling_params = SamplingParams(temperature=0.8, top_p=0.95)
llm = LLM(model="TheBloke/Swallow-70B-instruct-AWQ", quantization="awq", dtype="auto")
outputs = llm.generate(prompts, sampling_params)
# Print the outputs.
for output in outputs:
prompt = output.prompt
generated_text = output.outputs[0].text
print(f"Prompt: {prompt!r}, Generated text: {generated_text!r}")
```
<!-- README_AWQ.md-use-from-vllm start -->
<!-- README_AWQ.md-use-from-tgi start -->
## Multi-user inference server: Hugging Face Text Generation Inference (TGI)
Use TGI version 1.1.0 or later. The official Docker container is: `ghcr.io/huggingface/text-generation-inference:1.1.0`
Example Docker parameters:
```shell
--model-id TheBloke/Swallow-70B-instruct-AWQ --port 3000 --quantize awq --max-input-length 3696 --max-total-tokens 4096 --max-batch-prefill-tokens 4096
```
Example Python code for interfacing with TGI (requires [huggingface-hub](https://github.com/huggingface/huggingface_hub) 0.17.0 or later):
```shell
pip3 install huggingface-hub
```
```python
from huggingface_hub import InferenceClient
endpoint_url = "https://your-endpoint-url-here"
prompt = "Tell me about AI"
prompt_template=f'''以下に、あるタスクを説明する指示があります。リクエストを適切に完了するための回答を記述してください。\n\n### 指示:\n{prompt}\n\n### 応答:
'''
client = InferenceClient(endpoint_url)
response = client.text_generation(prompt,
max_new_tokens=128,
do_sample=True,
temperature=0.7,
top_p=0.95,
top_k=40,
repetition_penalty=1.1)
print(f"Model output: ", response)
```
<!-- README_AWQ.md-use-from-tgi end -->
<!-- README_AWQ.md-use-from-python start -->
## Inference from Python code using Transformers
### Install the necessary packages
- Requires: [Transformers](https://huggingface.co/docs/transformers) 4.35.0 or later.
- Requires: [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) 0.1.6 or later.
```shell
pip3 install --upgrade "autoawq>=0.1.6" "transformers>=4.35.0"
```
Note that if you are using PyTorch 2.0.1, the above AutoAWQ command will automatically upgrade you to PyTorch 2.1.0.
If you are using CUDA 11.8 and wish to continue using PyTorch 2.0.1, instead run this command:
```shell
pip3 install https://github.com/casper-hansen/AutoAWQ/releases/download/v0.1.6/autoawq-0.1.6+cu118-cp310-cp310-linux_x86_64.whl
```
If you have problems installing [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) using the pre-built wheels, install it from source instead:
```shell
pip3 uninstall -y autoawq
git clone https://github.com/casper-hansen/AutoAWQ
cd AutoAWQ
pip3 install .
```
### Transformers example code (requires Transformers 4.35.0 and later)
```python
from transformers import AutoModelForCausalLM, AutoTokenizer, TextStreamer
model_name_or_path = "TheBloke/Swallow-70B-instruct-AWQ"
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path)
model = AutoModelForCausalLM.from_pretrained(
model_name_or_path,
low_cpu_mem_usage=True,
device_map="cuda:0"
)
# Using the text streamer to stream output one token at a time
streamer = TextStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True)
prompt = "Tell me about AI"
prompt_template=f'''以下に、あるタスクを説明する指示があります。リクエストを適切に完了するための回答を記述してください。\n\n### 指示:\n{prompt}\n\n### 応答:
'''
# Convert prompt to tokens
tokens = tokenizer(
prompt_template,
return_tensors='pt'
).input_ids.cuda()
generation_params = {
"do_sample": True,
"temperature": 0.7,
"top_p": 0.95,
"top_k": 40,
"max_new_tokens": 512,
"repetition_penalty": 1.1
}
# Generate streamed output, visible one token at a time
generation_output = model.generate(
tokens,
streamer=streamer,
**generation_params
)
# Generation without a streamer, which will include the prompt in the output
generation_output = model.generate(
tokens,
**generation_params
)
# Get the tokens from the output, decode them, print them
token_output = generation_output[0]
text_output = tokenizer.decode(token_output)
print("model.generate output: ", text_output)
# Inference is also possible via Transformers' pipeline
from transformers import pipeline
pipe = pipeline(
"text-generation",
model=model,
tokenizer=tokenizer,
**generation_params
)
pipe_output = pipe(prompt_template)[0]['generated_text']
print("pipeline output: ", pipe_output)
```
<!-- README_AWQ.md-use-from-python end -->
<!-- README_AWQ.md-compatibility start -->
## Compatibility
The files provided are tested to work with:
- [text-generation-webui](https://github.com/oobabooga/text-generation-webui) using `Loader: AutoAWQ`.
- [vLLM](https://github.com/vllm-project/vllm) version 0.2.0 and later.
- [Hugging Face Text Generation Inference (TGI)](https://github.com/huggingface/text-generation-inference) version 1.1.0 and later.
- [Transformers](https://huggingface.co/docs/transformers) version 4.35.0 and later.
- [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) version 0.1.1 and later.
<!-- README_AWQ.md-compatibility end -->
<!-- footer start -->
<!-- 200823 -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute
Thanks to the [chirper.ai](https://chirper.ai) team!
Thanks to Clay from [gpus.llm-utils.org](llm-utils)!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Aemon Algiz.
**Patreon special mentions**: Michael Levine, 阿明, Trailburnt, Nikolai Manek, John Detwiler, Randy H, Will Dee, Sebastain Graf, NimbleBox.ai, Eugene Pentland, Emad Mostaque, Ai Maven, Jim Angel, Jeff Scroggin, Michael Davis, Manuel Alberto Morcote, Stephen Murray, Robert, Justin Joy, Luke @flexchar, Brandon Frisco, Elijah Stavena, S_X, Dan Guido, Undi ., Komninos Chatzipapas, Shadi, theTransient, Lone Striker, Raven Klaugh, jjj, Cap'n Zoog, Michel-Marie MAUDET (LINAGORA), Matthew Berman, David, Fen Risland, Omer Bin Jawed, Luke Pendergrass, Kalila, OG, Erik Bjäreholt, Rooh Singh, Joseph William Delisle, Dan Lewis, TL, John Villwock, AzureBlack, Brad, Pedro Madruga, Caitlyn Gatomon, K, jinyuan sun, Mano Prime, Alex, Jeffrey Morgan, Alicia Loh, Illia Dulskyi, Chadd, transmissions 11, fincy, Rainer Wilmers, ReadyPlayerEmma, knownsqashed, Mandus, biorpg, Deo Leter, Brandon Phillips, SuperWojo, Sean Connelly, Iucharbius, Jack West, Harry Royden McLaughlin, Nicholas, terasurfer, Vitor Caleffi, Duane Dunston, Johann-Peter Hartmann, David Ziegler, Olakabola, Ken Nordquist, Trenton Dambrowitz, Tom X Nguyen, Vadim, Ajan Kanaga, Leonard Tan, Clay Pascal, Alexandros Triantafyllidis, JM33133, Xule, vamX, ya boyyy, subjectnull, Talal Aujan, Alps Aficionado, wassieverse, Ari Malik, James Bentley, Woland, Spencer Kim, Michael Dempsey, Fred von Graf, Elle, zynix, William Richards, Stanislav Ovsiannikov, Edmond Seymore, Jonathan Leane, Martin Kemka, usrbinkat, Enrico Ros
Thank you to all my generous patrons and donaters!
And thank you again to a16z for their generous grant.
<!-- footer end -->
# Original model card: tokyotech-llm's Swallow 70B Instruct
# Swallow
Our Swallow model has undergone continuous pre-training from the Llama 2 family, primarily with the addition of Japanese language data. The tuned versions use supervised fine-tuning (SFT).
Links to other models can be found in the index.
## Swallow Model Index
|Model|Swallow-hf|Swallow-instruct-hf|
|---|---|---|
|7B| [Link](https://huggingface.co/tokyotech-llm/Swallow-7b-hf) | [Link](https://huggingface.co/tokyotech-llm/Swallow-7b-instruct-hf)|
|13B| [Link](https://huggingface.co/tokyotech-llm/Swallow-13b-hf) | [Link](https://huggingface.co/tokyotech-llm/Swallow-13b-instruct-hf)|
|70B| [Link](https://huggingface.co/tokyotech-llm/Swallow-70b-hf) | [Link](https://huggingface.co/tokyotech-llm/Swallow-70b-instruct-hf)|
This repository provides large language models developed by [TokyoTech-LLM](https://tokyotech-llm.github.io/).
Read our [blog post](https://zenn.dev/tokyotech_lm/articles/d6cb3a8fdfc907) or our paper (preprint coming soon) for more details!
## Model Details
* **Model type**: Please refer to LLaMA-2 technical report for details on the model architecture.
* **Language(s)**: Japanese English
* **Library**: [Megatron-LM](https://github.com/rioyokotalab/Megatron-Llama2)
* **Tokenizer**: This model employs a tokenizer that features a broadened vocabulary based on Japanese data. This allows for a more efficient representation of text using fewer tokens, leading to a notably faster inference process.
* **Contact**: swallow[at]nlp.c.titech.ac.jp
## Base Model Performance
### Japanese version
|Model|Size|JCommonsenseQA|JEMHopQA|NIILC|JSQuAD|XL-Sum|MGSM|WMT20-en-ja|WMT20-ja-en|
|---|---|---|---|---|---|---|---|---|---|
| | |4-shot|4-shot|4-shot|4-shot|1-shot|4-shot|4-shot|4-shot|
|Llama 2|7B|0.3852|0.4240|0.3410|0.7917|0.1905|0.0760|0.1783|0.1738|
|Swallow|7B|0.4808|0.5078|0.5968|0.8573|0.1830|0.1240|0.2510|0.1511|
|Llama 2|13B|0.6997|0.4415|0.4170|0.8533|0.2139|0.1320|0.2146|0.1982|
|Swallow|13B|0.7837|0.5063|0.6398|0.9005|0.2168|0.2040|0.2720|0.1771|
|Llama 2|70B|0.8686|0.4656|0.5256|0.9080|**0.2361**|0.3560|0.2643|**0.2398**|
|Swallow|70B|**0.9348**|**0.6290**|**0.6960**|**0.9176**|0.2266|**0.4840**|**0.3043**|0.2298|
## Usage
First install additional dependencies in [requirements.txt](./requirements.txt):
```sh
pip install -r requirements.txt
```
### Use the instruct model
```python
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
model_name = "tokyotech-llm/Swallow-7b-instruct-hf"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.bfloat16, low_cpu_mem_usage=True, device_map="auto")
PROMPT_DICT = {
"prompt_input": (
"以下に、あるタスクを説明する指示があり、それに付随する入力が更なる文脈を提供しています。"
"リクエストを適切に完了するための回答を記述してください。\n\n"
"### 指示:\n{instruction}\n\n### 入力:\n{input}\n\n### 応答:"
),
"prompt_no_input": (
"以下に、あるタスクを説明する指示があります。"
"リクエストを適切に完了するための回答を記述してください。\n\n"
"### 指示:\n{instruction}\n\n### 応答:"
),
}
def create_prompt(instruction, input=None):
"""
Generates a prompt based on the given instruction and an optional input.
If input is provided, it uses the 'prompt_input' template from PROMPT_DICT.
If no input is provided, it uses the 'prompt_no_input' template.
Args:
instruction (str): The instruction describing the task.
input (str, optional): Additional input providing context for the task. Default is None.
Returns:
str: The generated prompt.
"""
if input:
# Use the 'prompt_input' template when additional input is provided
return PROMPT_DICT["prompt_input"].format(instruction=instruction, input=input)
else:
# Use the 'prompt_no_input' template when no additional input is provided
return PROMPT_DICT["prompt_no_input"].format(instruction=instruction)
# Example usage
instruction_example = "以下のトピックに関する詳細な情報を提供してください。"
input_example = "東京工業大学の主なキャンパスについて教えてください"
prompt = create_prompt(instruction_example, input_example)
input_ids = tokenizer.encode(
prompt,
add_special_tokens=False,
return_tensors="pt"
)
tokens = model.generate(
input_ids.to(device=model.device),
max_new_tokens=128,
temperature=0.99,
top_p=0.95,
do_sample=True,
)
out = tokenizer.decode(tokens[0], skip_special_tokens=True)
print(out)
```
### Use the base model
```python
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
model_name = "tokyotech-llm/Swallow-7b-hf"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.bfloat16, device_map="auto")
prompt = "東京工業大学の主なキャンパスは、"
input_ids = tokenizer.encode(
prompt,
add_special_tokens=False,
return_tensors="pt"
)
tokens = model.generate(
input_ids.to(device=model.device),
max_new_tokens=128,
temperature=0.99,
top_p=0.95,
do_sample=True,
)
out = tokenizer.decode(tokens[0], skip_special_tokens=True)
print(out)
```
## Training Datasets
### Continual Pre-Training
The following datasets were used for continual pre-training.
- [Japanese Wikipedia](https://dumps.wikimedia.org/other/cirrussearch)
- [RefinedWeb](https://huggingface.co/datasets/tiiuae/falcon-refinedweb)
- Swallow Corpus
- [The Pile](https://huggingface.co/datasets/EleutherAI/pile)
### Instruction Tuning
The following datasets were used for the instruction tuning.
- [Anthropic HH-RLHF](https://huggingface.co/datasets/kunishou/hh-rlhf-49k-ja)
- [Databricks Dolly 15-k](https://huggingface.co/datasets/kunishou/databricks-dolly-15k-ja)
- [OpenAssistant Conversations Dataset](https://huggingface.co/datasets/kunishou/oasst1-89k-ja)
## Risks and Limitations
The models released here are still in the early stages of our research and development and have not been tuned to ensure outputs align with human intent and safety considerations.
## Acknowledgements
We thank Meta Research for releasing Llama 2 under an open license for others to build on.
Our project is supported by the [ABCI Large-scale Language Model Building Support Program](https://abci.ai/en/link/llm_support_program.html) of the National Institute of Advanced Industrial Science and Technology.
## License
Llama 2 is licensed under the LLAMA 2 Community License, Copyright © Meta Platforms, Inc. All Rights Reserved.
## Authors
Here are the team members:
- From [Okazaki Laboratory](https://www.nlp.c.titech.ac.jp/index.en.html), the following members:
- [Naoaki Okazaki](https://www.chokkan.org/index.ja.html)
- [Sakae Mizuki](https://s-mizuki-nlp.github.io/)
- [Hiroki Iida](https://meshidenn.github.io/)
- [Mengsay Loem](https://loem-ms.github.io/)
- [Shota Hirai](https://huggingface.co/Kotemo428)
- [Kakeru Hattori](https://aya-se.vercel.app/)
- [Masanari Ohi](https://twitter.com/stjohn2007)
- From [YOKOTA Laboratory](https://www.rio.gsic.titech.ac.jp/en/index.html), the following members:
- [Rio Yokota](https://twitter.com/rioyokota)
- [Kazuki Fujii](https://twitter.com/okoge_kaz)
- [Taishi Nakamura](https://twitter.com/Setuna7777_2)
|
MichalGas/ycjn-unff-povv-0
|
MichalGas
| 2023-12-19T19:49:12Z | 30 | 0 |
transformers
|
[
"transformers",
"safetensors",
"resnet",
"image-classification",
"autotrain",
"dataset:MichalGas/autotrain-data-ycjn-unff-povv",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
image-classification
| 2023-12-19T19:49:08Z |
---
tags:
- autotrain
- image-classification
widget:
- src: https://huggingface.co/datasets/mishig/sample_images/resolve/main/tiger.jpg
example_title: Tiger
- src: https://huggingface.co/datasets/mishig/sample_images/resolve/main/teapot.jpg
example_title: Teapot
- src: https://huggingface.co/datasets/mishig/sample_images/resolve/main/palace.jpg
example_title: Palace
datasets:
- MichalGas/autotrain-data-ycjn-unff-povv
---
# Model Trained Using AutoTrain
- Problem type: Image Classification
## Validation Metricsg
loss: nan
f1_macro: 0.05573770491803279
f1_micro: 0.20078740157480315
f1_weighted: 0.06714857364140958
precision_macro: 0.03346456692913386
precision_micro: 0.20078740157480315
precision_weighted: 0.040315580631161266
recall_macro: 0.16666666666666666
recall_micro: 0.20078740157480315
recall_weighted: 0.20078740157480315
accuracy: 0.20078740157480315
|
Hanaamiri/flan-t5-small-samsum
|
Hanaamiri
| 2023-12-19T19:48:43Z | 11 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"t5",
"text2text-generation",
"generated_from_trainer",
"dataset:samsum",
"base_model:google/flan-t5-small",
"base_model:finetune:google/flan-t5-small",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2023-12-19T19:23:56Z |
---
license: apache-2.0
base_model: google/flan-t5-small
tags:
- generated_from_trainer
datasets:
- samsum
metrics:
- rouge
model-index:
- name: flan-t5-small-samsum
results:
- task:
name: Sequence-to-sequence Language Modeling
type: text2text-generation
dataset:
name: samsum
type: samsum
config: samsum
split: test
args: samsum
metrics:
- name: Rouge1
type: rouge
value: 42.4655
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# flan-t5-small-samsum
This model is a fine-tuned version of [google/flan-t5-small](https://huggingface.co/google/flan-t5-small) on the samsum dataset.
It achieves the following results on the evaluation set:
- Loss: 1.6732
- Rouge1: 42.4655
- Rouge2: 18.4875
- Rougel: 35.2198
- Rougelsum: 38.6465
- Gen Len: 16.8486
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:-------:|
| 1.8853 | 0.22 | 100 | 1.7046 | 42.3969 | 18.365 | 35.0091 | 38.6527 | 16.6703 |
| 1.8463 | 0.43 | 200 | 1.6954 | 42.5607 | 18.4425 | 35.1088 | 38.8749 | 17.3565 |
| 1.8549 | 0.65 | 300 | 1.6794 | 42.5148 | 18.4716 | 35.1769 | 38.7018 | 17.1123 |
| 1.8361 | 0.87 | 400 | 1.6775 | 42.3899 | 18.4343 | 35.134 | 38.5732 | 16.6215 |
| 1.8132 | 1.08 | 500 | 1.6732 | 42.4655 | 18.4875 | 35.2198 | 38.6465 | 16.8486 |
| 1.8073 | 1.3 | 600 | 1.6708 | 42.4741 | 18.3824 | 35.1819 | 38.6066 | 16.9475 |
| 1.7973 | 1.52 | 700 | 1.6686 | 42.8206 | 18.7011 | 35.3874 | 38.9173 | 16.7595 |
| 1.798 | 1.74 | 800 | 1.6666 | 42.7779 | 18.6627 | 35.323 | 38.9467 | 16.9389 |
| 1.79 | 1.95 | 900 | 1.6668 | 42.8071 | 18.7113 | 35.2872 | 38.8641 | 16.9426 |
### Framework versions
- Transformers 4.36.0
- Pytorch 2.0.0
- Datasets 2.15.0
- Tokenizers 0.15.0
|
gsoaresbaptista/themis-instruct
|
gsoaresbaptista
| 2023-12-19T19:41:02Z | 0 | 0 |
peft
|
[
"peft",
"safetensors",
"generated_from_trainer",
"license:llama2",
"region:us"
] | null | 2023-12-19T00:45:46Z |
---
license: llama2
library_name: peft
tags:
- generated_from_trainer
base_model: dominguesm/canarim-7b-instruct
model-index:
- name: themis-instruct
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# themis-instruct
This model is a fine-tuned version of [dominguesm/canarim-7b-instruct](https://huggingface.co/dominguesm/canarim-7b-instruct) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.8855
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2.5e-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 1
- num_epochs: 1
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 2.0042 | 0.03 | 25 | 1.3872 |
| 1.2414 | 0.05 | 50 | 1.1247 |
| 1.0936 | 0.08 | 75 | 1.0210 |
| 1.0607 | 0.11 | 100 | 0.9881 |
| 1.0186 | 0.13 | 125 | 0.9693 |
| 0.9851 | 0.16 | 150 | 0.9552 |
| 1.0123 | 0.18 | 175 | 0.9469 |
| 0.9737 | 0.21 | 200 | 0.9398 |
| 0.9408 | 0.24 | 225 | 0.9371 |
| 0.9129 | 0.26 | 250 | 0.9300 |
| 0.9441 | 0.29 | 275 | 0.9254 |
| 0.9577 | 0.32 | 300 | 0.9198 |
| 0.9509 | 0.34 | 325 | 0.9162 |
| 0.9023 | 0.37 | 350 | 0.9135 |
| 0.8924 | 0.4 | 375 | 0.9109 |
| 0.9207 | 0.42 | 400 | 0.9086 |
| 0.9436 | 0.45 | 425 | 0.9058 |
| 0.8637 | 0.47 | 450 | 0.9047 |
| 0.9207 | 0.5 | 475 | 0.9033 |
| 0.9475 | 0.53 | 500 | 0.9002 |
| 0.9548 | 0.55 | 525 | 0.8981 |
| 0.8806 | 0.58 | 550 | 0.8969 |
| 0.9475 | 0.61 | 575 | 0.8949 |
| 0.8505 | 0.63 | 600 | 0.8932 |
| 0.8999 | 0.66 | 625 | 0.8926 |
| 0.9018 | 0.68 | 650 | 0.8906 |
| 0.9107 | 0.71 | 675 | 0.8901 |
| 0.8557 | 0.74 | 700 | 0.8888 |
| 0.8903 | 0.76 | 725 | 0.8881 |
| 0.8718 | 0.79 | 750 | 0.8875 |
| 0.9002 | 0.82 | 775 | 0.8870 |
| 0.9086 | 0.84 | 800 | 0.8867 |
| 0.8983 | 0.87 | 825 | 0.8863 |
| 0.9401 | 0.9 | 850 | 0.8861 |
| 0.9434 | 0.92 | 875 | 0.8857 |
| 0.8987 | 0.95 | 900 | 0.8855 |
| 0.9008 | 0.97 | 925 | 0.8855 |
### Framework versions
- PEFT 0.7.2.dev0
- Transformers 4.37.0.dev0
- Pytorch 2.1.0+cu121
- Datasets 2.15.0
- Tokenizers 0.15.0
|
TheBloke/Yi-34B-200K-DARE-merge-v5-AWQ
|
TheBloke
| 2023-12-19T19:40:51Z | 8 | 0 |
transformers
|
[
"transformers",
"safetensors",
"llama",
"text-generation",
"text-generation-inference",
"merge",
"en",
"base_model:brucethemoose/Yi-34B-200K-DARE-merge-v5",
"base_model:quantized:brucethemoose/Yi-34B-200K-DARE-merge-v5",
"license:other",
"autotrain_compatible",
"4-bit",
"awq",
"region:us"
] |
text-generation
| 2023-12-19T18:28:15Z |
---
base_model: brucethemoose/Yi-34B-200K-DARE-merge-v5
inference: false
language:
- en
library_name: transformers
license: other
license_link: https://huggingface.co/01-ai/Yi-34B/blob/main/LICENSE
license_name: yi-license
model_creator: brucethemoose
model_name: Yi 34B 200K DARE Merge v5
model_type: yi
pipeline_tag: text-generation
prompt_template: 'SYSTEM: {system_message}
USER: {prompt}
ASSISTANT:
'
quantized_by: TheBloke
tags:
- text-generation-inference
- merge
---
<!-- markdownlint-disable MD041 -->
<!-- header start -->
<!-- 200823 -->
<div style="width: auto; margin-left: auto; margin-right: auto">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<div style="text-align:center; margin-top: 0em; margin-bottom: 0em"><p style="margin-top: 0.25em; margin-bottom: 0em;">TheBloke's LLM work is generously supported by a grant from <a href="https://a16z.com">andreessen horowitz (a16z)</a></p></div>
<hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
<!-- header end -->
# Yi 34B 200K DARE Merge v5 - AWQ
- Model creator: [brucethemoose](https://huggingface.co/brucethemoose)
- Original model: [Yi 34B 200K DARE Merge v5](https://huggingface.co/brucethemoose/Yi-34B-200K-DARE-merge-v5)
<!-- description start -->
## Description
This repo contains AWQ model files for [brucethemoose's Yi 34B 200K DARE Merge v5](https://huggingface.co/brucethemoose/Yi-34B-200K-DARE-merge-v5).
These files were quantised using hardware kindly provided by [Massed Compute](https://massedcompute.com/).
### About AWQ
AWQ is an efficient, accurate and blazing-fast low-bit weight quantization method, currently supporting 4-bit quantization. Compared to GPTQ, it offers faster Transformers-based inference with equivalent or better quality compared to the most commonly used GPTQ settings.
AWQ models are currently supported on Linux and Windows, with NVidia GPUs only. macOS users: please use GGUF models instead.
It is supported by:
- [Text Generation Webui](https://github.com/oobabooga/text-generation-webui) - using Loader: AutoAWQ
- [vLLM](https://github.com/vllm-project/vllm) - version 0.2.2 or later for support for all model types.
- [Hugging Face Text Generation Inference (TGI)](https://github.com/huggingface/text-generation-inference)
- [Transformers](https://huggingface.co/docs/transformers) version 4.35.0 and later, from any code or client that supports Transformers
- [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) - for use from Python code
<!-- description end -->
<!-- repositories-available start -->
## Repositories available
* [AWQ model(s) for GPU inference.](https://huggingface.co/TheBloke/Yi-34B-200K-DARE-merge-v5-AWQ)
* [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/Yi-34B-200K-DARE-merge-v5-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/Yi-34B-200K-DARE-merge-v5-GGUF)
* [brucethemoose's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/brucethemoose/Yi-34B-200K-DARE-merge-v5)
<!-- repositories-available end -->
<!-- prompt-template start -->
## Prompt template: Orca-Vicuna
```
SYSTEM: {system_message}
USER: {prompt}
ASSISTANT:
```
<!-- prompt-template end -->
<!-- README_AWQ.md-provided-files start -->
## Provided files, and AWQ parameters
I currently release 128g GEMM models only. The addition of group_size 32 models, and GEMV kernel models, is being actively considered.
Models are released as sharded safetensors files.
| Branch | Bits | GS | AWQ Dataset | Seq Len | Size |
| ------ | ---- | -- | ----------- | ------- | ---- |
| [main](https://huggingface.co/TheBloke/Yi-34B-200K-DARE-merge-v5-AWQ/tree/main) | 4 | 128 | [VMware Open Instruct](https://huggingface.co/datasets/VMware/open-instruct/viewer/) | 8192 | 19.23 GB
<!-- README_AWQ.md-provided-files end -->
<!-- README_AWQ.md-text-generation-webui start -->
## How to easily download and use this model in [text-generation-webui](https://github.com/oobabooga/text-generation-webui)
Please make sure you're using the latest version of [text-generation-webui](https://github.com/oobabooga/text-generation-webui).
It is strongly recommended to use the text-generation-webui one-click-installers unless you're sure you know how to make a manual install.
1. Click the **Model tab**.
2. Under **Download custom model or LoRA**, enter `TheBloke/Yi-34B-200K-DARE-merge-v5-AWQ`.
3. Click **Download**.
4. The model will start downloading. Once it's finished it will say "Done".
5. In the top left, click the refresh icon next to **Model**.
6. In the **Model** dropdown, choose the model you just downloaded: `Yi-34B-200K-DARE-merge-v5-AWQ`
7. Select **Loader: AutoAWQ**.
8. Click Load, and the model will load and is now ready for use.
9. If you want any custom settings, set them and then click **Save settings for this model** followed by **Reload the Model** in the top right.
10. Once you're ready, click the **Text Generation** tab and enter a prompt to get started!
<!-- README_AWQ.md-text-generation-webui end -->
<!-- README_AWQ.md-use-from-vllm start -->
## Multi-user inference server: vLLM
Documentation on installing and using vLLM [can be found here](https://vllm.readthedocs.io/en/latest/).
- Please ensure you are using vLLM version 0.2 or later.
- When using vLLM as a server, pass the `--quantization awq` parameter.
For example:
```shell
python3 -m vllm.entrypoints.api_server --model TheBloke/Yi-34B-200K-DARE-merge-v5-AWQ --quantization awq --dtype auto
```
- When using vLLM from Python code, again set `quantization=awq`.
For example:
```python
from vllm import LLM, SamplingParams
prompts = [
"Tell me about AI",
"Write a story about llamas",
"What is 291 - 150?",
"How much wood would a woodchuck chuck if a woodchuck could chuck wood?",
]
prompt_template=f'''SYSTEM: {system_message}
USER: {prompt}
ASSISTANT:
'''
prompts = [prompt_template.format(prompt=prompt) for prompt in prompts]
sampling_params = SamplingParams(temperature=0.8, top_p=0.95)
llm = LLM(model="TheBloke/Yi-34B-200K-DARE-merge-v5-AWQ", quantization="awq", dtype="auto")
outputs = llm.generate(prompts, sampling_params)
# Print the outputs.
for output in outputs:
prompt = output.prompt
generated_text = output.outputs[0].text
print(f"Prompt: {prompt!r}, Generated text: {generated_text!r}")
```
<!-- README_AWQ.md-use-from-vllm start -->
<!-- README_AWQ.md-use-from-tgi start -->
## Multi-user inference server: Hugging Face Text Generation Inference (TGI)
Use TGI version 1.1.0 or later. The official Docker container is: `ghcr.io/huggingface/text-generation-inference:1.1.0`
Example Docker parameters:
```shell
--model-id TheBloke/Yi-34B-200K-DARE-merge-v5-AWQ --port 3000 --quantize awq --max-input-length 3696 --max-total-tokens 4096 --max-batch-prefill-tokens 4096
```
Example Python code for interfacing with TGI (requires [huggingface-hub](https://github.com/huggingface/huggingface_hub) 0.17.0 or later):
```shell
pip3 install huggingface-hub
```
```python
from huggingface_hub import InferenceClient
endpoint_url = "https://your-endpoint-url-here"
prompt = "Tell me about AI"
prompt_template=f'''SYSTEM: {system_message}
USER: {prompt}
ASSISTANT:
'''
client = InferenceClient(endpoint_url)
response = client.text_generation(prompt,
max_new_tokens=128,
do_sample=True,
temperature=0.7,
top_p=0.95,
top_k=40,
repetition_penalty=1.1)
print(f"Model output: ", response)
```
<!-- README_AWQ.md-use-from-tgi end -->
<!-- README_AWQ.md-use-from-python start -->
## Inference from Python code using Transformers
### Install the necessary packages
- Requires: [Transformers](https://huggingface.co/docs/transformers) 4.35.0 or later.
- Requires: [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) 0.1.6 or later.
```shell
pip3 install --upgrade "autoawq>=0.1.6" "transformers>=4.35.0"
```
Note that if you are using PyTorch 2.0.1, the above AutoAWQ command will automatically upgrade you to PyTorch 2.1.0.
If you are using CUDA 11.8 and wish to continue using PyTorch 2.0.1, instead run this command:
```shell
pip3 install https://github.com/casper-hansen/AutoAWQ/releases/download/v0.1.6/autoawq-0.1.6+cu118-cp310-cp310-linux_x86_64.whl
```
If you have problems installing [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) using the pre-built wheels, install it from source instead:
```shell
pip3 uninstall -y autoawq
git clone https://github.com/casper-hansen/AutoAWQ
cd AutoAWQ
pip3 install .
```
### Transformers example code (requires Transformers 4.35.0 and later)
```python
from transformers import AutoModelForCausalLM, AutoTokenizer, TextStreamer
model_name_or_path = "TheBloke/Yi-34B-200K-DARE-merge-v5-AWQ"
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path)
model = AutoModelForCausalLM.from_pretrained(
model_name_or_path,
low_cpu_mem_usage=True,
device_map="cuda:0"
)
# Using the text streamer to stream output one token at a time
streamer = TextStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True)
prompt = "Tell me about AI"
prompt_template=f'''SYSTEM: {system_message}
USER: {prompt}
ASSISTANT:
'''
# Convert prompt to tokens
tokens = tokenizer(
prompt_template,
return_tensors='pt'
).input_ids.cuda()
generation_params = {
"do_sample": True,
"temperature": 0.7,
"top_p": 0.95,
"top_k": 40,
"max_new_tokens": 512,
"repetition_penalty": 1.1
}
# Generate streamed output, visible one token at a time
generation_output = model.generate(
tokens,
streamer=streamer,
**generation_params
)
# Generation without a streamer, which will include the prompt in the output
generation_output = model.generate(
tokens,
**generation_params
)
# Get the tokens from the output, decode them, print them
token_output = generation_output[0]
text_output = tokenizer.decode(token_output)
print("model.generate output: ", text_output)
# Inference is also possible via Transformers' pipeline
from transformers import pipeline
pipe = pipeline(
"text-generation",
model=model,
tokenizer=tokenizer,
**generation_params
)
pipe_output = pipe(prompt_template)[0]['generated_text']
print("pipeline output: ", pipe_output)
```
<!-- README_AWQ.md-use-from-python end -->
<!-- README_AWQ.md-compatibility start -->
## Compatibility
The files provided are tested to work with:
- [text-generation-webui](https://github.com/oobabooga/text-generation-webui) using `Loader: AutoAWQ`.
- [vLLM](https://github.com/vllm-project/vllm) version 0.2.0 and later.
- [Hugging Face Text Generation Inference (TGI)](https://github.com/huggingface/text-generation-inference) version 1.1.0 and later.
- [Transformers](https://huggingface.co/docs/transformers) version 4.35.0 and later.
- [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) version 0.1.1 and later.
<!-- README_AWQ.md-compatibility end -->
<!-- footer start -->
<!-- 200823 -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute
Thanks to the [chirper.ai](https://chirper.ai) team!
Thanks to Clay from [gpus.llm-utils.org](llm-utils)!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Aemon Algiz.
**Patreon special mentions**: Michael Levine, 阿明, Trailburnt, Nikolai Manek, John Detwiler, Randy H, Will Dee, Sebastain Graf, NimbleBox.ai, Eugene Pentland, Emad Mostaque, Ai Maven, Jim Angel, Jeff Scroggin, Michael Davis, Manuel Alberto Morcote, Stephen Murray, Robert, Justin Joy, Luke @flexchar, Brandon Frisco, Elijah Stavena, S_X, Dan Guido, Undi ., Komninos Chatzipapas, Shadi, theTransient, Lone Striker, Raven Klaugh, jjj, Cap'n Zoog, Michel-Marie MAUDET (LINAGORA), Matthew Berman, David, Fen Risland, Omer Bin Jawed, Luke Pendergrass, Kalila, OG, Erik Bjäreholt, Rooh Singh, Joseph William Delisle, Dan Lewis, TL, John Villwock, AzureBlack, Brad, Pedro Madruga, Caitlyn Gatomon, K, jinyuan sun, Mano Prime, Alex, Jeffrey Morgan, Alicia Loh, Illia Dulskyi, Chadd, transmissions 11, fincy, Rainer Wilmers, ReadyPlayerEmma, knownsqashed, Mandus, biorpg, Deo Leter, Brandon Phillips, SuperWojo, Sean Connelly, Iucharbius, Jack West, Harry Royden McLaughlin, Nicholas, terasurfer, Vitor Caleffi, Duane Dunston, Johann-Peter Hartmann, David Ziegler, Olakabola, Ken Nordquist, Trenton Dambrowitz, Tom X Nguyen, Vadim, Ajan Kanaga, Leonard Tan, Clay Pascal, Alexandros Triantafyllidis, JM33133, Xule, vamX, ya boyyy, subjectnull, Talal Aujan, Alps Aficionado, wassieverse, Ari Malik, James Bentley, Woland, Spencer Kim, Michael Dempsey, Fred von Graf, Elle, zynix, William Richards, Stanislav Ovsiannikov, Edmond Seymore, Jonathan Leane, Martin Kemka, usrbinkat, Enrico Ros
Thank you to all my generous patrons and donaters!
And thank you again to a16z for their generous grant.
<!-- footer end -->
# Original model card: brucethemoose's Yi 34B 200K DARE Merge v5
[**Nous-Capybara-34B**](https://huggingface.co/NousResearch/Nous-Capybara-34B/), [**Tess-M-v1.4**](https://huggingface.co/migtissera/Tess-34B-v1.4), [**Airoboros-3_1-yi-34b-200k**](https://huggingface.co/bhenrym14/airoboros-3_1-yi-34b-200k), [**PlatYi-34B-200K-Q**](https://huggingface.co/kyujinpy/PlatYi-34B-200k-Q-FastChat), [**Pallas-0.4**](https://huggingface.co/Mihaiii/Pallas-0.4), [**Yi-34B-200K-AEZAKMI-v2**](https://huggingface.co/adamo1139/Yi-34B-200K-AEZAKMI-v2), and a tiny bit of [**SUS-Chat-34B**](https://huggingface.co/SUSTech/SUS-Chat-34B) merged with a new, experimental implementation of "dare ties" via mergekit. See:
> [Language Models are Super Mario: Absorbing Abilities from Homologous Models as a Free Lunch](https://github.com/yule-BUAA/MergeLM)
> https://github.com/cg123/mergekit/tree/dare
***
## Prompt template: Orca-Vicuna
```
SYSTEM: {system_message}
USER: {prompt}
ASSISTANT:
```
It might recognize ChatML, or maybe Llama-chat from Airoboros.
Sometimes the model "spells out" the stop token as `</s>` like Capybara, so you may need to add `</s>` as an additional stopping condition.
***
## Running
Being a Yi model, try running a lower temperature with 0.05-0.1 MinP, a little repetition penalty, and no other samplers. Yi tends to run "hot" by default, and it really needs MinP to cull the huge vocabulary.
24GB GPUs can run Yi-34B-200K models at **45K-75K context** with exllamav2, and performant UIs like [exui](https://github.com/turboderp/exui). I go into more detail in this [post](https://old.reddit.com/r/LocalLLaMA/comments/1896igc/how_i_run_34b_models_at_75k_context_on_24gb_fast/)
I recommend exl2 quantizations profiled on data similar to the desired task. It is especially sensitive to the quantization data at low bpw. I've published my own fiction-oriented quantizations here: https://huggingface.co/collections/brucethemoose/most-recent-merge-65742644ca03b6c514afa204
To load this in full-context backends like transformers, you *must* change `max_position_embeddings` in config.json to a lower value than 200,000, otherwise you will OOM!
***
## Testing Notes
Merged in mergekit with the following config, and the tokenizer from chargoddard's Yi-Llama:
```
models:
- model: /home/alpha/Storage/Models/Raw/chargoddard_Yi-34B-200K-Llama
# No parameters necessary for base model
- model: /home/alpha/Storage/Models/Raw/migtissera_Tess-34B-v1.4
# Less weight than previous merge since Pallas is a finetune of Tess
parameters:
weight: 0.14
density: 0.62
- model: /home/alpha/FastModels/Mihaiii_Pallas-0.4
parameters:
weight: 0.14
density: 0.62
- model: /home/alpha//Storage/Models/Raw/bhenrym14_airoboros-3_1-yi-34b-200k
parameters:
weight: 0.14
density: 0.52
- model: /home/alpha/Storage/Models/Raw/Nous-Capybara-34B
parameters:
weight: 0.22
density: 0.62
- model: /home/alpha/Storage/Models/Raw/kyujinpy_PlatYi-34B-200k-Q-FastChat
parameters:
weight: 0.14
density: 0.52
#- model: /home/alpha/Storage/Models/Raw/ehartford_dolphin-2.2-yi-34b-200k
# Dolphin 200K seems to be broken according to multiple leaderboards and perplexity tests?
# parameters:
# weight: 0.15
# density: 0.6
- model: /home/alpha/Models/Raw/adamo1139_Yi-34B-200K-AEZAKMI-v2
parameters:
weight: 0.14
density: 0.52
- model: /home/alpha/Models/Raw/SUSTech_SUS-Chat-34B/
# Very low density and low weight since its a Yi 4K finetune, to try and preserve long context performance while "keeping" some of SUS
parameters:
weight: 0.08
density: 0.08
merge_method: dare_ties
base_model: /home/alpha/Storage/Models/Raw/chargoddard_Yi-34B-200K-Llama
parameters:
int8_mask: true
dtype: bfloat16
```
Various densities were tested with perplexity tests and long context prompts. Relatively high densities seem to perform better, contrary to the findings of the Super Mario paper.
This particular version is merged with more than the "recommended" max density of 0.5. It seems to result in even better perplexity, but I'm not sure if this translates to better output.
Weights that add up to 1 seems to be optimal.
Dare Ties is also resulting in seemingly better, lower perplexity merges than a regular ties merge, task arithmetic or a slerp merge.
SUS Chat is not a 200K model, hence it was merged at a very low density to try and preserve Yi 200K's long context performance while still inheriting some of SUS's performance.
Dolphin 200K was taken out of this merge because it seems to be performing poorly for a 34B Dolphin model, like something went wrong during training?
I chose not to include other finetunes because they aren't trained on the 200K base. If any other 200K finetunes pop up, let me know.
***
## Credits:
https://github.com/cg123/mergekit/tree/dare
https://huggingface.co/NousResearch/Nous-Capybara-34B/
https://huggingface.co/bhenrym14/airoboros-3_1-yi-34b-200k
https://huggingface.co/migtissera/Tess-M-v1.4
https://huggingface.co/kyujinpy/PlatYi-34B-200k-Q-FastChat
https://huggingface.co/adamo1139/Yi-34B-200K-AEZAKMI-v2
https://huggingface.co/Mihaiii/Pallas-0.4
https://huggingface.co/SUSTech/SUS-Chat-34B
https://huggingface.co/chargoddard/Yi-34B-200K-Llama
https://huggingface.co/01-ai/Yi-34B-200K
|
ali-khoshtinat/mlm
|
ali-khoshtinat
| 2023-12-19T19:38:32Z | 4 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"roberta",
"fill-mask",
"generated_from_trainer",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2023-12-19T19:38:19Z |
---
tags:
- generated_from_trainer
model-index:
- name: mlm
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# mlm
This model is a fine-tuned version of [](https://huggingface.co/) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 4.5771
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0001
- train_batch_size: 64
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 200
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 6.0866 | 10.87 | 500 | 5.8241 |
| 5.1736 | 21.74 | 1000 | 5.3014 |
| 4.5378 | 32.61 | 1500 | 4.9272 |
| 4.019 | 43.48 | 2000 | 4.8975 |
| 3.595 | 54.35 | 2500 | 4.4323 |
| 3.2182 | 65.22 | 3000 | 4.5295 |
| 2.887 | 76.09 | 3500 | 4.6930 |
| 2.596 | 86.96 | 4000 | 4.5771 |
### Framework versions
- Transformers 4.35.2
- Pytorch 2.1.0+cu121
- Datasets 2.15.0
- Tokenizers 0.15.0
|
Szycha/flan-t5-base-samsum
|
Szycha
| 2023-12-19T19:33:45Z | 4 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"t5",
"text2text-generation",
"generated_from_trainer",
"dataset:samsum",
"base_model:google/flan-t5-base",
"base_model:finetune:google/flan-t5-base",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2023-12-19T19:25:49Z |
---
license: apache-2.0
base_model: google/flan-t5-base
tags:
- generated_from_trainer
datasets:
- samsum
model-index:
- name: flan-t5-base-samsum
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# flan-t5-base-samsum
This model is a fine-tuned version of [google/flan-t5-base](https://huggingface.co/google/flan-t5-base) on the samsum dataset.
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5
### Framework versions
- Transformers 4.36.2
- Pytorch 2.1.2+cu121
- Datasets 2.15.0
- Tokenizers 0.15.0
|
ahmedabdelwahed/SFT-base-5-epochs
|
ahmedabdelwahed
| 2023-12-19T19:32:16Z | 0 | 0 |
peft
|
[
"peft",
"safetensors",
"arxiv:1910.09700",
"base_model:google/mt5-base",
"base_model:adapter:google/mt5-base",
"region:us"
] | null | 2023-12-19T19:32:15Z |
---
library_name: peft
base_model: google/mt5-base
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
### Framework versions
- PEFT 0.7.1
|
behzadnet/Llama-2-7b-chat-hf-sharded-bf16-fine-tuned_SystemError0.6_Seed104
|
behzadnet
| 2023-12-19T19:26:36Z | 1 | 0 |
peft
|
[
"peft",
"arxiv:1910.09700",
"base_model:Trelis/Llama-2-7b-chat-hf-sharded-bf16",
"base_model:adapter:Trelis/Llama-2-7b-chat-hf-sharded-bf16",
"region:us"
] | null | 2023-12-19T19:26:33Z |
---
library_name: peft
base_model: Trelis/Llama-2-7b-chat-hf-sharded-bf16
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
- **Developed by:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Data Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Data Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- load_in_8bit: False
- load_in_4bit: True
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: nf4
- bnb_4bit_use_double_quant: True
- bnb_4bit_compute_dtype: bfloat16
### Framework versions
- PEFT 0.7.0.dev0
|
sk-2302/flan-t5-small-samsum
|
sk-2302
| 2023-12-19T19:26:32Z | 5 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"t5",
"text2text-generation",
"generated_from_trainer",
"dataset:samsum",
"base_model:google/flan-t5-small",
"base_model:finetune:google/flan-t5-small",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2023-12-19T18:24:50Z |
---
license: apache-2.0
base_model: google/flan-t5-small
tags:
- generated_from_trainer
datasets:
- samsum
metrics:
- rouge
model-index:
- name: flan-t5-small-samsum
results:
- task:
name: Sequence-to-sequence Language Modeling
type: text2text-generation
dataset:
name: samsum
type: samsum
config: samsum
split: test
args: samsum
metrics:
- name: Rouge1
type: rouge
value: 42.6
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# flan-t5-small-samsum
This model is a fine-tuned version of [google/flan-t5-small](https://huggingface.co/google/flan-t5-small) on the samsum dataset.
It achieves the following results on the evaluation set:
- Loss: 1.6729
- Rouge1: 42.6
- Rouge2: 18.7153
- Rougel: 35.4138
- Rougelsum: 38.8543
- Gen Len: 16.9170
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:-------:|
| 1.8863 | 0.22 | 100 | 1.7049 | 42.0859 | 18.0002 | 34.7349 | 38.3446 | 16.5788 |
| 1.8463 | 0.43 | 200 | 1.6947 | 42.4056 | 18.3005 | 34.9821 | 38.8013 | 17.3614 |
| 1.8548 | 0.65 | 300 | 1.6792 | 42.585 | 18.5643 | 35.2235 | 38.8298 | 17.1514 |
| 1.8358 | 0.87 | 400 | 1.6772 | 42.1544 | 18.2303 | 34.8971 | 38.3609 | 16.5873 |
| 1.8129 | 1.08 | 500 | 1.6729 | 42.6 | 18.7153 | 35.4138 | 38.8543 | 16.9170 |
| 1.8068 | 1.3 | 600 | 1.6709 | 42.5217 | 18.3285 | 35.1455 | 38.5954 | 16.9451 |
| 1.7973 | 1.52 | 700 | 1.6687 | 42.8667 | 18.624 | 35.3429 | 38.9322 | 16.7546 |
| 1.7979 | 1.74 | 800 | 1.6668 | 42.919 | 18.7388 | 35.4528 | 39.0561 | 16.8791 |
| 1.7899 | 1.95 | 900 | 1.6670 | 43.0931 | 18.741 | 35.5047 | 39.2321 | 16.9109 |
### Framework versions
- Transformers 4.36.0
- Pytorch 2.0.0
- Datasets 2.15.0
- Tokenizers 0.15.0
|
behzadnet/Llama-2-7b-chat-hf-sharded-bf16-fine-tuned-adapters_SystemError0.6_Seed104
|
behzadnet
| 2023-12-19T19:26:25Z | 1 | 0 |
peft
|
[
"peft",
"arxiv:1910.09700",
"base_model:Trelis/Llama-2-7b-chat-hf-sharded-bf16",
"base_model:adapter:Trelis/Llama-2-7b-chat-hf-sharded-bf16",
"region:us"
] | null | 2023-12-19T19:26:16Z |
---
library_name: peft
base_model: Trelis/Llama-2-7b-chat-hf-sharded-bf16
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
- **Developed by:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Data Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Data Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- load_in_8bit: False
- load_in_4bit: True
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: nf4
- bnb_4bit_use_double_quant: True
- bnb_4bit_compute_dtype: bfloat16
### Framework versions
- PEFT 0.7.0.dev0
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- load_in_8bit: False
- load_in_4bit: True
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: nf4
- bnb_4bit_use_double_quant: True
- bnb_4bit_compute_dtype: bfloat16
### Framework versions
- PEFT 0.7.0.dev0
|
bryanlincoln/tmp_trainer
|
bryanlincoln
| 2023-12-19T19:24:55Z | 6 | 0 |
transformers
|
[
"transformers",
"safetensors",
"bert",
"text-classification",
"generated_from_trainer",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2023-12-19T19:24:02Z |
---
tags:
- generated_from_trainer
model-index:
- name: tmp_trainer
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# tmp_trainer
This model was trained from scratch on an unknown dataset.
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
### Framework versions
- Transformers 4.36.2
- Pytorch 2.1.1+cu121
- Datasets 2.15.0
- Tokenizers 0.15.0
|
dpogreb/ai
|
dpogreb
| 2023-12-19T19:22:57Z | 0 | 0 | null |
[
"code",
"zero-shot-classification",
"en",
"dataset:wikimedia/wikipedia",
"dataset:nvidia/HelpSteer",
"license:unlicense",
"region:us"
] |
zero-shot-classification
| 2023-12-19T19:21:53Z |
---
license: unlicense
datasets:
- wikimedia/wikipedia
- nvidia/HelpSteer
language:
- en
metrics:
- accuracy
- code_eval
pipeline_tag: zero-shot-classification
tags:
- code
---
|
MichalGas/instrument-0
|
MichalGas
| 2023-12-19T19:22:17Z | 26 | 0 |
transformers
|
[
"transformers",
"safetensors",
"resnet",
"image-classification",
"autotrain",
"dataset:MichalGas/autotrain-data-instrument",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
image-classification
| 2023-12-19T19:22:12Z |
---
tags:
- autotrain
- image-classification
widget:
- src: https://huggingface.co/datasets/mishig/sample_images/resolve/main/tiger.jpg
example_title: Tiger
- src: https://huggingface.co/datasets/mishig/sample_images/resolve/main/teapot.jpg
example_title: Teapot
- src: https://huggingface.co/datasets/mishig/sample_images/resolve/main/palace.jpg
example_title: Palace
datasets:
- MichalGas/autotrain-data-instrument
---
# Model Trained Using AutoTrain
- Problem type: Image Classification
## Validation Metricsg
loss: nan
f1_macro: 0.05573770491803279
f1_micro: 0.20078740157480315
f1_weighted: 0.06714857364140958
precision_macro: 0.03346456692913386
precision_micro: 0.20078740157480315
precision_weighted: 0.040315580631161266
recall_macro: 0.16666666666666666
recall_micro: 0.20078740157480315
recall_weighted: 0.20078740157480315
accuracy: 0.20078740157480315
|
xpyct01/bruch
|
xpyct01
| 2023-12-19T19:17:21Z | 0 | 0 |
stable-baselines3
|
[
"stable-baselines3",
"LunarLander-v2",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-12-19T19:16:36Z |
---
library_name: stable-baselines3
tags:
- LunarLander-v2
- deep-reinforcement-learning
- reinforcement-learning
- stable-baselines3
model-index:
- name: PPO
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: LunarLander-v2
type: LunarLander-v2
metrics:
- type: mean_reward
value: 254.19 +/- 17.37
name: mean_reward
verified: false
---
# **PPO** Agent playing **LunarLander-v2**
This is a trained model of a **PPO** agent playing **LunarLander-v2**
using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3).
## Usage (with Stable-baselines3)
TODO: Add your code
```python
from stable_baselines3 import ...
from huggingface_sb3 import load_from_hub
...
```
|
AlexRez/distiluse-base-multilingual-cased-v2-4epochs-CITIESFINDER
|
AlexRez
| 2023-12-19T19:08:55Z | 13 | 0 |
sentence-transformers
|
[
"sentence-transformers",
"safetensors",
"distilbert",
"feature-extraction",
"sentence-similarity",
"autotrain_compatible",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
] |
sentence-similarity
| 2023-12-19T16:27:46Z |
---
pipeline_tag: sentence-similarity
tags:
- sentence-transformers
- feature-extraction
- sentence-similarity
---
# {MODEL_NAME}
This is a [sentence-transformers](https://www.SBERT.net) model: It maps sentences & paragraphs to a 512 dimensional dense vector space and can be used for tasks like clustering or semantic search.
<!--- Describe your model here -->
## Usage (Sentence-Transformers)
Using this model becomes easy when you have [sentence-transformers](https://www.SBERT.net) installed:
```
pip install -U sentence-transformers
```
Then you can use the model like this:
```python
from sentence_transformers import SentenceTransformer
sentences = ["This is an example sentence", "Each sentence is converted"]
model = SentenceTransformer('{MODEL_NAME}')
embeddings = model.encode(sentences)
print(embeddings)
```
## Evaluation Results
<!--- Describe how your model was evaluated -->
For an automated evaluation of this model, see the *Sentence Embeddings Benchmark*: [https://seb.sbert.net](https://seb.sbert.net?model_name={MODEL_NAME})
## Training
The model was trained with the parameters:
**DataLoader**:
`torch.utils.data.dataloader.DataLoader` of length 2040 with parameters:
```
{'batch_size': 32, 'sampler': 'torch.utils.data.sampler.RandomSampler', 'batch_sampler': 'torch.utils.data.sampler.BatchSampler'}
```
**Loss**:
`sentence_transformers.losses.MultipleNegativesRankingLoss.MultipleNegativesRankingLoss` with parameters:
```
{'scale': 20.0, 'similarity_fct': 'cos_sim'}
```
Parameters of the fit()-Method:
```
{
"epochs": 1,
"evaluation_steps": 0,
"evaluator": "NoneType",
"max_grad_norm": 1,
"optimizer_class": "<class 'torch.optim.adamw.AdamW'>",
"optimizer_params": {
"lr": 2e-05
},
"scheduler": "WarmupLinear",
"steps_per_epoch": null,
"warmup_steps": 10000,
"weight_decay": 0.01
}
```
## Full Model Architecture
```
SentenceTransformer(
(0): Transformer({'max_seq_length': 128, 'do_lower_case': False}) with Transformer model: DistilBertModel
(1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False})
(2): Dense({'in_features': 768, 'out_features': 512, 'bias': True, 'activation_function': 'torch.nn.modules.activation.Tanh'})
)
```
## Citing & Authors
<!--- Describe where people can find more information -->
|
TheBloke/Swallow-70B-instruct-GGUF
|
TheBloke
| 2023-12-19T19:01:33Z | 130 | 8 |
transformers
|
[
"transformers",
"gguf",
"llama",
"text-generation",
"en",
"ja",
"base_model:tokyotech-llm/Swallow-70b-instruct-hf",
"base_model:quantized:tokyotech-llm/Swallow-70b-instruct-hf",
"license:llama2",
"region:us"
] |
text-generation
| 2023-12-19T18:37:03Z |
---
base_model: tokyotech-llm/Swallow-70b-instruct-hf
inference: false
language:
- en
- ja
library_name: transformers
license: llama2
model_creator: tokyotech-llm
model_name: Swallow 70B Instruct
model_type: llama
pipeline_tag: text-generation
prompt_template: "\u4EE5\u4E0B\u306B\u3001\u3042\u308B\u30BF\u30B9\u30AF\u3092\u8AAC\
\u660E\u3059\u308B\u6307\u793A\u304C\u3042\u308A\u307E\u3059\u3002\u30EA\u30AF\u30A8\
\u30B9\u30C8\u3092\u9069\u5207\u306B\u5B8C\u4E86\u3059\u308B\u305F\u3081\u306E\u56DE\
\u7B54\u3092\u8A18\u8FF0\u3057\u3066\u304F\u3060\u3055\u3044\u3002\\n\\n### \u6307\
\u793A:\\n{prompt}\\n\\n### \u5FDC\u7B54:\n"
quantized_by: TheBloke
---
<!-- markdownlint-disable MD041 -->
<!-- header start -->
<!-- 200823 -->
<div style="width: auto; margin-left: auto; margin-right: auto">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<div style="text-align:center; margin-top: 0em; margin-bottom: 0em"><p style="margin-top: 0.25em; margin-bottom: 0em;">TheBloke's LLM work is generously supported by a grant from <a href="https://a16z.com">andreessen horowitz (a16z)</a></p></div>
<hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
<!-- header end -->
# Swallow 70B Instruct - GGUF
- Model creator: [tokyotech-llm](https://huggingface.co/tokyotech-llm)
- Original model: [Swallow 70B Instruct](https://huggingface.co/tokyotech-llm/Swallow-70b-instruct-hf)
<!-- description start -->
## Description
This repo contains GGUF format model files for [tokyotech-llm's Swallow 70B Instruct](https://huggingface.co/tokyotech-llm/Swallow-70b-instruct-hf).
<!-- description end -->
<!-- README_GGUF.md-about-gguf start -->
### About GGUF
GGUF is a new format introduced by the llama.cpp team on August 21st 2023. It is a replacement for GGML, which is no longer supported by llama.cpp.
Here is an incomplete list of clients and libraries that are known to support GGUF:
* [llama.cpp](https://github.com/ggerganov/llama.cpp). The source project for GGUF. Offers a CLI and a server option.
* [text-generation-webui](https://github.com/oobabooga/text-generation-webui), the most widely used web UI, with many features and powerful extensions. Supports GPU acceleration.
* [KoboldCpp](https://github.com/LostRuins/koboldcpp), a fully featured web UI, with GPU accel across all platforms and GPU architectures. Especially good for story telling.
* [GPT4All](https://gpt4all.io/index.html), a free and open source local running GUI, supporting Windows, Linux and macOS with full GPU accel.
* [LM Studio](https://lmstudio.ai/), an easy-to-use and powerful local GUI for Windows and macOS (Silicon), with GPU acceleration. Linux available, in beta as of 27/11/2023.
* [LoLLMS Web UI](https://github.com/ParisNeo/lollms-webui), a great web UI with many interesting and unique features, including a full model library for easy model selection.
* [Faraday.dev](https://faraday.dev/), an attractive and easy to use character-based chat GUI for Windows and macOS (both Silicon and Intel), with GPU acceleration.
* [llama-cpp-python](https://github.com/abetlen/llama-cpp-python), a Python library with GPU accel, LangChain support, and OpenAI-compatible API server.
* [candle](https://github.com/huggingface/candle), a Rust ML framework with a focus on performance, including GPU support, and ease of use.
* [ctransformers](https://github.com/marella/ctransformers), a Python library with GPU accel, LangChain support, and OpenAI-compatible AI server. Note, as of time of writing (November 27th 2023), ctransformers has not been updated in a long time and does not support many recent models.
<!-- README_GGUF.md-about-gguf end -->
<!-- repositories-available start -->
## Repositories available
* [AWQ model(s) for GPU inference.](https://huggingface.co/TheBloke/Swallow-70B-instruct-AWQ)
* [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/Swallow-70B-instruct-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/Swallow-70B-instruct-GGUF)
* [tokyotech-llm's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/tokyotech-llm/Swallow-70b-instruct-hf)
<!-- repositories-available end -->
<!-- prompt-template start -->
## Prompt template: Swallow-Instruct
```
以下に、あるタスクを説明する指示があります。リクエストを適切に完了するための回答を記述してください。\n\n### 指示:\n{prompt}\n\n### 応答:
```
<!-- prompt-template end -->
<!-- compatibility_gguf start -->
## Compatibility
These quantised GGUFv2 files are compatible with llama.cpp from August 27th onwards, as of commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221)
They are also compatible with many third party UIs and libraries - please see the list at the top of this README.
## Explanation of quantisation methods
<details>
<summary>Click to see details</summary>
The new methods available are:
* GGML_TYPE_Q2_K - "type-1" 2-bit quantization in super-blocks containing 16 blocks, each block having 16 weight. Block scales and mins are quantized with 4 bits. This ends up effectively using 2.5625 bits per weight (bpw)
* GGML_TYPE_Q3_K - "type-0" 3-bit quantization in super-blocks containing 16 blocks, each block having 16 weights. Scales are quantized with 6 bits. This end up using 3.4375 bpw.
* GGML_TYPE_Q4_K - "type-1" 4-bit quantization in super-blocks containing 8 blocks, each block having 32 weights. Scales and mins are quantized with 6 bits. This ends up using 4.5 bpw.
* GGML_TYPE_Q5_K - "type-1" 5-bit quantization. Same super-block structure as GGML_TYPE_Q4_K resulting in 5.5 bpw
* GGML_TYPE_Q6_K - "type-0" 6-bit quantization. Super-blocks with 16 blocks, each block having 16 weights. Scales are quantized with 8 bits. This ends up using 6.5625 bpw
Refer to the Provided Files table below to see what files use which methods, and how.
</details>
<!-- compatibility_gguf end -->
<!-- README_GGUF.md-provided-files start -->
## Provided files
| Name | Quant method | Bits | Size | Max RAM required | Use case |
| ---- | ---- | ---- | ---- | ---- | ----- |
| [swallow-70b-instruct.Q2_K.gguf](https://huggingface.co/TheBloke/Swallow-70B-instruct-GGUF/blob/main/swallow-70b-instruct.Q2_K.gguf) | Q2_K | 2 | 29.38 GB| 31.88 GB | smallest, significant quality loss - not recommended for most purposes |
| [swallow-70b-instruct.Q3_K_S.gguf](https://huggingface.co/TheBloke/Swallow-70B-instruct-GGUF/blob/main/swallow-70b-instruct.Q3_K_S.gguf) | Q3_K_S | 3 | 30.03 GB| 32.53 GB | very small, high quality loss |
| [swallow-70b-instruct.Q3_K_M.gguf](https://huggingface.co/TheBloke/Swallow-70B-instruct-GGUF/blob/main/swallow-70b-instruct.Q3_K_M.gguf) | Q3_K_M | 3 | 33.30 GB| 35.80 GB | very small, high quality loss |
| [swallow-70b-instruct.Q3_K_L.gguf](https://huggingface.co/TheBloke/Swallow-70B-instruct-GGUF/blob/main/swallow-70b-instruct.Q3_K_L.gguf) | Q3_K_L | 3 | 36.26 GB| 38.76 GB | small, substantial quality loss |
| [swallow-70b-instruct.Q4_0.gguf](https://huggingface.co/TheBloke/Swallow-70B-instruct-GGUF/blob/main/swallow-70b-instruct.Q4_0.gguf) | Q4_0 | 4 | 39.00 GB| 41.50 GB | legacy; small, very high quality loss - prefer using Q3_K_M |
| [swallow-70b-instruct.Q4_K_S.gguf](https://huggingface.co/TheBloke/Swallow-70B-instruct-GGUF/blob/main/swallow-70b-instruct.Q4_K_S.gguf) | Q4_K_S | 4 | 39.20 GB| 41.70 GB | small, greater quality loss |
| [swallow-70b-instruct.Q4_K_M.gguf](https://huggingface.co/TheBloke/Swallow-70B-instruct-GGUF/blob/main/swallow-70b-instruct.Q4_K_M.gguf) | Q4_K_M | 4 | 41.55 GB| 44.05 GB | medium, balanced quality - recommended |
| [swallow-70b-instruct.Q5_0.gguf](https://huggingface.co/TheBloke/Swallow-70B-instruct-GGUF/blob/main/swallow-70b-instruct.Q5_0.gguf) | Q5_0 | 5 | 47.60 GB| 50.10 GB | legacy; medium, balanced quality - prefer using Q4_K_M |
| [swallow-70b-instruct.Q5_K_S.gguf](https://huggingface.co/TheBloke/Swallow-70B-instruct-GGUF/blob/main/swallow-70b-instruct.Q5_K_S.gguf) | Q5_K_S | 5 | 47.60 GB| 50.10 GB | large, low quality loss - recommended |
| [swallow-70b-instruct.Q5_K_M.gguf](https://huggingface.co/TheBloke/Swallow-70B-instruct-GGUF/blob/main/swallow-70b-instruct.Q5_K_M.gguf) | Q5_K_M | 5 | 48.89 GB| 51.39 GB | large, very low quality loss - recommended |
| swallow-70b-instruct.Q6_K.gguf | Q6_K | 6 | 56.74 GB| 59.24 GB | very large, extremely low quality loss |
| swallow-70b-instruct.Q8_0.gguf | Q8_0 | 8 | 73.49 GB| 75.99 GB | very large, extremely low quality loss - not recommended |
**Note**: the above RAM figures assume no GPU offloading. If layers are offloaded to the GPU, this will reduce RAM usage and use VRAM instead.
### Q6_K and Q8_0 files are split and require joining
**Note:** HF does not support uploading files larger than 50GB. Therefore I have uploaded the Q6_K and Q8_0 files as split files.
<details>
<summary>Click for instructions regarding Q6_K and Q8_0 files</summary>
### q6_K
Please download:
* `swallow-70b-instruct.Q6_K.gguf-split-a`
* `swallow-70b-instruct.Q6_K.gguf-split-b`
### q8_0
Please download:
* `swallow-70b-instruct.Q8_0.gguf-split-a`
* `swallow-70b-instruct.Q8_0.gguf-split-b`
To join the files, do the following:
Linux and macOS:
```
cat swallow-70b-instruct.Q6_K.gguf-split-* > swallow-70b-instruct.Q6_K.gguf && rm swallow-70b-instruct.Q6_K.gguf-split-*
cat swallow-70b-instruct.Q8_0.gguf-split-* > swallow-70b-instruct.Q8_0.gguf && rm swallow-70b-instruct.Q8_0.gguf-split-*
```
Windows command line:
```
COPY /B swallow-70b-instruct.Q6_K.gguf-split-a + swallow-70b-instruct.Q6_K.gguf-split-b swallow-70b-instruct.Q6_K.gguf
del swallow-70b-instruct.Q6_K.gguf-split-a swallow-70b-instruct.Q6_K.gguf-split-b
COPY /B swallow-70b-instruct.Q8_0.gguf-split-a + swallow-70b-instruct.Q8_0.gguf-split-b swallow-70b-instruct.Q8_0.gguf
del swallow-70b-instruct.Q8_0.gguf-split-a swallow-70b-instruct.Q8_0.gguf-split-b
```
</details>
<!-- README_GGUF.md-provided-files end -->
<!-- README_GGUF.md-how-to-download start -->
## How to download GGUF files
**Note for manual downloaders:** You almost never want to clone the entire repo! Multiple different quantisation formats are provided, and most users only want to pick and download a single file.
The following clients/libraries will automatically download models for you, providing a list of available models to choose from:
* LM Studio
* LoLLMS Web UI
* Faraday.dev
### In `text-generation-webui`
Under Download Model, you can enter the model repo: TheBloke/Swallow-70B-instruct-GGUF and below it, a specific filename to download, such as: swallow-70b-instruct.Q4_K_M.gguf.
Then click Download.
### On the command line, including multiple files at once
I recommend using the `huggingface-hub` Python library:
```shell
pip3 install huggingface-hub
```
Then you can download any individual model file to the current directory, at high speed, with a command like this:
```shell
huggingface-cli download TheBloke/Swallow-70B-instruct-GGUF swallow-70b-instruct.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False
```
<details>
<summary>More advanced huggingface-cli download usage (click to read)</summary>
You can also download multiple files at once with a pattern:
```shell
huggingface-cli download TheBloke/Swallow-70B-instruct-GGUF --local-dir . --local-dir-use-symlinks False --include='*Q4_K*gguf'
```
For more documentation on downloading with `huggingface-cli`, please see: [HF -> Hub Python Library -> Download files -> Download from the CLI](https://huggingface.co/docs/huggingface_hub/guides/download#download-from-the-cli).
To accelerate downloads on fast connections (1Gbit/s or higher), install `hf_transfer`:
```shell
pip3 install hf_transfer
```
And set environment variable `HF_HUB_ENABLE_HF_TRANSFER` to `1`:
```shell
HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download TheBloke/Swallow-70B-instruct-GGUF swallow-70b-instruct.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False
```
Windows Command Line users: You can set the environment variable by running `set HF_HUB_ENABLE_HF_TRANSFER=1` before the download command.
</details>
<!-- README_GGUF.md-how-to-download end -->
<!-- README_GGUF.md-how-to-run start -->
## Example `llama.cpp` command
Make sure you are using `llama.cpp` from commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221) or later.
```shell
./main -ngl 35 -m swallow-70b-instruct.Q4_K_M.gguf --color -c 4096 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "以下に、あるタスクを説明する指示があります。リクエストを適切に完了するための回答を記述してください。\n\n### 指示:\n{prompt}\n\n### 応答:"
```
Change `-ngl 32` to the number of layers to offload to GPU. Remove it if you don't have GPU acceleration.
Change `-c 4096` to the desired sequence length. For extended sequence models - eg 8K, 16K, 32K - the necessary RoPE scaling parameters are read from the GGUF file and set by llama.cpp automatically. Note that longer sequence lengths require much more resources, so you may need to reduce this value.
If you want to have a chat-style conversation, replace the `-p <PROMPT>` argument with `-i -ins`
For other parameters and how to use them, please refer to [the llama.cpp documentation](https://github.com/ggerganov/llama.cpp/blob/master/examples/main/README.md)
## How to run in `text-generation-webui`
Further instructions can be found in the text-generation-webui documentation, here: [text-generation-webui/docs/04 ‐ Model Tab.md](https://github.com/oobabooga/text-generation-webui/blob/main/docs/04%20%E2%80%90%20Model%20Tab.md#llamacpp).
## How to run from Python code
You can use GGUF models from Python using the [llama-cpp-python](https://github.com/abetlen/llama-cpp-python) or [ctransformers](https://github.com/marella/ctransformers) libraries. Note that at the time of writing (Nov 27th 2023), ctransformers has not been updated for some time and is not compatible with some recent models. Therefore I recommend you use llama-cpp-python.
### How to load this model in Python code, using llama-cpp-python
For full documentation, please see: [llama-cpp-python docs](https://abetlen.github.io/llama-cpp-python/).
#### First install the package
Run one of the following commands, according to your system:
```shell
# Base ctransformers with no GPU acceleration
pip install llama-cpp-python
# With NVidia CUDA acceleration
CMAKE_ARGS="-DLLAMA_CUBLAS=on" pip install llama-cpp-python
# Or with OpenBLAS acceleration
CMAKE_ARGS="-DLLAMA_BLAS=ON -DLLAMA_BLAS_VENDOR=OpenBLAS" pip install llama-cpp-python
# Or with CLBLast acceleration
CMAKE_ARGS="-DLLAMA_CLBLAST=on" pip install llama-cpp-python
# Or with AMD ROCm GPU acceleration (Linux only)
CMAKE_ARGS="-DLLAMA_HIPBLAS=on" pip install llama-cpp-python
# Or with Metal GPU acceleration for macOS systems only
CMAKE_ARGS="-DLLAMA_METAL=on" pip install llama-cpp-python
# In windows, to set the variables CMAKE_ARGS in PowerShell, follow this format; eg for NVidia CUDA:
$env:CMAKE_ARGS = "-DLLAMA_OPENBLAS=on"
pip install llama-cpp-python
```
#### Simple llama-cpp-python example code
```python
from llama_cpp import Llama
# Set gpu_layers to the number of layers to offload to GPU. Set to 0 if no GPU acceleration is available on your system.
llm = Llama(
model_path="./swallow-70b-instruct.Q4_K_M.gguf", # Download the model file first
n_ctx=4096, # The max sequence length to use - note that longer sequence lengths require much more resources
n_threads=8, # The number of CPU threads to use, tailor to your system and the resulting performance
n_gpu_layers=35 # The number of layers to offload to GPU, if you have GPU acceleration available
)
# Simple inference example
output = llm(
"以下に、あるタスクを説明する指示があります。リクエストを適切に完了するための回答を記述してください。\n\n### 指示:\n{prompt}\n\n### 応答:", # Prompt
max_tokens=512, # Generate up to 512 tokens
stop=["</s>"], # Example stop token - not necessarily correct for this specific model! Please check before using.
echo=True # Whether to echo the prompt
)
# Chat Completion API
llm = Llama(model_path="./swallow-70b-instruct.Q4_K_M.gguf", chat_format="llama-2") # Set chat_format according to the model you are using
llm.create_chat_completion(
messages = [
{"role": "system", "content": "You are a story writing assistant."},
{
"role": "user",
"content": "Write a story about llamas."
}
]
)
```
## How to use with LangChain
Here are guides on using llama-cpp-python and ctransformers with LangChain:
* [LangChain + llama-cpp-python](https://python.langchain.com/docs/integrations/llms/llamacpp)
* [LangChain + ctransformers](https://python.langchain.com/docs/integrations/providers/ctransformers)
<!-- README_GGUF.md-how-to-run end -->
<!-- footer start -->
<!-- 200823 -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute
Thanks to the [chirper.ai](https://chirper.ai) team!
Thanks to Clay from [gpus.llm-utils.org](llm-utils)!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Aemon Algiz.
**Patreon special mentions**: Michael Levine, 阿明, Trailburnt, Nikolai Manek, John Detwiler, Randy H, Will Dee, Sebastain Graf, NimbleBox.ai, Eugene Pentland, Emad Mostaque, Ai Maven, Jim Angel, Jeff Scroggin, Michael Davis, Manuel Alberto Morcote, Stephen Murray, Robert, Justin Joy, Luke @flexchar, Brandon Frisco, Elijah Stavena, S_X, Dan Guido, Undi ., Komninos Chatzipapas, Shadi, theTransient, Lone Striker, Raven Klaugh, jjj, Cap'n Zoog, Michel-Marie MAUDET (LINAGORA), Matthew Berman, David, Fen Risland, Omer Bin Jawed, Luke Pendergrass, Kalila, OG, Erik Bjäreholt, Rooh Singh, Joseph William Delisle, Dan Lewis, TL, John Villwock, AzureBlack, Brad, Pedro Madruga, Caitlyn Gatomon, K, jinyuan sun, Mano Prime, Alex, Jeffrey Morgan, Alicia Loh, Illia Dulskyi, Chadd, transmissions 11, fincy, Rainer Wilmers, ReadyPlayerEmma, knownsqashed, Mandus, biorpg, Deo Leter, Brandon Phillips, SuperWojo, Sean Connelly, Iucharbius, Jack West, Harry Royden McLaughlin, Nicholas, terasurfer, Vitor Caleffi, Duane Dunston, Johann-Peter Hartmann, David Ziegler, Olakabola, Ken Nordquist, Trenton Dambrowitz, Tom X Nguyen, Vadim, Ajan Kanaga, Leonard Tan, Clay Pascal, Alexandros Triantafyllidis, JM33133, Xule, vamX, ya boyyy, subjectnull, Talal Aujan, Alps Aficionado, wassieverse, Ari Malik, James Bentley, Woland, Spencer Kim, Michael Dempsey, Fred von Graf, Elle, zynix, William Richards, Stanislav Ovsiannikov, Edmond Seymore, Jonathan Leane, Martin Kemka, usrbinkat, Enrico Ros
Thank you to all my generous patrons and donaters!
And thank you again to a16z for their generous grant.
<!-- footer end -->
<!-- original-model-card start -->
# Original model card: tokyotech-llm's Swallow 70B Instruct
# Swallow
Our Swallow model has undergone continuous pre-training from the Llama 2 family, primarily with the addition of Japanese language data. The tuned versions use supervised fine-tuning (SFT).
Links to other models can be found in the index.
## Swallow Model Index
|Model|Swallow-hf|Swallow-instruct-hf|
|---|---|---|
|7B| [Link](https://huggingface.co/tokyotech-llm/Swallow-7b-hf) | [Link](https://huggingface.co/tokyotech-llm/Swallow-7b-instruct-hf)|
|13B| [Link](https://huggingface.co/tokyotech-llm/Swallow-13b-hf) | [Link](https://huggingface.co/tokyotech-llm/Swallow-13b-instruct-hf)|
|70B| [Link](https://huggingface.co/tokyotech-llm/Swallow-70b-hf) | [Link](https://huggingface.co/tokyotech-llm/Swallow-70b-instruct-hf)|
This repository provides large language models developed by [TokyoTech-LLM](https://tokyotech-llm.github.io/).
Read our [blog post](https://zenn.dev/tokyotech_lm/articles/d6cb3a8fdfc907) or our paper (preprint coming soon) for more details!
## Model Details
* **Model type**: Please refer to LLaMA-2 technical report for details on the model architecture.
* **Language(s)**: Japanese English
* **Library**: [Megatron-LM](https://github.com/rioyokotalab/Megatron-Llama2)
* **Tokenizer**: This model employs a tokenizer that features a broadened vocabulary based on Japanese data. This allows for a more efficient representation of text using fewer tokens, leading to a notably faster inference process.
* **Contact**: swallow[at]nlp.c.titech.ac.jp
## Base Model Performance
### Japanese version
|Model|Size|JCommonsenseQA|JEMHopQA|NIILC|JSQuAD|XL-Sum|MGSM|WMT20-en-ja|WMT20-ja-en|
|---|---|---|---|---|---|---|---|---|---|
| | |4-shot|4-shot|4-shot|4-shot|1-shot|4-shot|4-shot|4-shot|
|Llama 2|7B|0.3852|0.4240|0.3410|0.7917|0.1905|0.0760|0.1783|0.1738|
|Swallow|7B|0.4808|0.5078|0.5968|0.8573|0.1830|0.1240|0.2510|0.1511|
|Llama 2|13B|0.6997|0.4415|0.4170|0.8533|0.2139|0.1320|0.2146|0.1982|
|Swallow|13B|0.7837|0.5063|0.6398|0.9005|0.2168|0.2040|0.2720|0.1771|
|Llama 2|70B|0.8686|0.4656|0.5256|0.9080|**0.2361**|0.3560|0.2643|**0.2398**|
|Swallow|70B|**0.9348**|**0.6290**|**0.6960**|**0.9176**|0.2266|**0.4840**|**0.3043**|0.2298|
## Usage
First install additional dependencies in [requirements.txt](./requirements.txt):
```sh
pip install -r requirements.txt
```
### Use the instruct model
```python
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
model_name = "tokyotech-llm/Swallow-7b-instruct-hf"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.bfloat16, low_cpu_mem_usage=True, device_map="auto")
PROMPT_DICT = {
"prompt_input": (
"以下に、あるタスクを説明する指示があり、それに付随する入力が更なる文脈を提供しています。"
"リクエストを適切に完了するための回答を記述してください。\n\n"
"### 指示:\n{instruction}\n\n### 入力:\n{input}\n\n### 応答:"
),
"prompt_no_input": (
"以下に、あるタスクを説明する指示があります。"
"リクエストを適切に完了するための回答を記述してください。\n\n"
"### 指示:\n{instruction}\n\n### 応答:"
),
}
def create_prompt(instruction, input=None):
"""
Generates a prompt based on the given instruction and an optional input.
If input is provided, it uses the 'prompt_input' template from PROMPT_DICT.
If no input is provided, it uses the 'prompt_no_input' template.
Args:
instruction (str): The instruction describing the task.
input (str, optional): Additional input providing context for the task. Default is None.
Returns:
str: The generated prompt.
"""
if input:
# Use the 'prompt_input' template when additional input is provided
return PROMPT_DICT["prompt_input"].format(instruction=instruction, input=input)
else:
# Use the 'prompt_no_input' template when no additional input is provided
return PROMPT_DICT["prompt_no_input"].format(instruction=instruction)
# Example usage
instruction_example = "以下のトピックに関する詳細な情報を提供してください。"
input_example = "東京工業大学の主なキャンパスについて教えてください"
prompt = create_prompt(instruction_example, input_example)
input_ids = tokenizer.encode(
prompt,
add_special_tokens=False,
return_tensors="pt"
)
tokens = model.generate(
input_ids.to(device=model.device),
max_new_tokens=128,
temperature=0.99,
top_p=0.95,
do_sample=True,
)
out = tokenizer.decode(tokens[0], skip_special_tokens=True)
print(out)
```
### Use the base model
```python
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
model_name = "tokyotech-llm/Swallow-7b-hf"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.bfloat16, device_map="auto")
prompt = "東京工業大学の主なキャンパスは、"
input_ids = tokenizer.encode(
prompt,
add_special_tokens=False,
return_tensors="pt"
)
tokens = model.generate(
input_ids.to(device=model.device),
max_new_tokens=128,
temperature=0.99,
top_p=0.95,
do_sample=True,
)
out = tokenizer.decode(tokens[0], skip_special_tokens=True)
print(out)
```
## Training Datasets
### Continual Pre-Training
The following datasets were used for continual pre-training.
- [Japanese Wikipedia](https://dumps.wikimedia.org/other/cirrussearch)
- [RefinedWeb](https://huggingface.co/datasets/tiiuae/falcon-refinedweb)
- Swallow Corpus
- [The Pile](https://huggingface.co/datasets/EleutherAI/pile)
### Instruction Tuning
The following datasets were used for the instruction tuning.
- [Anthropic HH-RLHF](https://huggingface.co/datasets/kunishou/hh-rlhf-49k-ja)
- [Databricks Dolly 15-k](https://huggingface.co/datasets/kunishou/databricks-dolly-15k-ja)
- [OpenAssistant Conversations Dataset](https://huggingface.co/datasets/kunishou/oasst1-89k-ja)
## Risks and Limitations
The models released here are still in the early stages of our research and development and have not been tuned to ensure outputs align with human intent and safety considerations.
## Acknowledgements
We thank Meta Research for releasing Llama 2 under an open license for others to build on.
Our project is supported by the [ABCI Large-scale Language Model Building Support Program](https://abci.ai/en/link/llm_support_program.html) of the National Institute of Advanced Industrial Science and Technology.
## License
Llama 2 is licensed under the LLAMA 2 Community License, Copyright © Meta Platforms, Inc. All Rights Reserved.
## Authors
Here are the team members:
- From [Okazaki Laboratory](https://www.nlp.c.titech.ac.jp/index.en.html), the following members:
- [Naoaki Okazaki](https://www.chokkan.org/index.ja.html)
- [Sakae Mizuki](https://s-mizuki-nlp.github.io/)
- [Hiroki Iida](https://meshidenn.github.io/)
- [Mengsay Loem](https://loem-ms.github.io/)
- [Shota Hirai](https://huggingface.co/Kotemo428)
- [Kakeru Hattori](https://aya-se.vercel.app/)
- [Masanari Ohi](https://twitter.com/stjohn2007)
- From [YOKOTA Laboratory](https://www.rio.gsic.titech.ac.jp/en/index.html), the following members:
- [Rio Yokota](https://twitter.com/rioyokota)
- [Kazuki Fujii](https://twitter.com/okoge_kaz)
- [Taishi Nakamura](https://twitter.com/Setuna7777_2)
<!-- original-model-card end -->
|
liorfieldwire/question_answer_model
|
liorfieldwire
| 2023-12-19T18:59:41Z | 13 | 0 |
transformers
|
[
"transformers",
"safetensors",
"t5",
"text2text-generation",
"generated_from_trainer",
"base_model:google-t5/t5-small",
"base_model:finetune:google-t5/t5-small",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2023-12-19T14:36:08Z |
---
license: apache-2.0
base_model: t5-small
tags:
- generated_from_trainer
metrics:
- rouge
model-index:
- name: question_answer_model
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# question_answer_model
This model is a fine-tuned version of [t5-small](https://huggingface.co/t5-small) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 1.2860
- Rouge1: 0.2907
- Rouge2: 0.1375
- Rougel: 0.2517
- Rougelsum: 0.2517
- Gen Len: 19.0
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|:-------------:|:-----:|:----:|:---------------:|:------:|:------:|:------:|:---------:|:-------:|
| No log | 1.0 | 104 | 1.3606 | 0.2902 | 0.1333 | 0.2522 | 0.2518 | 19.0 |
| No log | 2.0 | 208 | 1.2995 | 0.2919 | 0.1376 | 0.2528 | 0.2526 | 19.0 |
| No log | 3.0 | 312 | 1.2860 | 0.2907 | 0.1375 | 0.2517 | 0.2517 | 19.0 |
### Framework versions
- Transformers 4.36.2
- Pytorch 2.1.2
- Datasets 2.15.0
- Tokenizers 0.15.0
|
jrdndj/flan-t5-small-samsum
|
jrdndj
| 2023-12-19T18:58:20Z | 4 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"t5",
"text2text-generation",
"generated_from_trainer",
"dataset:samsum",
"base_model:google/flan-t5-small",
"base_model:finetune:google/flan-t5-small",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2023-12-19T18:39:56Z |
---
license: apache-2.0
base_model: google/flan-t5-small
tags:
- generated_from_trainer
datasets:
- samsum
metrics:
- rouge
model-index:
- name: flan-t5-small-samsum
results:
- task:
name: Sequence-to-sequence Language Modeling
type: text2text-generation
dataset:
name: samsum
type: samsum
config: samsum
split: test
args: samsum
metrics:
- name: Rouge1
type: rouge
value: 42.6
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# flan-t5-small-samsum
This model is a fine-tuned version of [google/flan-t5-small](https://huggingface.co/google/flan-t5-small) on the samsum dataset.
It achieves the following results on the evaluation set:
- Loss: 1.6729
- Rouge1: 42.6
- Rouge2: 18.7153
- Rougel: 35.4138
- Rougelsum: 38.8543
- Gen Len: 16.9170
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:-------:|
| 1.8863 | 0.22 | 100 | 1.7049 | 42.0859 | 18.0002 | 34.7349 | 38.3446 | 16.5788 |
| 1.8463 | 0.43 | 200 | 1.6947 | 42.4056 | 18.3005 | 34.9821 | 38.8013 | 17.3614 |
| 1.8548 | 0.65 | 300 | 1.6792 | 42.585 | 18.5643 | 35.2235 | 38.8298 | 17.1514 |
| 1.8358 | 0.87 | 400 | 1.6772 | 42.1544 | 18.2303 | 34.8971 | 38.3609 | 16.5873 |
| 1.8129 | 1.08 | 500 | 1.6729 | 42.6 | 18.7153 | 35.4138 | 38.8543 | 16.9170 |
| 1.8068 | 1.3 | 600 | 1.6709 | 42.5217 | 18.3285 | 35.1455 | 38.5954 | 16.9451 |
| 1.7973 | 1.52 | 700 | 1.6687 | 42.8667 | 18.624 | 35.3429 | 38.9322 | 16.7546 |
| 1.7979 | 1.74 | 800 | 1.6668 | 42.919 | 18.7388 | 35.4528 | 39.0561 | 16.8791 |
| 1.7899 | 1.95 | 900 | 1.6670 | 43.0931 | 18.741 | 35.5047 | 39.2321 | 16.9109 |
### Framework versions
- Transformers 4.36.0
- Pytorch 2.0.0
- Datasets 2.15.0
- Tokenizers 0.15.0
|
TheBloke/Yi-34B-200K-DARE-merge-v5-GGUF
|
TheBloke
| 2023-12-19T18:54:52Z | 44 | 6 |
transformers
|
[
"transformers",
"gguf",
"yi",
"text-generation-inference",
"merge",
"text-generation",
"en",
"base_model:brucethemoose/Yi-34B-200K-DARE-merge-v5",
"base_model:quantized:brucethemoose/Yi-34B-200K-DARE-merge-v5",
"license:other",
"region:us"
] |
text-generation
| 2023-12-19T18:28:15Z |
---
base_model: brucethemoose/Yi-34B-200K-DARE-merge-v5
inference: false
language:
- en
library_name: transformers
license: other
license_link: https://huggingface.co/01-ai/Yi-34B/blob/main/LICENSE
license_name: yi-license
model_creator: brucethemoose
model_name: Yi 34B 200K DARE Merge v5
model_type: yi
pipeline_tag: text-generation
prompt_template: 'SYSTEM: {system_message}
USER: {prompt}
ASSISTANT:
'
quantized_by: TheBloke
tags:
- text-generation-inference
- merge
---
<!-- markdownlint-disable MD041 -->
<!-- header start -->
<!-- 200823 -->
<div style="width: auto; margin-left: auto; margin-right: auto">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<div style="text-align:center; margin-top: 0em; margin-bottom: 0em"><p style="margin-top: 0.25em; margin-bottom: 0em;">TheBloke's LLM work is generously supported by a grant from <a href="https://a16z.com">andreessen horowitz (a16z)</a></p></div>
<hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
<!-- header end -->
# Yi 34B 200K DARE Merge v5 - GGUF
- Model creator: [brucethemoose](https://huggingface.co/brucethemoose)
- Original model: [Yi 34B 200K DARE Merge v5](https://huggingface.co/brucethemoose/Yi-34B-200K-DARE-merge-v5)
<!-- description start -->
## Description
This repo contains GGUF format model files for [brucethemoose's Yi 34B 200K DARE Merge v5](https://huggingface.co/brucethemoose/Yi-34B-200K-DARE-merge-v5).
These files were quantised using hardware kindly provided by [Massed Compute](https://massedcompute.com/).
<!-- description end -->
<!-- README_GGUF.md-about-gguf start -->
### About GGUF
GGUF is a new format introduced by the llama.cpp team on August 21st 2023. It is a replacement for GGML, which is no longer supported by llama.cpp.
Here is an incomplete list of clients and libraries that are known to support GGUF:
* [llama.cpp](https://github.com/ggerganov/llama.cpp). The source project for GGUF. Offers a CLI and a server option.
* [text-generation-webui](https://github.com/oobabooga/text-generation-webui), the most widely used web UI, with many features and powerful extensions. Supports GPU acceleration.
* [KoboldCpp](https://github.com/LostRuins/koboldcpp), a fully featured web UI, with GPU accel across all platforms and GPU architectures. Especially good for story telling.
* [GPT4All](https://gpt4all.io/index.html), a free and open source local running GUI, supporting Windows, Linux and macOS with full GPU accel.
* [LM Studio](https://lmstudio.ai/), an easy-to-use and powerful local GUI for Windows and macOS (Silicon), with GPU acceleration. Linux available, in beta as of 27/11/2023.
* [LoLLMS Web UI](https://github.com/ParisNeo/lollms-webui), a great web UI with many interesting and unique features, including a full model library for easy model selection.
* [Faraday.dev](https://faraday.dev/), an attractive and easy to use character-based chat GUI for Windows and macOS (both Silicon and Intel), with GPU acceleration.
* [llama-cpp-python](https://github.com/abetlen/llama-cpp-python), a Python library with GPU accel, LangChain support, and OpenAI-compatible API server.
* [candle](https://github.com/huggingface/candle), a Rust ML framework with a focus on performance, including GPU support, and ease of use.
* [ctransformers](https://github.com/marella/ctransformers), a Python library with GPU accel, LangChain support, and OpenAI-compatible AI server. Note, as of time of writing (November 27th 2023), ctransformers has not been updated in a long time and does not support many recent models.
<!-- README_GGUF.md-about-gguf end -->
<!-- repositories-available start -->
## Repositories available
* [AWQ model(s) for GPU inference.](https://huggingface.co/TheBloke/Yi-34B-200K-DARE-merge-v5-AWQ)
* [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/Yi-34B-200K-DARE-merge-v5-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/Yi-34B-200K-DARE-merge-v5-GGUF)
* [brucethemoose's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/brucethemoose/Yi-34B-200K-DARE-merge-v5)
<!-- repositories-available end -->
<!-- prompt-template start -->
## Prompt template: Orca-Vicuna
```
SYSTEM: {system_message}
USER: {prompt}
ASSISTANT:
```
<!-- prompt-template end -->
<!-- compatibility_gguf start -->
## Compatibility
These quantised GGUFv2 files are compatible with llama.cpp from August 27th onwards, as of commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221)
They are also compatible with many third party UIs and libraries - please see the list at the top of this README.
## Explanation of quantisation methods
<details>
<summary>Click to see details</summary>
The new methods available are:
* GGML_TYPE_Q2_K - "type-1" 2-bit quantization in super-blocks containing 16 blocks, each block having 16 weight. Block scales and mins are quantized with 4 bits. This ends up effectively using 2.5625 bits per weight (bpw)
* GGML_TYPE_Q3_K - "type-0" 3-bit quantization in super-blocks containing 16 blocks, each block having 16 weights. Scales are quantized with 6 bits. This end up using 3.4375 bpw.
* GGML_TYPE_Q4_K - "type-1" 4-bit quantization in super-blocks containing 8 blocks, each block having 32 weights. Scales and mins are quantized with 6 bits. This ends up using 4.5 bpw.
* GGML_TYPE_Q5_K - "type-1" 5-bit quantization. Same super-block structure as GGML_TYPE_Q4_K resulting in 5.5 bpw
* GGML_TYPE_Q6_K - "type-0" 6-bit quantization. Super-blocks with 16 blocks, each block having 16 weights. Scales are quantized with 8 bits. This ends up using 6.5625 bpw
Refer to the Provided Files table below to see what files use which methods, and how.
</details>
<!-- compatibility_gguf end -->
<!-- README_GGUF.md-provided-files start -->
## Provided files
| Name | Quant method | Bits | Size | Max RAM required | Use case |
| ---- | ---- | ---- | ---- | ---- | ----- |
| [yi-34b-200k-dare-merge-v5.Q2_K.gguf](https://huggingface.co/TheBloke/Yi-34B-200K-DARE-merge-v5-GGUF/blob/main/yi-34b-200k-dare-merge-v5.Q2_K.gguf) | Q2_K | 2 | 14.56 GB| 17.06 GB | smallest, significant quality loss - not recommended for most purposes |
| [yi-34b-200k-dare-merge-v5.Q3_K_S.gguf](https://huggingface.co/TheBloke/Yi-34B-200K-DARE-merge-v5-GGUF/blob/main/yi-34b-200k-dare-merge-v5.Q3_K_S.gguf) | Q3_K_S | 3 | 14.96 GB| 17.46 GB | very small, high quality loss |
| [yi-34b-200k-dare-merge-v5.Q3_K_M.gguf](https://huggingface.co/TheBloke/Yi-34B-200K-DARE-merge-v5-GGUF/blob/main/yi-34b-200k-dare-merge-v5.Q3_K_M.gguf) | Q3_K_M | 3 | 16.64 GB| 19.14 GB | very small, high quality loss |
| [yi-34b-200k-dare-merge-v5.Q3_K_L.gguf](https://huggingface.co/TheBloke/Yi-34B-200K-DARE-merge-v5-GGUF/blob/main/yi-34b-200k-dare-merge-v5.Q3_K_L.gguf) | Q3_K_L | 3 | 18.14 GB| 20.64 GB | small, substantial quality loss |
| [yi-34b-200k-dare-merge-v5.Q4_0.gguf](https://huggingface.co/TheBloke/Yi-34B-200K-DARE-merge-v5-GGUF/blob/main/yi-34b-200k-dare-merge-v5.Q4_0.gguf) | Q4_0 | 4 | 19.47 GB| 21.97 GB | legacy; small, very high quality loss - prefer using Q3_K_M |
| [yi-34b-200k-dare-merge-v5.Q4_K_S.gguf](https://huggingface.co/TheBloke/Yi-34B-200K-DARE-merge-v5-GGUF/blob/main/yi-34b-200k-dare-merge-v5.Q4_K_S.gguf) | Q4_K_S | 4 | 19.55 GB| 22.05 GB | small, greater quality loss |
| [yi-34b-200k-dare-merge-v5.Q4_K_M.gguf](https://huggingface.co/TheBloke/Yi-34B-200K-DARE-merge-v5-GGUF/blob/main/yi-34b-200k-dare-merge-v5.Q4_K_M.gguf) | Q4_K_M | 4 | 20.66 GB| 23.16 GB | medium, balanced quality - recommended |
| [yi-34b-200k-dare-merge-v5.Q5_0.gguf](https://huggingface.co/TheBloke/Yi-34B-200K-DARE-merge-v5-GGUF/blob/main/yi-34b-200k-dare-merge-v5.Q5_0.gguf) | Q5_0 | 5 | 23.71 GB| 26.21 GB | legacy; medium, balanced quality - prefer using Q4_K_M |
| [yi-34b-200k-dare-merge-v5.Q5_K_S.gguf](https://huggingface.co/TheBloke/Yi-34B-200K-DARE-merge-v5-GGUF/blob/main/yi-34b-200k-dare-merge-v5.Q5_K_S.gguf) | Q5_K_S | 5 | 23.71 GB| 26.21 GB | large, low quality loss - recommended |
| [yi-34b-200k-dare-merge-v5.Q5_K_M.gguf](https://huggingface.co/TheBloke/Yi-34B-200K-DARE-merge-v5-GGUF/blob/main/yi-34b-200k-dare-merge-v5.Q5_K_M.gguf) | Q5_K_M | 5 | 24.32 GB| 26.82 GB | large, very low quality loss - recommended |
| [yi-34b-200k-dare-merge-v5.Q6_K.gguf](https://huggingface.co/TheBloke/Yi-34B-200K-DARE-merge-v5-GGUF/blob/main/yi-34b-200k-dare-merge-v5.Q6_K.gguf) | Q6_K | 6 | 28.22 GB| 30.72 GB | very large, extremely low quality loss |
| [yi-34b-200k-dare-merge-v5.Q8_0.gguf](https://huggingface.co/TheBloke/Yi-34B-200K-DARE-merge-v5-GGUF/blob/main/yi-34b-200k-dare-merge-v5.Q8_0.gguf) | Q8_0 | 8 | 36.54 GB| 39.04 GB | very large, extremely low quality loss - not recommended |
**Note**: the above RAM figures assume no GPU offloading. If layers are offloaded to the GPU, this will reduce RAM usage and use VRAM instead.
<!-- README_GGUF.md-provided-files end -->
<!-- README_GGUF.md-how-to-download start -->
## How to download GGUF files
**Note for manual downloaders:** You almost never want to clone the entire repo! Multiple different quantisation formats are provided, and most users only want to pick and download a single file.
The following clients/libraries will automatically download models for you, providing a list of available models to choose from:
* LM Studio
* LoLLMS Web UI
* Faraday.dev
### In `text-generation-webui`
Under Download Model, you can enter the model repo: TheBloke/Yi-34B-200K-DARE-merge-v5-GGUF and below it, a specific filename to download, such as: yi-34b-200k-dare-merge-v5.Q4_K_M.gguf.
Then click Download.
### On the command line, including multiple files at once
I recommend using the `huggingface-hub` Python library:
```shell
pip3 install huggingface-hub
```
Then you can download any individual model file to the current directory, at high speed, with a command like this:
```shell
huggingface-cli download TheBloke/Yi-34B-200K-DARE-merge-v5-GGUF yi-34b-200k-dare-merge-v5.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False
```
<details>
<summary>More advanced huggingface-cli download usage (click to read)</summary>
You can also download multiple files at once with a pattern:
```shell
huggingface-cli download TheBloke/Yi-34B-200K-DARE-merge-v5-GGUF --local-dir . --local-dir-use-symlinks False --include='*Q4_K*gguf'
```
For more documentation on downloading with `huggingface-cli`, please see: [HF -> Hub Python Library -> Download files -> Download from the CLI](https://huggingface.co/docs/huggingface_hub/guides/download#download-from-the-cli).
To accelerate downloads on fast connections (1Gbit/s or higher), install `hf_transfer`:
```shell
pip3 install hf_transfer
```
And set environment variable `HF_HUB_ENABLE_HF_TRANSFER` to `1`:
```shell
HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download TheBloke/Yi-34B-200K-DARE-merge-v5-GGUF yi-34b-200k-dare-merge-v5.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False
```
Windows Command Line users: You can set the environment variable by running `set HF_HUB_ENABLE_HF_TRANSFER=1` before the download command.
</details>
<!-- README_GGUF.md-how-to-download end -->
<!-- README_GGUF.md-how-to-run start -->
## Example `llama.cpp` command
Make sure you are using `llama.cpp` from commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221) or later.
```shell
./main -ngl 35 -m yi-34b-200k-dare-merge-v5.Q4_K_M.gguf --color -c 200000 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "SYSTEM: {system_message}\nUSER: {prompt}\nASSISTANT:"
```
Change `-ngl 32` to the number of layers to offload to GPU. Remove it if you don't have GPU acceleration.
Change `-c 200000` to the desired sequence length. For extended sequence models - eg 8K, 16K, 32K - the necessary RoPE scaling parameters are read from the GGUF file and set by llama.cpp automatically. Note that longer sequence lengths require much more resources, so you may need to reduce this value.
If you want to have a chat-style conversation, replace the `-p <PROMPT>` argument with `-i -ins`
For other parameters and how to use them, please refer to [the llama.cpp documentation](https://github.com/ggerganov/llama.cpp/blob/master/examples/main/README.md)
## How to run in `text-generation-webui`
Further instructions can be found in the text-generation-webui documentation, here: [text-generation-webui/docs/04 ‐ Model Tab.md](https://github.com/oobabooga/text-generation-webui/blob/main/docs/04%20%E2%80%90%20Model%20Tab.md#llamacpp).
## How to run from Python code
You can use GGUF models from Python using the [llama-cpp-python](https://github.com/abetlen/llama-cpp-python) or [ctransformers](https://github.com/marella/ctransformers) libraries. Note that at the time of writing (Nov 27th 2023), ctransformers has not been updated for some time and is not compatible with some recent models. Therefore I recommend you use llama-cpp-python.
### How to load this model in Python code, using llama-cpp-python
For full documentation, please see: [llama-cpp-python docs](https://abetlen.github.io/llama-cpp-python/).
#### First install the package
Run one of the following commands, according to your system:
```shell
# Base ctransformers with no GPU acceleration
pip install llama-cpp-python
# With NVidia CUDA acceleration
CMAKE_ARGS="-DLLAMA_CUBLAS=on" pip install llama-cpp-python
# Or with OpenBLAS acceleration
CMAKE_ARGS="-DLLAMA_BLAS=ON -DLLAMA_BLAS_VENDOR=OpenBLAS" pip install llama-cpp-python
# Or with CLBLast acceleration
CMAKE_ARGS="-DLLAMA_CLBLAST=on" pip install llama-cpp-python
# Or with AMD ROCm GPU acceleration (Linux only)
CMAKE_ARGS="-DLLAMA_HIPBLAS=on" pip install llama-cpp-python
# Or with Metal GPU acceleration for macOS systems only
CMAKE_ARGS="-DLLAMA_METAL=on" pip install llama-cpp-python
# In windows, to set the variables CMAKE_ARGS in PowerShell, follow this format; eg for NVidia CUDA:
$env:CMAKE_ARGS = "-DLLAMA_OPENBLAS=on"
pip install llama-cpp-python
```
#### Simple llama-cpp-python example code
```python
from llama_cpp import Llama
# Set gpu_layers to the number of layers to offload to GPU. Set to 0 if no GPU acceleration is available on your system.
llm = Llama(
model_path="./yi-34b-200k-dare-merge-v5.Q4_K_M.gguf", # Download the model file first
n_ctx=200000, # The max sequence length to use - note that longer sequence lengths require much more resources
n_threads=8, # The number of CPU threads to use, tailor to your system and the resulting performance
n_gpu_layers=35 # The number of layers to offload to GPU, if you have GPU acceleration available
)
# Simple inference example
output = llm(
"SYSTEM: {system_message}\nUSER: {prompt}\nASSISTANT:", # Prompt
max_tokens=512, # Generate up to 512 tokens
stop=["</s>"], # Example stop token - not necessarily correct for this specific model! Please check before using.
echo=True # Whether to echo the prompt
)
# Chat Completion API
llm = Llama(model_path="./yi-34b-200k-dare-merge-v5.Q4_K_M.gguf", chat_format="llama-2") # Set chat_format according to the model you are using
llm.create_chat_completion(
messages = [
{"role": "system", "content": "You are a story writing assistant."},
{
"role": "user",
"content": "Write a story about llamas."
}
]
)
```
## How to use with LangChain
Here are guides on using llama-cpp-python and ctransformers with LangChain:
* [LangChain + llama-cpp-python](https://python.langchain.com/docs/integrations/llms/llamacpp)
* [LangChain + ctransformers](https://python.langchain.com/docs/integrations/providers/ctransformers)
<!-- README_GGUF.md-how-to-run end -->
<!-- footer start -->
<!-- 200823 -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute
Thanks to the [chirper.ai](https://chirper.ai) team!
Thanks to Clay from [gpus.llm-utils.org](llm-utils)!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Aemon Algiz.
**Patreon special mentions**: Michael Levine, 阿明, Trailburnt, Nikolai Manek, John Detwiler, Randy H, Will Dee, Sebastain Graf, NimbleBox.ai, Eugene Pentland, Emad Mostaque, Ai Maven, Jim Angel, Jeff Scroggin, Michael Davis, Manuel Alberto Morcote, Stephen Murray, Robert, Justin Joy, Luke @flexchar, Brandon Frisco, Elijah Stavena, S_X, Dan Guido, Undi ., Komninos Chatzipapas, Shadi, theTransient, Lone Striker, Raven Klaugh, jjj, Cap'n Zoog, Michel-Marie MAUDET (LINAGORA), Matthew Berman, David, Fen Risland, Omer Bin Jawed, Luke Pendergrass, Kalila, OG, Erik Bjäreholt, Rooh Singh, Joseph William Delisle, Dan Lewis, TL, John Villwock, AzureBlack, Brad, Pedro Madruga, Caitlyn Gatomon, K, jinyuan sun, Mano Prime, Alex, Jeffrey Morgan, Alicia Loh, Illia Dulskyi, Chadd, transmissions 11, fincy, Rainer Wilmers, ReadyPlayerEmma, knownsqashed, Mandus, biorpg, Deo Leter, Brandon Phillips, SuperWojo, Sean Connelly, Iucharbius, Jack West, Harry Royden McLaughlin, Nicholas, terasurfer, Vitor Caleffi, Duane Dunston, Johann-Peter Hartmann, David Ziegler, Olakabola, Ken Nordquist, Trenton Dambrowitz, Tom X Nguyen, Vadim, Ajan Kanaga, Leonard Tan, Clay Pascal, Alexandros Triantafyllidis, JM33133, Xule, vamX, ya boyyy, subjectnull, Talal Aujan, Alps Aficionado, wassieverse, Ari Malik, James Bentley, Woland, Spencer Kim, Michael Dempsey, Fred von Graf, Elle, zynix, William Richards, Stanislav Ovsiannikov, Edmond Seymore, Jonathan Leane, Martin Kemka, usrbinkat, Enrico Ros
Thank you to all my generous patrons and donaters!
And thank you again to a16z for their generous grant.
<!-- footer end -->
<!-- original-model-card start -->
# Original model card: brucethemoose's Yi 34B 200K DARE Merge v5
[**Nous-Capybara-34B**](https://huggingface.co/NousResearch/Nous-Capybara-34B/), [**Tess-M-v1.4**](https://huggingface.co/migtissera/Tess-34B-v1.4), [**Airoboros-3_1-yi-34b-200k**](https://huggingface.co/bhenrym14/airoboros-3_1-yi-34b-200k), [**PlatYi-34B-200K-Q**](https://huggingface.co/kyujinpy/PlatYi-34B-200k-Q-FastChat), [**Pallas-0.4**](https://huggingface.co/Mihaiii/Pallas-0.4), [**Yi-34B-200K-AEZAKMI-v2**](https://huggingface.co/adamo1139/Yi-34B-200K-AEZAKMI-v2), and a tiny bit of [**SUS-Chat-34B**](https://huggingface.co/SUSTech/SUS-Chat-34B) merged with a new, experimental implementation of "dare ties" via mergekit. See:
> [Language Models are Super Mario: Absorbing Abilities from Homologous Models as a Free Lunch](https://github.com/yule-BUAA/MergeLM)
> https://github.com/cg123/mergekit/tree/dare
***
## Prompt template: Orca-Vicuna
```
SYSTEM: {system_message}
USER: {prompt}
ASSISTANT:
```
It might recognize ChatML, or maybe Llama-chat from Airoboros.
Sometimes the model "spells out" the stop token as `</s>` like Capybara, so you may need to add `</s>` as an additional stopping condition.
***
## Running
Being a Yi model, try running a lower temperature with 0.05-0.1 MinP, a little repetition penalty, and no other samplers. Yi tends to run "hot" by default, and it really needs MinP to cull the huge vocabulary.
24GB GPUs can run Yi-34B-200K models at **45K-75K context** with exllamav2, and performant UIs like [exui](https://github.com/turboderp/exui). I go into more detail in this [post](https://old.reddit.com/r/LocalLLaMA/comments/1896igc/how_i_run_34b_models_at_75k_context_on_24gb_fast/)
I recommend exl2 quantizations profiled on data similar to the desired task. It is especially sensitive to the quantization data at low bpw. I've published my own fiction-oriented quantizations here: https://huggingface.co/collections/brucethemoose/most-recent-merge-65742644ca03b6c514afa204
To load this in full-context backends like transformers, you *must* change `max_position_embeddings` in config.json to a lower value than 200,000, otherwise you will OOM!
***
## Testing Notes
Merged in mergekit with the following config, and the tokenizer from chargoddard's Yi-Llama:
```
models:
- model: /home/alpha/Storage/Models/Raw/chargoddard_Yi-34B-200K-Llama
# No parameters necessary for base model
- model: /home/alpha/Storage/Models/Raw/migtissera_Tess-34B-v1.4
# Less weight than previous merge since Pallas is a finetune of Tess
parameters:
weight: 0.14
density: 0.62
- model: /home/alpha/FastModels/Mihaiii_Pallas-0.4
parameters:
weight: 0.14
density: 0.62
- model: /home/alpha//Storage/Models/Raw/bhenrym14_airoboros-3_1-yi-34b-200k
parameters:
weight: 0.14
density: 0.52
- model: /home/alpha/Storage/Models/Raw/Nous-Capybara-34B
parameters:
weight: 0.22
density: 0.62
- model: /home/alpha/Storage/Models/Raw/kyujinpy_PlatYi-34B-200k-Q-FastChat
parameters:
weight: 0.14
density: 0.52
#- model: /home/alpha/Storage/Models/Raw/ehartford_dolphin-2.2-yi-34b-200k
# Dolphin 200K seems to be broken according to multiple leaderboards and perplexity tests?
# parameters:
# weight: 0.15
# density: 0.6
- model: /home/alpha/Models/Raw/adamo1139_Yi-34B-200K-AEZAKMI-v2
parameters:
weight: 0.14
density: 0.52
- model: /home/alpha/Models/Raw/SUSTech_SUS-Chat-34B/
# Very low density and low weight since its a Yi 4K finetune, to try and preserve long context performance while "keeping" some of SUS
parameters:
weight: 0.08
density: 0.08
merge_method: dare_ties
base_model: /home/alpha/Storage/Models/Raw/chargoddard_Yi-34B-200K-Llama
parameters:
int8_mask: true
dtype: bfloat16
```
Various densities were tested with perplexity tests and long context prompts. Relatively high densities seem to perform better, contrary to the findings of the Super Mario paper.
This particular version is merged with more than the "recommended" max density of 0.5. It seems to result in even better perplexity, but I'm not sure if this translates to better output.
Weights that add up to 1 seems to be optimal.
Dare Ties is also resulting in seemingly better, lower perplexity merges than a regular ties merge, task arithmetic or a slerp merge.
SUS Chat is not a 200K model, hence it was merged at a very low density to try and preserve Yi 200K's long context performance while still inheriting some of SUS's performance.
Dolphin 200K was taken out of this merge because it seems to be performing poorly for a 34B Dolphin model, like something went wrong during training?
I chose not to include other finetunes because they aren't trained on the 200K base. If any other 200K finetunes pop up, let me know.
***
## Credits:
https://github.com/cg123/mergekit/tree/dare
https://huggingface.co/NousResearch/Nous-Capybara-34B/
https://huggingface.co/bhenrym14/airoboros-3_1-yi-34b-200k
https://huggingface.co/migtissera/Tess-M-v1.4
https://huggingface.co/kyujinpy/PlatYi-34B-200k-Q-FastChat
https://huggingface.co/adamo1139/Yi-34B-200K-AEZAKMI-v2
https://huggingface.co/Mihaiii/Pallas-0.4
https://huggingface.co/SUSTech/SUS-Chat-34B
https://huggingface.co/chargoddard/Yi-34B-200K-Llama
https://huggingface.co/01-ai/Yi-34B-200K
<!-- original-model-card end -->
|
SovaGamer/flan-t5-small-samsum
|
SovaGamer
| 2023-12-19T18:50:38Z | 5 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"t5",
"text2text-generation",
"generated_from_trainer",
"dataset:samsum",
"base_model:google/flan-t5-small",
"base_model:finetune:google/flan-t5-small",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2023-12-19T18:31:10Z |
---
license: apache-2.0
base_model: google/flan-t5-small
tags:
- generated_from_trainer
datasets:
- samsum
metrics:
- rouge
model-index:
- name: flan-t5-small-samsum
results:
- task:
name: Sequence-to-sequence Language Modeling
type: text2text-generation
dataset:
name: samsum
type: samsum
config: samsum
split: test
args: samsum
metrics:
- name: Rouge1
type: rouge
value: 42.6775
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# flan-t5-small-samsum
This model is a fine-tuned version of [google/flan-t5-small](https://huggingface.co/google/flan-t5-small) on the samsum dataset.
It achieves the following results on the evaluation set:
- Loss: 1.6754
- Rouge1: 42.6775
- Rouge2: 18.3775
- Rougel: 35.281
- Rougelsum: 38.9344
- Gen Len: 16.8474
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 52
- eval_batch_size: 52
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:-------:|
| 1.8824 | 0.35 | 100 | 1.7015 | 42.4867 | 18.3364 | 35.0855 | 38.8376 | 16.6532 |
| 1.8578 | 0.7 | 200 | 1.6878 | 42.0066 | 18.2434 | 34.9702 | 38.5065 | 16.7216 |
| 1.835 | 1.06 | 300 | 1.6823 | 42.7374 | 18.6187 | 35.4371 | 38.993 | 16.9048 |
| 1.8144 | 1.41 | 400 | 1.6786 | 42.6189 | 18.3792 | 35.3473 | 38.9192 | 16.6618 |
| 1.8094 | 1.76 | 500 | 1.6754 | 42.6775 | 18.3775 | 35.281 | 38.9344 | 16.8474 |
### Framework versions
- Transformers 4.36.0
- Pytorch 2.0.0
- Datasets 2.15.0
- Tokenizers 0.15.0
|
mrupar/flan-t5-small-samsum
|
mrupar
| 2023-12-19T18:44:17Z | 5 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"t5",
"text2text-generation",
"generated_from_trainer",
"dataset:samsum",
"base_model:google/flan-t5-small",
"base_model:finetune:google/flan-t5-small",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2023-12-19T18:25:59Z |
---
license: apache-2.0
base_model: google/flan-t5-small
tags:
- generated_from_trainer
datasets:
- samsum
metrics:
- rouge
model-index:
- name: flan-t5-small-samsum
results:
- task:
name: Sequence-to-sequence Language Modeling
type: text2text-generation
dataset:
name: samsum
type: samsum
config: samsum
split: test
args: samsum
metrics:
- name: Rouge1
type: rouge
value: 42.6693
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# flan-t5-small-samsum
This model is a fine-tuned version of [google/flan-t5-small](https://huggingface.co/google/flan-t5-small) on the samsum dataset.
It achieves the following results on the evaluation set:
- Loss: 1.6754
- Rouge1: 42.6693
- Rouge2: 18.3378
- Rougel: 35.2729
- Rougelsum: 38.9033
- Gen Len: 16.8474
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 52
- eval_batch_size: 52
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:-------:|
| 1.8824 | 0.35 | 100 | 1.7015 | 42.4703 | 18.3068 | 35.1199 | 38.8083 | 16.6532 |
| 1.8578 | 0.7 | 200 | 1.6878 | 42.0064 | 18.2236 | 34.9497 | 38.4611 | 16.7216 |
| 1.835 | 1.06 | 300 | 1.6823 | 42.7407 | 18.5955 | 35.4344 | 38.9663 | 16.9048 |
| 1.8144 | 1.41 | 400 | 1.6786 | 42.6272 | 18.3894 | 35.34 | 38.8868 | 16.6618 |
| 1.8094 | 1.76 | 500 | 1.6754 | 42.6693 | 18.3378 | 35.2729 | 38.9033 | 16.8474 |
### Framework versions
- Transformers 4.36.0
- Pytorch 2.0.0
- Datasets 2.15.0
- Tokenizers 0.15.0
|
temova/flan-t5-small-samsum
|
temova
| 2023-12-19T18:44:11Z | 5 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"t5",
"text2text-generation",
"generated_from_trainer",
"dataset:samsum",
"base_model:google/flan-t5-small",
"base_model:finetune:google/flan-t5-small",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2023-12-19T18:24:33Z |
---
license: apache-2.0
base_model: google/flan-t5-small
tags:
- generated_from_trainer
datasets:
- samsum
metrics:
- rouge
model-index:
- name: flan-t5-small-samsum
results:
- task:
name: Sequence-to-sequence Language Modeling
type: text2text-generation
dataset:
name: samsum
type: samsum
config: samsum
split: test
args: samsum
metrics:
- name: Rouge1
type: rouge
value: 42.6907
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# flan-t5-small-samsum
This model is a fine-tuned version of [google/flan-t5-small](https://huggingface.co/google/flan-t5-small) on the samsum dataset.
It achieves the following results on the evaluation set:
- Loss: 1.6754
- Rouge1: 42.6907
- Rouge2: 18.3626
- Rougel: 35.2723
- Rougelsum: 38.9062
- Gen Len: 16.8474
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 52
- eval_batch_size: 52
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:-------:|
| 1.8824 | 0.35 | 100 | 1.7015 | 42.4935 | 18.3634 | 35.0823 | 38.8358 | 16.6532 |
| 1.8578 | 0.7 | 200 | 1.6878 | 42.0329 | 18.2685 | 34.9421 | 38.4636 | 16.7216 |
| 1.835 | 1.06 | 300 | 1.6823 | 42.7493 | 18.6379 | 35.4001 | 38.9845 | 16.9048 |
| 1.8144 | 1.41 | 400 | 1.6786 | 42.6157 | 18.4093 | 35.3149 | 38.8787 | 16.6618 |
| 1.8094 | 1.76 | 500 | 1.6754 | 42.6907 | 18.3626 | 35.2723 | 38.9062 | 16.8474 |
### Framework versions
- Transformers 4.36.0
- Pytorch 2.0.0
- Datasets 2.15.0
- Tokenizers 0.15.0
|
ElEnAaaaaaa/flan-t5-small-samsum
|
ElEnAaaaaaa
| 2023-12-19T18:44:09Z | 4 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"t5",
"text2text-generation",
"generated_from_trainer",
"dataset:samsum",
"base_model:google/flan-t5-small",
"base_model:finetune:google/flan-t5-small",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2023-12-19T18:24:39Z |
---
license: apache-2.0
base_model: google/flan-t5-small
tags:
- generated_from_trainer
datasets:
- samsum
metrics:
- rouge
model-index:
- name: flan-t5-small-samsum
results:
- task:
name: Sequence-to-sequence Language Modeling
type: text2text-generation
dataset:
name: samsum
type: samsum
config: samsum
split: test
args: samsum
metrics:
- name: Rouge1
type: rouge
value: 42.7055
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# flan-t5-small-samsum
This model is a fine-tuned version of [google/flan-t5-small](https://huggingface.co/google/flan-t5-small) on the samsum dataset.
It achieves the following results on the evaluation set:
- Loss: 1.6754
- Rouge1: 42.7055
- Rouge2: 18.3564
- Rougel: 35.2909
- Rougelsum: 38.9643
- Gen Len: 16.8474
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 52
- eval_batch_size: 52
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:-------:|
| 1.8824 | 0.35 | 100 | 1.7015 | 42.5036 | 18.333 | 35.1513 | 38.8251 | 16.6532 |
| 1.8578 | 0.7 | 200 | 1.6878 | 42.0047 | 18.2467 | 35.0002 | 38.5021 | 16.7216 |
| 1.835 | 1.06 | 300 | 1.6823 | 42.7828 | 18.6243 | 35.4419 | 38.9984 | 16.9048 |
| 1.8144 | 1.41 | 400 | 1.6786 | 42.6579 | 18.3971 | 35.3512 | 38.9118 | 16.6618 |
| 1.8094 | 1.76 | 500 | 1.6754 | 42.7055 | 18.3564 | 35.2909 | 38.9643 | 16.8474 |
### Framework versions
- Transformers 4.36.0
- Pytorch 2.0.0
- Datasets 2.15.0
- Tokenizers 0.15.0
|
Petrovilija/flan-t5-small-samsum
|
Petrovilija
| 2023-12-19T18:44:00Z | 4 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"t5",
"text2text-generation",
"generated_from_trainer",
"dataset:samsum",
"base_model:google/flan-t5-small",
"base_model:finetune:google/flan-t5-small",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2023-12-19T18:24:40Z |
---
license: apache-2.0
base_model: google/flan-t5-small
tags:
- generated_from_trainer
datasets:
- samsum
metrics:
- rouge
model-index:
- name: flan-t5-small-samsum
results:
- task:
name: Sequence-to-sequence Language Modeling
type: text2text-generation
dataset:
name: samsum
type: samsum
config: samsum
split: test
args: samsum
metrics:
- name: Rouge1
type: rouge
value: 42.739
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# flan-t5-small-samsum
This model is a fine-tuned version of [google/flan-t5-small](https://huggingface.co/google/flan-t5-small) on the samsum dataset.
It achieves the following results on the evaluation set:
- Loss: 1.6754
- Rouge1: 42.739
- Rouge2: 18.3741
- Rougel: 35.2588
- Rougelsum: 38.893
- Gen Len: 16.8474
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 52
- eval_batch_size: 52
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:-------:|
| 1.8824 | 0.35 | 100 | 1.7015 | 42.5324 | 18.3468 | 35.0528 | 38.7814 | 16.6532 |
| 1.8578 | 0.7 | 200 | 1.6878 | 42.0766 | 18.2423 | 34.9442 | 38.4806 | 16.7216 |
| 1.835 | 1.06 | 300 | 1.6823 | 42.8147 | 18.6292 | 35.4054 | 38.956 | 16.9048 |
| 1.8144 | 1.41 | 400 | 1.6786 | 42.6886 | 18.402 | 35.3235 | 38.8638 | 16.6618 |
| 1.8094 | 1.76 | 500 | 1.6754 | 42.739 | 18.3741 | 35.2588 | 38.893 | 16.8474 |
### Framework versions
- Transformers 4.36.0
- Pytorch 2.0.0
- Datasets 2.15.0
- Tokenizers 0.15.0
|
nikolamurgo/flan-t5-small-samsum
|
nikolamurgo
| 2023-12-19T18:43:59Z | 4 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"t5",
"text2text-generation",
"generated_from_trainer",
"dataset:samsum",
"base_model:google/flan-t5-small",
"base_model:finetune:google/flan-t5-small",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2023-12-19T18:24:35Z |
---
license: apache-2.0
base_model: google/flan-t5-small
tags:
- generated_from_trainer
datasets:
- samsum
metrics:
- rouge
model-index:
- name: flan-t5-small-samsum
results:
- task:
name: Sequence-to-sequence Language Modeling
type: text2text-generation
dataset:
name: samsum
type: samsum
config: samsum
split: test
args: samsum
metrics:
- name: Rouge1
type: rouge
value: 42.6857
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# flan-t5-small-samsum
This model is a fine-tuned version of [google/flan-t5-small](https://huggingface.co/google/flan-t5-small) on the samsum dataset.
It achieves the following results on the evaluation set:
- Loss: 1.6754
- Rouge1: 42.6857
- Rouge2: 18.3487
- Rougel: 35.3275
- Rougelsum: 38.8882
- Gen Len: 16.8474
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 52
- eval_batch_size: 52
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:-------:|
| 1.8824 | 0.35 | 100 | 1.7015 | 42.4897 | 18.3435 | 35.1269 | 38.7613 | 16.6532 |
| 1.8578 | 0.7 | 200 | 1.6878 | 42.0049 | 18.2725 | 34.9941 | 38.4372 | 16.7216 |
| 1.835 | 1.06 | 300 | 1.6823 | 42.7659 | 18.64 | 35.4591 | 38.9559 | 16.9048 |
| 1.8144 | 1.41 | 400 | 1.6786 | 42.61 | 18.3931 | 35.3777 | 38.8516 | 16.6618 |
| 1.8094 | 1.76 | 500 | 1.6754 | 42.6857 | 18.3487 | 35.3275 | 38.8882 | 16.8474 |
### Framework versions
- Transformers 4.36.0
- Pytorch 2.0.0
- Datasets 2.15.0
- Tokenizers 0.15.0
|
zstoimchev/flan-t5-small-samsum
|
zstoimchev
| 2023-12-19T18:43:56Z | 4 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"t5",
"text2text-generation",
"generated_from_trainer",
"dataset:samsum",
"base_model:google/flan-t5-small",
"base_model:finetune:google/flan-t5-small",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2023-12-19T18:24:39Z |
---
license: apache-2.0
base_model: google/flan-t5-small
tags:
- generated_from_trainer
datasets:
- samsum
metrics:
- rouge
model-index:
- name: flan-t5-small-samsum
results:
- task:
name: Sequence-to-sequence Language Modeling
type: text2text-generation
dataset:
name: samsum
type: samsum
config: samsum
split: test
args: samsum
metrics:
- name: Rouge1
type: rouge
value: 42.6976
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# flan-t5-small-samsum
This model is a fine-tuned version of [google/flan-t5-small](https://huggingface.co/google/flan-t5-small) on the samsum dataset.
It achieves the following results on the evaluation set:
- Loss: 1.6754
- Rouge1: 42.6976
- Rouge2: 18.3401
- Rougel: 35.2388
- Rougelsum: 38.9044
- Gen Len: 16.8474
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 52
- eval_batch_size: 52
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:-------:|
| 1.8824 | 0.35 | 100 | 1.7015 | 42.4907 | 18.3171 | 35.0684 | 38.8292 | 16.6532 |
| 1.8578 | 0.7 | 200 | 1.6878 | 42.0288 | 18.2503 | 34.9411 | 38.4881 | 16.7216 |
| 1.835 | 1.06 | 300 | 1.6823 | 42.742 | 18.6174 | 35.3759 | 38.992 | 16.9048 |
| 1.8144 | 1.41 | 400 | 1.6786 | 42.6286 | 18.3873 | 35.3084 | 38.8797 | 16.6618 |
| 1.8094 | 1.76 | 500 | 1.6754 | 42.6976 | 18.3401 | 35.2388 | 38.9044 | 16.8474 |
### Framework versions
- Transformers 4.36.0
- Pytorch 2.0.0
- Datasets 2.15.0
- Tokenizers 0.15.0
|
EmilijaTR/flan-t5-small-samsum
|
EmilijaTR
| 2023-12-19T18:43:55Z | 4 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"t5",
"text2text-generation",
"generated_from_trainer",
"dataset:samsum",
"base_model:google/flan-t5-small",
"base_model:finetune:google/flan-t5-small",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2023-12-19T18:24:30Z |
---
license: apache-2.0
base_model: google/flan-t5-small
tags:
- generated_from_trainer
datasets:
- samsum
metrics:
- rouge
model-index:
- name: flan-t5-small-samsum
results:
- task:
name: Sequence-to-sequence Language Modeling
type: text2text-generation
dataset:
name: samsum
type: samsum
config: samsum
split: test
args: samsum
metrics:
- name: Rouge1
type: rouge
value: 42.6766
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# flan-t5-small-samsum
This model is a fine-tuned version of [google/flan-t5-small](https://huggingface.co/google/flan-t5-small) on the samsum dataset.
It achieves the following results on the evaluation set:
- Loss: 1.6754
- Rouge1: 42.6766
- Rouge2: 18.3823
- Rougel: 35.2377
- Rougelsum: 38.9432
- Gen Len: 16.8474
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 52
- eval_batch_size: 52
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:-------:|
| 1.8824 | 0.35 | 100 | 1.7015 | 42.4882 | 18.3559 | 35.0843 | 38.8515 | 16.6532 |
| 1.8578 | 0.7 | 200 | 1.6878 | 42.0007 | 18.2614 | 34.9474 | 38.4992 | 16.7216 |
| 1.835 | 1.06 | 300 | 1.6823 | 42.7535 | 18.6418 | 35.3807 | 39.0098 | 16.9048 |
| 1.8144 | 1.41 | 400 | 1.6786 | 42.6326 | 18.4197 | 35.2995 | 38.9165 | 16.6618 |
| 1.8094 | 1.76 | 500 | 1.6754 | 42.6766 | 18.3823 | 35.2377 | 38.9432 | 16.8474 |
### Framework versions
- Transformers 4.36.0
- Pytorch 2.0.0
- Datasets 2.15.0
- Tokenizers 0.15.0
|
Veskic/flan-t5-small-samsum
|
Veskic
| 2023-12-19T18:43:52Z | 4 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"t5",
"text2text-generation",
"generated_from_trainer",
"dataset:samsum",
"base_model:google/flan-t5-small",
"base_model:finetune:google/flan-t5-small",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2023-12-19T18:24:33Z |
---
license: apache-2.0
base_model: google/flan-t5-small
tags:
- generated_from_trainer
datasets:
- samsum
model-index:
- name: flan-t5-small-samsum
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# flan-t5-small-samsum
This model is a fine-tuned version of [google/flan-t5-small](https://huggingface.co/google/flan-t5-small) on the samsum dataset.
It achieves the following results on the evaluation set:
- eval_loss: 1.6754
- eval_rouge1: 42.7098
- eval_rouge2: 18.3566
- eval_rougeL: 35.2282
- eval_rougeLsum: 38.9027
- eval_gen_len: 16.8474
- eval_runtime: 23.9949
- eval_samples_per_second: 34.132
- eval_steps_per_second: 0.667
- step: 0
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 52
- eval_batch_size: 52
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
- mixed_precision_training: Native AMP
### Framework versions
- Transformers 4.36.0
- Pytorch 2.0.0
- Datasets 2.15.0
- Tokenizers 0.15.0
|
Wazzzabeee/flan-t5-small-samsum
|
Wazzzabeee
| 2023-12-19T18:43:43Z | 6 | 1 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"t5",
"text2text-generation",
"generated_from_trainer",
"dataset:samsum",
"base_model:google/flan-t5-small",
"base_model:finetune:google/flan-t5-small",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2023-12-19T18:24:29Z |
---
license: apache-2.0
base_model: google/flan-t5-small
tags:
- generated_from_trainer
datasets:
- samsum
metrics:
- rouge
model-index:
- name: flan-t5-small-samsum
results:
- task:
name: Sequence-to-sequence Language Modeling
type: text2text-generation
dataset:
name: samsum
type: samsum
config: samsum
split: test
args: samsum
metrics:
- name: Rouge1
type: rouge
value: 42.6222
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# flan-t5-small-samsum
This model is a fine-tuned version of [google/flan-t5-small](https://huggingface.co/google/flan-t5-small) on the samsum dataset.
It achieves the following results on the evaluation set:
- Loss: 1.6729
- Rouge1: 42.6222
- Rouge2: 18.682
- Rougel: 35.3954
- Rougelsum: 38.9104
- Gen Len: 16.9170
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:-------:|
| 1.8863 | 0.22 | 100 | 1.7049 | 42.1145 | 18.0254 | 34.733 | 38.4052 | 16.5788 |
| 1.8463 | 0.43 | 200 | 1.6947 | 42.4119 | 18.2925 | 34.9702 | 38.8535 | 17.3614 |
| 1.8548 | 0.65 | 300 | 1.6792 | 42.5967 | 18.5244 | 35.1965 | 38.9087 | 17.1514 |
| 1.8358 | 0.87 | 400 | 1.6772 | 42.167 | 18.2032 | 34.8647 | 38.4144 | 16.5873 |
| 1.8129 | 1.08 | 500 | 1.6729 | 42.6222 | 18.682 | 35.3954 | 38.9104 | 16.9170 |
| 1.8068 | 1.3 | 600 | 1.6709 | 42.5238 | 18.311 | 35.1257 | 38.6584 | 16.9451 |
| 1.7973 | 1.52 | 700 | 1.6687 | 42.8715 | 18.6133 | 35.3054 | 38.971 | 16.7546 |
| 1.7979 | 1.74 | 800 | 1.6668 | 42.9038 | 18.7483 | 35.4156 | 39.1118 | 16.8791 |
| 1.7899 | 1.95 | 900 | 1.6670 | 43.1142 | 18.7369 | 35.4796 | 39.2724 | 16.9109 |
### Framework versions
- Transformers 4.36.0
- Pytorch 2.0.0
- Datasets 2.15.0
- Tokenizers 0.15.0
|
jjfumero/distilbert-base-uncased-finetuned-emotions
|
jjfumero
| 2023-12-19T18:39:12Z | 7 | 0 |
transformers
|
[
"transformers",
"safetensors",
"distilbert",
"text-classification",
"generated_from_trainer",
"dataset:emotion",
"base_model:distilbert/distilbert-base-uncased",
"base_model:finetune:distilbert/distilbert-base-uncased",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2023-12-19T18:37:12Z |
---
license: apache-2.0
base_model: distilbert-base-uncased
tags:
- generated_from_trainer
datasets:
- emotion
model-index:
- name: distilbert-base-uncased-finetuned-emotions
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# distilbert-base-uncased-finetuned-emotions
This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the emotion dataset.
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
### Training results
### Framework versions
- Transformers 4.36.2
- Pytorch 2.3.0.dev20231219+cu121
- Datasets 2.15.0
- Tokenizers 0.15.0
|
digiplay/Dolka_Rusalka_v0.5.1
|
digiplay
| 2023-12-19T18:31:42Z | 1,112 | 7 |
diffusers
|
[
"diffusers",
"safetensors",
"stable-diffusion",
"stable-diffusion-diffusers",
"text-to-image",
"license:other",
"autotrain_compatible",
"endpoints_compatible",
"diffusers:StableDiffusionPipeline",
"region:us"
] |
text-to-image
| 2023-12-19T18:05:05Z |
---
license: other
tags:
- stable-diffusion
- stable-diffusion-diffusers
- text-to-image
- diffusers
inference: true
---
Model info:
https://civitai.com/models/236251?modelVersionId=267511
Sample images generated by Hugginface's API:
4k ,lake,duck,1girl,picnic, close up ,
4k ,lake,duck,1girl,picnic, close up , sakura trees,
|
LaVuna47/ppo-LunarLander-v2
|
LaVuna47
| 2023-12-19T18:31:01Z | 0 | 0 |
stable-baselines3
|
[
"stable-baselines3",
"LunarLander-v2",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-12-19T17:36:08Z |
---
library_name: stable-baselines3
tags:
- LunarLander-v2
- deep-reinforcement-learning
- reinforcement-learning
- stable-baselines3
model-index:
- name: PPO
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: LunarLander-v2
type: LunarLander-v2
metrics:
- type: mean_reward
value: 254.22 +/- 27.00
name: mean_reward
verified: false
---
# **PPO** Agent playing **LunarLander-v2**
This is a trained model of a **PPO** agent playing **LunarLander-v2**
using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3).
## Usage (with Stable-baselines3)
TODO: Add your code
```python
from stable_baselines3 import ...
from huggingface_sb3 import load_from_hub
...
```
|
digiplay/YutaMix_realistic_v11
|
digiplay
| 2023-12-19T18:26:59Z | 1,460 | 4 |
diffusers
|
[
"diffusers",
"safetensors",
"stable-diffusion",
"stable-diffusion-diffusers",
"text-to-image",
"license:other",
"autotrain_compatible",
"endpoints_compatible",
"diffusers:StableDiffusionPipeline",
"region:us"
] |
text-to-image
| 2023-12-19T18:02:11Z |
---
license: other
tags:
- stable-diffusion
- stable-diffusion-diffusers
- text-to-image
- diffusers
inference: true
---
Model info:
https://civitai.com/models/237256/yutamix-realistic
|
manar21/model
|
manar21
| 2023-12-19T18:20:59Z | 0 | 0 |
keras
|
[
"keras",
"image-segmentation",
"en",
"license:apache-2.0",
"region:us"
] |
image-segmentation
| 2023-11-19T21:34:20Z |
---
pipeline_tag: image-segmentation
license: apache-2.0
language:
- en
metrics:
- accuracy
library_name: keras
---
|
maraoz/mistral_instruct_generation
|
maraoz
| 2023-12-19T18:06:05Z | 4 | 0 |
peft
|
[
"peft",
"tensorboard",
"safetensors",
"generated_from_trainer",
"base_model:mistralai/Mistral-7B-Instruct-v0.1",
"base_model:adapter:mistralai/Mistral-7B-Instruct-v0.1",
"license:apache-2.0",
"region:us"
] | null | 2023-12-19T18:02:01Z |
---
license: apache-2.0
library_name: peft
tags:
- generated_from_trainer
base_model: mistralai/Mistral-7B-Instruct-v0.1
model-index:
- name: mistral_instruct_generation
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# mistral_instruct_generation
This model is a fine-tuned version of [mistralai/Mistral-7B-Instruct-v0.1](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.1) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 1.4259
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0002
- train_batch_size: 1
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: constant
- lr_scheduler_warmup_steps: 0.03
- training_steps: 100
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 1.7768 | 0.02 | 20 | 1.5506 |
| 1.5974 | 0.04 | 40 | 1.4599 |
| 1.5168 | 0.06 | 60 | 1.4403 |
| 1.5212 | 0.08 | 80 | 1.4321 |
| 1.3018 | 0.1 | 100 | 1.4259 |
### Framework versions
- PEFT 0.7.1
- Transformers 4.36.2
- Pytorch 2.1.2+cu121
- Datasets 2.15.0
- Tokenizers 0.15.0
|
hkivancoral/smids_10x_deit_small_sgd_0001_fold5
|
hkivancoral
| 2023-12-19T18:02:16Z | 5 | 0 |
transformers
|
[
"transformers",
"pytorch",
"vit",
"image-classification",
"generated_from_trainer",
"dataset:imagefolder",
"base_model:facebook/deit-small-patch16-224",
"base_model:finetune:facebook/deit-small-patch16-224",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
image-classification
| 2023-12-19T17:06:54Z |
---
license: apache-2.0
base_model: facebook/deit-small-patch16-224
tags:
- generated_from_trainer
datasets:
- imagefolder
metrics:
- accuracy
model-index:
- name: smids_10x_deit_small_sgd_0001_fold5
results:
- task:
name: Image Classification
type: image-classification
dataset:
name: imagefolder
type: imagefolder
config: default
split: test
args: default
metrics:
- name: Accuracy
type: accuracy
value: 0.835
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# smids_10x_deit_small_sgd_0001_fold5
This model is a fine-tuned version of [facebook/deit-small-patch16-224](https://huggingface.co/facebook/deit-small-patch16-224) on the imagefolder dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4025
- Accuracy: 0.835
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0001
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 50
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|
| 0.999 | 1.0 | 750 | 1.0177 | 0.4867 |
| 0.9125 | 2.0 | 1500 | 0.9538 | 0.56 |
| 0.8354 | 3.0 | 2250 | 0.8848 | 0.64 |
| 0.7909 | 4.0 | 3000 | 0.8172 | 0.685 |
| 0.7315 | 5.0 | 3750 | 0.7535 | 0.7183 |
| 0.6641 | 6.0 | 4500 | 0.7023 | 0.7433 |
| 0.61 | 7.0 | 5250 | 0.6582 | 0.755 |
| 0.5883 | 8.0 | 6000 | 0.6232 | 0.7783 |
| 0.6057 | 9.0 | 6750 | 0.5936 | 0.79 |
| 0.5434 | 10.0 | 7500 | 0.5693 | 0.795 |
| 0.5298 | 11.0 | 8250 | 0.5500 | 0.7917 |
| 0.4881 | 12.0 | 9000 | 0.5324 | 0.8 |
| 0.5014 | 13.0 | 9750 | 0.5180 | 0.8 |
| 0.4862 | 14.0 | 10500 | 0.5060 | 0.8083 |
| 0.4712 | 15.0 | 11250 | 0.4949 | 0.81 |
| 0.4371 | 16.0 | 12000 | 0.4864 | 0.8117 |
| 0.4626 | 17.0 | 12750 | 0.4789 | 0.815 |
| 0.4294 | 18.0 | 13500 | 0.4706 | 0.815 |
| 0.4498 | 19.0 | 14250 | 0.4650 | 0.815 |
| 0.425 | 20.0 | 15000 | 0.4594 | 0.815 |
| 0.4212 | 21.0 | 15750 | 0.4532 | 0.8167 |
| 0.4517 | 22.0 | 16500 | 0.4489 | 0.82 |
| 0.4104 | 23.0 | 17250 | 0.4443 | 0.8167 |
| 0.4051 | 24.0 | 18000 | 0.4407 | 0.82 |
| 0.4019 | 25.0 | 18750 | 0.4371 | 0.8217 |
| 0.3884 | 26.0 | 19500 | 0.4338 | 0.825 |
| 0.3154 | 27.0 | 20250 | 0.4302 | 0.825 |
| 0.3994 | 28.0 | 21000 | 0.4273 | 0.8283 |
| 0.4061 | 29.0 | 21750 | 0.4246 | 0.83 |
| 0.4059 | 30.0 | 22500 | 0.4225 | 0.8283 |
| 0.3637 | 31.0 | 23250 | 0.4202 | 0.8267 |
| 0.3501 | 32.0 | 24000 | 0.4181 | 0.8283 |
| 0.4209 | 33.0 | 24750 | 0.4163 | 0.8317 |
| 0.3255 | 34.0 | 25500 | 0.4145 | 0.8317 |
| 0.3933 | 35.0 | 26250 | 0.4127 | 0.8317 |
| 0.3766 | 36.0 | 27000 | 0.4115 | 0.8317 |
| 0.3145 | 37.0 | 27750 | 0.4102 | 0.8317 |
| 0.3874 | 38.0 | 28500 | 0.4090 | 0.83 |
| 0.3898 | 39.0 | 29250 | 0.4079 | 0.83 |
| 0.365 | 40.0 | 30000 | 0.4069 | 0.8317 |
| 0.3728 | 41.0 | 30750 | 0.4059 | 0.8317 |
| 0.3865 | 42.0 | 31500 | 0.4051 | 0.8317 |
| 0.3813 | 43.0 | 32250 | 0.4045 | 0.8317 |
| 0.3607 | 44.0 | 33000 | 0.4040 | 0.8317 |
| 0.3955 | 45.0 | 33750 | 0.4034 | 0.8333 |
| 0.3317 | 46.0 | 34500 | 0.4031 | 0.835 |
| 0.4022 | 47.0 | 35250 | 0.4028 | 0.835 |
| 0.3888 | 48.0 | 36000 | 0.4026 | 0.835 |
| 0.3745 | 49.0 | 36750 | 0.4025 | 0.835 |
| 0.3 | 50.0 | 37500 | 0.4025 | 0.835 |
### Framework versions
- Transformers 4.32.1
- Pytorch 2.1.0+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2
|
martyn/sdxl-turbo-mario-merge
|
martyn
| 2023-12-19T18:01:47Z | 0 | 5 | null |
[
"dare",
"super mario merge",
"pytorch",
"sdxl",
"sdxl_turbo",
"merge",
"text-to-image",
"en",
"arxiv:2311.03099",
"license:mit",
"region:us"
] |
text-to-image
| 2023-11-29T01:13:53Z |
---
license: mit
language:
- en
pipeline_tag: text-to-image
inference: false
tags:
- dare
- super mario merge
- pytorch
- sdxl
- sdxl_turbo
- merge
---
# SDXL Mario Merged
Merged with https://github.com/martyn/safetensors-merge-supermario
Using this technique:
https://arxiv.org/pdf/2311.03099.pdf - Language Models are Super Mario: Absorbing Abilities from Homologous Models as a Free Lunch
## Parameters used:
```
python3 merge.py -p 0.13 -lambda 3.0 sdxl_base.safetensors sd_xl_turbo_1.0_fp16.safetensors sdxl_merged.safetensors
```
Both models were fp16.
|
martyn/sdxl-dpo-turbo-dare-v0
|
martyn
| 2023-12-19T18:01:04Z | 0 | 0 | null |
[
"dare",
"super mario merge",
"pytorch",
"sdxl",
"sdxl_dpo",
"sdxl_turbo",
"merge",
"text-to-image",
"en",
"license:mit",
"region:us"
] |
text-to-image
| 2023-12-19T17:53:44Z |
---
license: mit
language:
- en
pipeline_tag: text-to-image
inference: false
tags:
- dare
- super mario merge
- pytorch
- sdxl
- sdxl_dpo
- sdxl_turbo
- merge
---
# SDXL DPO Turbo Merge
The following were merged with DARE using [https://github.com/martyn/safetensors-merge-supermario](https://github.com/martyn/safetensors-merge-supermario)
## Mergelist
```
mhdang/dpo-sdxl-text2image-v1
stabilityai/sdxl-turbo
```
## Merge command
```
python3 merge.py -p 0.13 -lambda 3.0 stable_xl_dpo.safetensors sd_xl_turbo_1.0_fp16.safetensors [output]
```
|
AsIF90/ppo-LunarLander-v2
|
AsIF90
| 2023-12-19T18:00:04Z | 0 | 0 |
stable-baselines3
|
[
"stable-baselines3",
"LunarLander-v2",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-12-19T17:59:44Z |
---
library_name: stable-baselines3
tags:
- LunarLander-v2
- deep-reinforcement-learning
- reinforcement-learning
- stable-baselines3
model-index:
- name: PPO
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: LunarLander-v2
type: LunarLander-v2
metrics:
- type: mean_reward
value: 234.88 +/- 26.81
name: mean_reward
verified: false
---
# **PPO** Agent playing **LunarLander-v2**
This is a trained model of a **PPO** agent playing **LunarLander-v2**
using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3).
## Usage (with Stable-baselines3)
TODO: Add your code
```python
from stable_baselines3 import ...
from huggingface_sb3 import load_from_hub
...
```
|
extraltodeus/Bise_7B_m37_SSRD
|
extraltodeus
| 2023-12-19T17:43:45Z | 19 | 0 |
transformers
|
[
"transformers",
"safetensors",
"mistral",
"text-generation",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"4-bit",
"gptq",
"region:us"
] |
text-generation
| 2023-12-19T16:57:05Z |
A merge of the following models:
TheBloke_Mistral-7B-Claude-Chat-GPTQ
TheBloke_airoboros-mistral2.2-7B-GPTQ
TheBloke_ANIMA-Phi-Neptune-Mistral-7B-GPTQ
TheBloke_Arithmo-Mistral-7B-GPTQ
TheBloke_AshhLimaRP-Mistral-7B-GPTQ
TheBloke_Astrid-Mistral-7B-GPTQ
TheBloke_Autolycus-Mistral_7B-GPTQ
TheBloke_Barcenas-Mistral-7B-GPTQ
TheBloke_blossom-v3-mistral-7B-GPTQ
TheBloke_CollectiveCognition-v1.1-Mistral-7B-GPTQ
TheBloke_dolphin-2.2.1-mistral-7B-GPTQ
TheBloke_Free_Sydney_V2_Mistral_7b-GPTQ
TheBloke_Generate_Question_Mistral_7B-GPTQ
TheBloke_Hermes-Trismegistus-Mistral-7B-GPTQ
TheBloke_Karen_TheEditor_V2_CREATIVE_Mistral_7B-GPTQ
TheBloke_Kimiko-Mistral-7B-GPTQ
TheBloke_Leo-Mistral-Hessianai-7B-Chat-GPTQ
TheBloke_MetaMath-Mistral-7B-GPTQ
TheBloke_Mistral-7B-AEZAKMI-v1-GPTQ
TheBloke_mistral-7B-dpo-v5-GPTQ
TheBloke_Mistral-7B-OpenOrca-GPTQ
TheBloke_Mistral-ClaudeLimaRP-v3-7B-GPTQ
TheBloke_Mistral-Trismegistus-7B-GPTQ
TheBloke_MistralLite-7B-GPTQ
TheBloke_mistral_7b_norobots-GPTQ
TheBloke_NeuralHermes-2.5-Mistral-7B-GPTQ
TheBloke_openbuddy-mistral-7B-v13.1-GPTQ
TheBloke_OpenHermes-2.5-Mistral-7B-GPTQ
TheBloke_openinstruct-mistral-7B-GPTQ
TheBloke_PiVoT-10.7B-Mistral-v0.2-RP-GPTQ
TheBloke_saiga_mistral_7b-GPTQ
TheBloke_samantha-1.2-mistral-7B-GPTQ
TheBloke_SauerkrautLM-7B-v1-mistral-GPTQ
TheBloke_SlimOpenOrca-Mistral-7B-GPTQ
TheBloke_speechless-code-mistral-7B-v1.0-GPTQ
TheBloke_Thespis-Mistral-7B-v0.6-GPTQ
TheBloke_Writing_Partner_Mistral_7B-GPTQ
The method used was to select each value that had the smallest sum of relative absolute difference.
The config files are copies from the TheBloke_Mistral-7B-Claude-Chat-GPTQ repository.
|
TheBloke/GEITje-7B-chat-GGUF
|
TheBloke
| 2023-12-19T17:41:55Z | 142 | 3 |
transformers
|
[
"transformers",
"gguf",
"mistral",
"generated_from_trainer",
"GEITje",
"conversational",
"nl",
"dataset:Rijgersberg/no_robots_nl",
"dataset:Rijgersberg/ultrachat_10k_nl",
"base_model:Rijgersberg/GEITje-7B-chat",
"base_model:quantized:Rijgersberg/GEITje-7B-chat",
"license:apache-2.0",
"region:us"
] |
text-generation
| 2023-12-19T17:37:33Z |
---
base_model: Rijgersberg/GEITje-7B-chat
datasets:
- Rijgersberg/no_robots_nl
- Rijgersberg/ultrachat_10k_nl
inference: false
language:
- nl
license: apache-2.0
model-index:
- name: GEITje-7B-chat
results: []
model_creator: Edwin Rijgersberg
model_name: Geitje 7B Chat
model_type: mistral
pipeline_tag: conversational
prompt_template: '<|user|>
{prompt}
<|assistant|>
'
quantized_by: TheBloke
tags:
- generated_from_trainer
- GEITje
---
<!-- markdownlint-disable MD041 -->
<!-- header start -->
<!-- 200823 -->
<div style="width: auto; margin-left: auto; margin-right: auto">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<div style="text-align:center; margin-top: 0em; margin-bottom: 0em"><p style="margin-top: 0.25em; margin-bottom: 0em;">TheBloke's LLM work is generously supported by a grant from <a href="https://a16z.com">andreessen horowitz (a16z)</a></p></div>
<hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
<!-- header end -->
# Geitje 7B Chat - GGUF
- Model creator: [Edwin Rijgersberg](https://huggingface.co/Rijgersberg)
- Original model: [Geitje 7B Chat](https://huggingface.co/Rijgersberg/GEITje-7B-chat)
<!-- description start -->
## Description
This repo contains GGUF format model files for [Edwin Rijgersberg's Geitje 7B Chat](https://huggingface.co/Rijgersberg/GEITje-7B-chat).
These files were quantised using hardware kindly provided by [Massed Compute](https://massedcompute.com/).
<!-- description end -->
<!-- README_GGUF.md-about-gguf start -->
### About GGUF
GGUF is a new format introduced by the llama.cpp team on August 21st 2023. It is a replacement for GGML, which is no longer supported by llama.cpp.
Here is an incomplete list of clients and libraries that are known to support GGUF:
* [llama.cpp](https://github.com/ggerganov/llama.cpp). The source project for GGUF. Offers a CLI and a server option.
* [text-generation-webui](https://github.com/oobabooga/text-generation-webui), the most widely used web UI, with many features and powerful extensions. Supports GPU acceleration.
* [KoboldCpp](https://github.com/LostRuins/koboldcpp), a fully featured web UI, with GPU accel across all platforms and GPU architectures. Especially good for story telling.
* [GPT4All](https://gpt4all.io/index.html), a free and open source local running GUI, supporting Windows, Linux and macOS with full GPU accel.
* [LM Studio](https://lmstudio.ai/), an easy-to-use and powerful local GUI for Windows and macOS (Silicon), with GPU acceleration. Linux available, in beta as of 27/11/2023.
* [LoLLMS Web UI](https://github.com/ParisNeo/lollms-webui), a great web UI with many interesting and unique features, including a full model library for easy model selection.
* [Faraday.dev](https://faraday.dev/), an attractive and easy to use character-based chat GUI for Windows and macOS (both Silicon and Intel), with GPU acceleration.
* [llama-cpp-python](https://github.com/abetlen/llama-cpp-python), a Python library with GPU accel, LangChain support, and OpenAI-compatible API server.
* [candle](https://github.com/huggingface/candle), a Rust ML framework with a focus on performance, including GPU support, and ease of use.
* [ctransformers](https://github.com/marella/ctransformers), a Python library with GPU accel, LangChain support, and OpenAI-compatible AI server. Note, as of time of writing (November 27th 2023), ctransformers has not been updated in a long time and does not support many recent models.
<!-- README_GGUF.md-about-gguf end -->
<!-- repositories-available start -->
## Repositories available
* [AWQ model(s) for GPU inference.](https://huggingface.co/TheBloke/GEITje-7B-chat-AWQ)
* [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/GEITje-7B-chat-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/GEITje-7B-chat-GGUF)
* [Edwin Rijgersberg's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/Rijgersberg/GEITje-7B-chat)
<!-- repositories-available end -->
<!-- prompt-template start -->
## Prompt template: ToRA
```
<|user|>
{prompt}
<|assistant|>
```
<!-- prompt-template end -->
<!-- compatibility_gguf start -->
## Compatibility
These quantised GGUFv2 files are compatible with llama.cpp from August 27th onwards, as of commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221)
They are also compatible with many third party UIs and libraries - please see the list at the top of this README.
## Explanation of quantisation methods
<details>
<summary>Click to see details</summary>
The new methods available are:
* GGML_TYPE_Q2_K - "type-1" 2-bit quantization in super-blocks containing 16 blocks, each block having 16 weight. Block scales and mins are quantized with 4 bits. This ends up effectively using 2.5625 bits per weight (bpw)
* GGML_TYPE_Q3_K - "type-0" 3-bit quantization in super-blocks containing 16 blocks, each block having 16 weights. Scales are quantized with 6 bits. This end up using 3.4375 bpw.
* GGML_TYPE_Q4_K - "type-1" 4-bit quantization in super-blocks containing 8 blocks, each block having 32 weights. Scales and mins are quantized with 6 bits. This ends up using 4.5 bpw.
* GGML_TYPE_Q5_K - "type-1" 5-bit quantization. Same super-block structure as GGML_TYPE_Q4_K resulting in 5.5 bpw
* GGML_TYPE_Q6_K - "type-0" 6-bit quantization. Super-blocks with 16 blocks, each block having 16 weights. Scales are quantized with 8 bits. This ends up using 6.5625 bpw
Refer to the Provided Files table below to see what files use which methods, and how.
</details>
<!-- compatibility_gguf end -->
<!-- README_GGUF.md-provided-files start -->
## Provided files
| Name | Quant method | Bits | Size | Max RAM required | Use case |
| ---- | ---- | ---- | ---- | ---- | ----- |
| [geitje-7b-chat.Q2_K.gguf](https://huggingface.co/TheBloke/GEITje-7B-chat-GGUF/blob/main/geitje-7b-chat.Q2_K.gguf) | Q2_K | 2 | 3.08 GB| 5.58 GB | smallest, significant quality loss - not recommended for most purposes |
| [geitje-7b-chat.Q3_K_S.gguf](https://huggingface.co/TheBloke/GEITje-7B-chat-GGUF/blob/main/geitje-7b-chat.Q3_K_S.gguf) | Q3_K_S | 3 | 3.17 GB| 5.67 GB | very small, high quality loss |
| [geitje-7b-chat.Q3_K_M.gguf](https://huggingface.co/TheBloke/GEITje-7B-chat-GGUF/blob/main/geitje-7b-chat.Q3_K_M.gguf) | Q3_K_M | 3 | 3.52 GB| 6.02 GB | very small, high quality loss |
| [geitje-7b-chat.Q3_K_L.gguf](https://huggingface.co/TheBloke/GEITje-7B-chat-GGUF/blob/main/geitje-7b-chat.Q3_K_L.gguf) | Q3_K_L | 3 | 3.82 GB| 6.32 GB | small, substantial quality loss |
| [geitje-7b-chat.Q4_0.gguf](https://huggingface.co/TheBloke/GEITje-7B-chat-GGUF/blob/main/geitje-7b-chat.Q4_0.gguf) | Q4_0 | 4 | 4.11 GB| 6.61 GB | legacy; small, very high quality loss - prefer using Q3_K_M |
| [geitje-7b-chat.Q4_K_S.gguf](https://huggingface.co/TheBloke/GEITje-7B-chat-GGUF/blob/main/geitje-7b-chat.Q4_K_S.gguf) | Q4_K_S | 4 | 4.14 GB| 6.64 GB | small, greater quality loss |
| [geitje-7b-chat.Q4_K_M.gguf](https://huggingface.co/TheBloke/GEITje-7B-chat-GGUF/blob/main/geitje-7b-chat.Q4_K_M.gguf) | Q4_K_M | 4 | 4.37 GB| 6.87 GB | medium, balanced quality - recommended |
| [geitje-7b-chat.Q5_0.gguf](https://huggingface.co/TheBloke/GEITje-7B-chat-GGUF/blob/main/geitje-7b-chat.Q5_0.gguf) | Q5_0 | 5 | 5.00 GB| 7.50 GB | legacy; medium, balanced quality - prefer using Q4_K_M |
| [geitje-7b-chat.Q5_K_S.gguf](https://huggingface.co/TheBloke/GEITje-7B-chat-GGUF/blob/main/geitje-7b-chat.Q5_K_S.gguf) | Q5_K_S | 5 | 5.00 GB| 7.50 GB | large, low quality loss - recommended |
| [geitje-7b-chat.Q5_K_M.gguf](https://huggingface.co/TheBloke/GEITje-7B-chat-GGUF/blob/main/geitje-7b-chat.Q5_K_M.gguf) | Q5_K_M | 5 | 5.13 GB| 7.63 GB | large, very low quality loss - recommended |
| [geitje-7b-chat.Q6_K.gguf](https://huggingface.co/TheBloke/GEITje-7B-chat-GGUF/blob/main/geitje-7b-chat.Q6_K.gguf) | Q6_K | 6 | 5.94 GB| 8.44 GB | very large, extremely low quality loss |
| [geitje-7b-chat.Q8_0.gguf](https://huggingface.co/TheBloke/GEITje-7B-chat-GGUF/blob/main/geitje-7b-chat.Q8_0.gguf) | Q8_0 | 8 | 7.70 GB| 10.20 GB | very large, extremely low quality loss - not recommended |
**Note**: the above RAM figures assume no GPU offloading. If layers are offloaded to the GPU, this will reduce RAM usage and use VRAM instead.
<!-- README_GGUF.md-provided-files end -->
<!-- README_GGUF.md-how-to-download start -->
## How to download GGUF files
**Note for manual downloaders:** You almost never want to clone the entire repo! Multiple different quantisation formats are provided, and most users only want to pick and download a single file.
The following clients/libraries will automatically download models for you, providing a list of available models to choose from:
* LM Studio
* LoLLMS Web UI
* Faraday.dev
### In `text-generation-webui`
Under Download Model, you can enter the model repo: TheBloke/GEITje-7B-chat-GGUF and below it, a specific filename to download, such as: geitje-7b-chat.Q4_K_M.gguf.
Then click Download.
### On the command line, including multiple files at once
I recommend using the `huggingface-hub` Python library:
```shell
pip3 install huggingface-hub
```
Then you can download any individual model file to the current directory, at high speed, with a command like this:
```shell
huggingface-cli download TheBloke/GEITje-7B-chat-GGUF geitje-7b-chat.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False
```
<details>
<summary>More advanced huggingface-cli download usage (click to read)</summary>
You can also download multiple files at once with a pattern:
```shell
huggingface-cli download TheBloke/GEITje-7B-chat-GGUF --local-dir . --local-dir-use-symlinks False --include='*Q4_K*gguf'
```
For more documentation on downloading with `huggingface-cli`, please see: [HF -> Hub Python Library -> Download files -> Download from the CLI](https://huggingface.co/docs/huggingface_hub/guides/download#download-from-the-cli).
To accelerate downloads on fast connections (1Gbit/s or higher), install `hf_transfer`:
```shell
pip3 install hf_transfer
```
And set environment variable `HF_HUB_ENABLE_HF_TRANSFER` to `1`:
```shell
HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download TheBloke/GEITje-7B-chat-GGUF geitje-7b-chat.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False
```
Windows Command Line users: You can set the environment variable by running `set HF_HUB_ENABLE_HF_TRANSFER=1` before the download command.
</details>
<!-- README_GGUF.md-how-to-download end -->
<!-- README_GGUF.md-how-to-run start -->
## Example `llama.cpp` command
Make sure you are using `llama.cpp` from commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221) or later.
```shell
./main -ngl 35 -m geitje-7b-chat.Q4_K_M.gguf --color -c 32768 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "<|user|>\n{prompt}\n<|assistant|>"
```
Change `-ngl 32` to the number of layers to offload to GPU. Remove it if you don't have GPU acceleration.
Change `-c 32768` to the desired sequence length. For extended sequence models - eg 8K, 16K, 32K - the necessary RoPE scaling parameters are read from the GGUF file and set by llama.cpp automatically. Note that longer sequence lengths require much more resources, so you may need to reduce this value.
If you want to have a chat-style conversation, replace the `-p <PROMPT>` argument with `-i -ins`
For other parameters and how to use them, please refer to [the llama.cpp documentation](https://github.com/ggerganov/llama.cpp/blob/master/examples/main/README.md)
## How to run in `text-generation-webui`
Further instructions can be found in the text-generation-webui documentation, here: [text-generation-webui/docs/04 ‐ Model Tab.md](https://github.com/oobabooga/text-generation-webui/blob/main/docs/04%20%E2%80%90%20Model%20Tab.md#llamacpp).
## How to run from Python code
You can use GGUF models from Python using the [llama-cpp-python](https://github.com/abetlen/llama-cpp-python) or [ctransformers](https://github.com/marella/ctransformers) libraries. Note that at the time of writing (Nov 27th 2023), ctransformers has not been updated for some time and is not compatible with some recent models. Therefore I recommend you use llama-cpp-python.
### How to load this model in Python code, using llama-cpp-python
For full documentation, please see: [llama-cpp-python docs](https://abetlen.github.io/llama-cpp-python/).
#### First install the package
Run one of the following commands, according to your system:
```shell
# Base ctransformers with no GPU acceleration
pip install llama-cpp-python
# With NVidia CUDA acceleration
CMAKE_ARGS="-DLLAMA_CUBLAS=on" pip install llama-cpp-python
# Or with OpenBLAS acceleration
CMAKE_ARGS="-DLLAMA_BLAS=ON -DLLAMA_BLAS_VENDOR=OpenBLAS" pip install llama-cpp-python
# Or with CLBLast acceleration
CMAKE_ARGS="-DLLAMA_CLBLAST=on" pip install llama-cpp-python
# Or with AMD ROCm GPU acceleration (Linux only)
CMAKE_ARGS="-DLLAMA_HIPBLAS=on" pip install llama-cpp-python
# Or with Metal GPU acceleration for macOS systems only
CMAKE_ARGS="-DLLAMA_METAL=on" pip install llama-cpp-python
# In windows, to set the variables CMAKE_ARGS in PowerShell, follow this format; eg for NVidia CUDA:
$env:CMAKE_ARGS = "-DLLAMA_OPENBLAS=on"
pip install llama-cpp-python
```
#### Simple llama-cpp-python example code
```python
from llama_cpp import Llama
# Set gpu_layers to the number of layers to offload to GPU. Set to 0 if no GPU acceleration is available on your system.
llm = Llama(
model_path="./geitje-7b-chat.Q4_K_M.gguf", # Download the model file first
n_ctx=32768, # The max sequence length to use - note that longer sequence lengths require much more resources
n_threads=8, # The number of CPU threads to use, tailor to your system and the resulting performance
n_gpu_layers=35 # The number of layers to offload to GPU, if you have GPU acceleration available
)
# Simple inference example
output = llm(
"<|user|>\n{prompt}\n<|assistant|>", # Prompt
max_tokens=512, # Generate up to 512 tokens
stop=["</s>"], # Example stop token - not necessarily correct for this specific model! Please check before using.
echo=True # Whether to echo the prompt
)
# Chat Completion API
llm = Llama(model_path="./geitje-7b-chat.Q4_K_M.gguf", chat_format="llama-2") # Set chat_format according to the model you are using
llm.create_chat_completion(
messages = [
{"role": "system", "content": "You are a story writing assistant."},
{
"role": "user",
"content": "Write a story about llamas."
}
]
)
```
## How to use with LangChain
Here are guides on using llama-cpp-python and ctransformers with LangChain:
* [LangChain + llama-cpp-python](https://python.langchain.com/docs/integrations/llms/llamacpp)
* [LangChain + ctransformers](https://python.langchain.com/docs/integrations/providers/ctransformers)
<!-- README_GGUF.md-how-to-run end -->
<!-- footer start -->
<!-- 200823 -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute
Thanks to the [chirper.ai](https://chirper.ai) team!
Thanks to Clay from [gpus.llm-utils.org](llm-utils)!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Aemon Algiz.
**Patreon special mentions**: Michael Levine, 阿明, Trailburnt, Nikolai Manek, John Detwiler, Randy H, Will Dee, Sebastain Graf, NimbleBox.ai, Eugene Pentland, Emad Mostaque, Ai Maven, Jim Angel, Jeff Scroggin, Michael Davis, Manuel Alberto Morcote, Stephen Murray, Robert, Justin Joy, Luke @flexchar, Brandon Frisco, Elijah Stavena, S_X, Dan Guido, Undi ., Komninos Chatzipapas, Shadi, theTransient, Lone Striker, Raven Klaugh, jjj, Cap'n Zoog, Michel-Marie MAUDET (LINAGORA), Matthew Berman, David, Fen Risland, Omer Bin Jawed, Luke Pendergrass, Kalila, OG, Erik Bjäreholt, Rooh Singh, Joseph William Delisle, Dan Lewis, TL, John Villwock, AzureBlack, Brad, Pedro Madruga, Caitlyn Gatomon, K, jinyuan sun, Mano Prime, Alex, Jeffrey Morgan, Alicia Loh, Illia Dulskyi, Chadd, transmissions 11, fincy, Rainer Wilmers, ReadyPlayerEmma, knownsqashed, Mandus, biorpg, Deo Leter, Brandon Phillips, SuperWojo, Sean Connelly, Iucharbius, Jack West, Harry Royden McLaughlin, Nicholas, terasurfer, Vitor Caleffi, Duane Dunston, Johann-Peter Hartmann, David Ziegler, Olakabola, Ken Nordquist, Trenton Dambrowitz, Tom X Nguyen, Vadim, Ajan Kanaga, Leonard Tan, Clay Pascal, Alexandros Triantafyllidis, JM33133, Xule, vamX, ya boyyy, subjectnull, Talal Aujan, Alps Aficionado, wassieverse, Ari Malik, James Bentley, Woland, Spencer Kim, Michael Dempsey, Fred von Graf, Elle, zynix, William Richards, Stanislav Ovsiannikov, Edmond Seymore, Jonathan Leane, Martin Kemka, usrbinkat, Enrico Ros
Thank you to all my generous patrons and donaters!
And thank you again to a16z for their generous grant.
<!-- footer end -->
<!-- original-model-card start -->
# Original model card: Edwin Rijgersberg's Geitje 7B Chat
# GEITje-7B-chat
# GEITje-7B
GEITje is a large open Dutch language model with 7 billion parameters, based on Mistral 7B.
It has been further trained on 10 billion tokens of Dutch text.
This has improved its Dutch language skills and increased its knowledge of Dutch topics.
## Model description
### _Mistral_ – Base Model
GEITje is based on [Mistral 7B](https://mistral.ai/news/announcing-mistral-7b/).
It's a large open language model with 7 billion parameters,
trained by [Mistral AI](https://mistral.ai).
According to Mistral AI, the 7B model performs better than [Llama 2](https://ai.meta.com/llama/) 13B on all (English-language) benchmarks they tested it on.
Mistral 7B has been released under the Apache 2.0 open source license.
### _GEITje_ – Trained Further on Dutch Texts
GEITje was created by further training Mistral 7B on no less than 10 billion tokens of Dutch text from the [Dutch Gigacorpus](http://gigacorpus.nl) and the [MADLAD-400](https://huggingface.co/datasets/allenai/MADLAD-400) web crawling corpus.
It is a so-called _full-parameter finetune_:
performed on all parameters.
It is not a [PEFT](https://huggingface.co/blog/peft) or [LoRA](https://huggingface.co/docs/peft/conceptual_guides/lora) finetune.
Like Mistral, GEITje has a _context length_ of 8,192 tokens.
### _GEITje-chat_ – Finetuned for Dialogues
As a demonstration of GEITje's capabilities for chat applications, two initial chat variants of GEITje have also been finetuned: GEITje-chat and GEITje-chat-v2.
They can follow instructions, answer questions, and hold dialogues on a variety of topics.
## More info
Read more about GEITje-chat in the [📄 README](https://github.com/Rijgersberg/GEITje/blob/main/README-en.md) on GitHub.
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 1e-05
- train_batch_size: 2
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 8
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 3
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 1.0263 | 0.2 | 236 | 0.9482 |
| 1.0368 | 0.4 | 472 | 0.9574 |
| 0.9503 | 0.6 | 708 | 0.9492 |
| 1.1419 | 0.8 | 944 | 0.9406 |
| 1.2161 | 1.0 | 1180 | 0.9317 |
| 0.6695 | 1.2 | 1416 | 0.9407 |
| 0.7379 | 1.4 | 1652 | 0.9350 |
| 0.7695 | 1.6 | 1888 | 0.9282 |
| 0.6795 | 1.8 | 2124 | 0.9218 |
| 0.6217 | 2.0 | 2360 | 0.9174 |
| 0.438 | 2.2 | 2596 | 0.9546 |
| 0.3719 | 2.39 | 2832 | 0.9546 |
| 0.4853 | 2.59 | 3068 | 0.9548 |
| 0.3852 | 2.79 | 3304 | 0.9548 |
| 0.48 | 2.99 | 3540 | 0.9548 |
### Framework versions
- Transformers 4.36.0.dev0
- Pytorch 2.1.1+cu121
- Datasets 2.15.0
- Tokenizers 0.15.0
<!-- original-model-card end -->
|
kajol/mistral_math
|
kajol
| 2023-12-19T17:34:55Z | 0 | 0 |
peft
|
[
"peft",
"safetensors",
"arxiv:1910.09700",
"base_model:TheBloke/zephyr-7B-alpha-GPTQ",
"base_model:adapter:TheBloke/zephyr-7B-alpha-GPTQ",
"region:us"
] | null | 2023-12-19T17:34:05Z |
---
library_name: peft
base_model: TheBloke/zephyr-7B-alpha-GPTQ
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
### Framework versions
- PEFT 0.7.1
|
TheBloke/Metis-0.4-GPTQ
|
TheBloke
| 2023-12-19T17:34:05Z | 17 | 2 |
transformers
|
[
"transformers",
"safetensors",
"mistral",
"text-generation",
"conversational",
"base_model:Mihaiii/Metis-0.4",
"base_model:quantized:Mihaiii/Metis-0.4",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"4-bit",
"gptq",
"region:us"
] |
text-generation
| 2023-12-19T17:05:42Z |
---
base_model: Mihaiii/Metis-0.4
inference: false
license: apache-2.0
license_name: apache-2.0
metrics:
- accuracy
model_creator: Mihai
model_name: Metis 0.4
model_type: mistral
prompt_template: '<|system|>
{system_message}</s>
<|user|>
{prompt}</s>
<|assistant|>
'
quantized_by: TheBloke
---
<!-- markdownlint-disable MD041 -->
<!-- header start -->
<!-- 200823 -->
<div style="width: auto; margin-left: auto; margin-right: auto">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<div style="text-align:center; margin-top: 0em; margin-bottom: 0em"><p style="margin-top: 0.25em; margin-bottom: 0em;">TheBloke's LLM work is generously supported by a grant from <a href="https://a16z.com">andreessen horowitz (a16z)</a></p></div>
<hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
<!-- header end -->
# Metis 0.4 - GPTQ
- Model creator: [Mihai](https://huggingface.co/Mihaiii)
- Original model: [Metis 0.4](https://huggingface.co/Mihaiii/Metis-0.4)
<!-- description start -->
# Description
This repo contains GPTQ model files for [Mihai's Metis 0.4](https://huggingface.co/Mihaiii/Metis-0.4).
Multiple GPTQ parameter permutations are provided; see Provided Files below for details of the options provided, their parameters, and the software used to create them.
These files were quantised using hardware kindly provided by [Massed Compute](https://massedcompute.com/).
<!-- description end -->
<!-- repositories-available start -->
## Repositories available
* [AWQ model(s) for GPU inference.](https://huggingface.co/TheBloke/Metis-0.4-AWQ)
* [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/Metis-0.4-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/Metis-0.4-GGUF)
* [Mihai's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/Mihaiii/Metis-0.4)
<!-- repositories-available end -->
<!-- prompt-template start -->
## Prompt template: Zephyr
```
<|system|>
{system_message}</s>
<|user|>
{prompt}</s>
<|assistant|>
```
<!-- prompt-template end -->
<!-- README_GPTQ.md-compatible clients start -->
## Known compatible clients / servers
GPTQ models are currently supported on Linux (NVidia/AMD) and Windows (NVidia only). macOS users: please use GGUF models.
These GPTQ models are known to work in the following inference servers/webuis.
- [text-generation-webui](https://github.com/oobabooga/text-generation-webui)
- [KoboldAI United](https://github.com/henk717/koboldai)
- [LoLLMS Web UI](https://github.com/ParisNeo/lollms-webui)
- [Hugging Face Text Generation Inference (TGI)](https://github.com/huggingface/text-generation-inference)
This may not be a complete list; if you know of others, please let me know!
<!-- README_GPTQ.md-compatible clients end -->
<!-- README_GPTQ.md-provided-files start -->
## Provided files, and GPTQ parameters
Multiple quantisation parameters are provided, to allow you to choose the best one for your hardware and requirements.
Each separate quant is in a different branch. See below for instructions on fetching from different branches.
Most GPTQ files are made with AutoGPTQ. Mistral models are currently made with Transformers.
<details>
<summary>Explanation of GPTQ parameters</summary>
- Bits: The bit size of the quantised model.
- GS: GPTQ group size. Higher numbers use less VRAM, but have lower quantisation accuracy. "None" is the lowest possible value.
- Act Order: True or False. Also known as `desc_act`. True results in better quantisation accuracy. Some GPTQ clients have had issues with models that use Act Order plus Group Size, but this is generally resolved now.
- Damp %: A GPTQ parameter that affects how samples are processed for quantisation. 0.01 is default, but 0.1 results in slightly better accuracy.
- GPTQ dataset: The calibration dataset used during quantisation. Using a dataset more appropriate to the model's training can improve quantisation accuracy. Note that the GPTQ calibration dataset is not the same as the dataset used to train the model - please refer to the original model repo for details of the training dataset(s).
- Sequence Length: The length of the dataset sequences used for quantisation. Ideally this is the same as the model sequence length. For some very long sequence models (16+K), a lower sequence length may have to be used. Note that a lower sequence length does not limit the sequence length of the quantised model. It only impacts the quantisation accuracy on longer inference sequences.
- ExLlama Compatibility: Whether this file can be loaded with ExLlama, which currently only supports Llama and Mistral models in 4-bit.
</details>
| Branch | Bits | GS | Act Order | Damp % | GPTQ Dataset | Seq Len | Size | ExLlama | Desc |
| ------ | ---- | -- | --------- | ------ | ------------ | ------- | ---- | ------- | ---- |
| [main](https://huggingface.co/TheBloke/Metis-0.4-GPTQ/tree/main) | 4 | 128 | Yes | 0.1 | [VMware Open Instruct](https://huggingface.co/datasets/VMware/open-instruct/viewer/) | 4096 | 4.16 GB | Yes | 4-bit, with Act Order and group size 128g. Uses even less VRAM than 64g, but with slightly lower accuracy. |
| [gptq-4bit-32g-actorder_True](https://huggingface.co/TheBloke/Metis-0.4-GPTQ/tree/gptq-4bit-32g-actorder_True) | 4 | 32 | Yes | 0.1 | [VMware Open Instruct](https://huggingface.co/datasets/VMware/open-instruct/viewer/) | 4096 | 4.57 GB | Yes | 4-bit, with Act Order and group size 32g. Gives highest possible inference quality, with maximum VRAM usage. |
| [gptq-8bit--1g-actorder_True](https://huggingface.co/TheBloke/Metis-0.4-GPTQ/tree/gptq-8bit--1g-actorder_True) | 8 | None | Yes | 0.1 | [VMware Open Instruct](https://huggingface.co/datasets/VMware/open-instruct/viewer/) | 4096 | 7.52 GB | No | 8-bit, with Act Order. No group size, to lower VRAM requirements. |
| [gptq-8bit-128g-actorder_True](https://huggingface.co/TheBloke/Metis-0.4-GPTQ/tree/gptq-8bit-128g-actorder_True) | 8 | 128 | Yes | 0.1 | [VMware Open Instruct](https://huggingface.co/datasets/VMware/open-instruct/viewer/) | 4096 | 7.68 GB | No | 8-bit, with group size 128g for higher inference quality and with Act Order for even higher accuracy. |
| [gptq-8bit-32g-actorder_True](https://huggingface.co/TheBloke/Metis-0.4-GPTQ/tree/gptq-8bit-32g-actorder_True) | 8 | 32 | Yes | 0.1 | [VMware Open Instruct](https://huggingface.co/datasets/VMware/open-instruct/viewer/) | 4096 | 8.17 GB | No | 8-bit, with group size 32g and Act Order for maximum inference quality. |
| [gptq-4bit-64g-actorder_True](https://huggingface.co/TheBloke/Metis-0.4-GPTQ/tree/gptq-4bit-64g-actorder_True) | 4 | 64 | Yes | 0.1 | [VMware Open Instruct](https://huggingface.co/datasets/VMware/open-instruct/viewer/) | 4096 | 4.29 GB | Yes | 4-bit, with Act Order and group size 64g. Uses less VRAM than 32g, but with slightly lower accuracy. |
<!-- README_GPTQ.md-provided-files end -->
<!-- README_GPTQ.md-download-from-branches start -->
## How to download, including from branches
### In text-generation-webui
To download from the `main` branch, enter `TheBloke/Metis-0.4-GPTQ` in the "Download model" box.
To download from another branch, add `:branchname` to the end of the download name, eg `TheBloke/Metis-0.4-GPTQ:gptq-4bit-32g-actorder_True`
### From the command line
I recommend using the `huggingface-hub` Python library:
```shell
pip3 install huggingface-hub
```
To download the `main` branch to a folder called `Metis-0.4-GPTQ`:
```shell
mkdir Metis-0.4-GPTQ
huggingface-cli download TheBloke/Metis-0.4-GPTQ --local-dir Metis-0.4-GPTQ --local-dir-use-symlinks False
```
To download from a different branch, add the `--revision` parameter:
```shell
mkdir Metis-0.4-GPTQ
huggingface-cli download TheBloke/Metis-0.4-GPTQ --revision gptq-4bit-32g-actorder_True --local-dir Metis-0.4-GPTQ --local-dir-use-symlinks False
```
<details>
<summary>More advanced huggingface-cli download usage</summary>
If you remove the `--local-dir-use-symlinks False` parameter, the files will instead be stored in the central Hugging Face cache directory (default location on Linux is: `~/.cache/huggingface`), and symlinks will be added to the specified `--local-dir`, pointing to their real location in the cache. This allows for interrupted downloads to be resumed, and allows you to quickly clone the repo to multiple places on disk without triggering a download again. The downside, and the reason why I don't list that as the default option, is that the files are then hidden away in a cache folder and it's harder to know where your disk space is being used, and to clear it up if/when you want to remove a download model.
The cache location can be changed with the `HF_HOME` environment variable, and/or the `--cache-dir` parameter to `huggingface-cli`.
For more documentation on downloading with `huggingface-cli`, please see: [HF -> Hub Python Library -> Download files -> Download from the CLI](https://huggingface.co/docs/huggingface_hub/guides/download#download-from-the-cli).
To accelerate downloads on fast connections (1Gbit/s or higher), install `hf_transfer`:
```shell
pip3 install hf_transfer
```
And set environment variable `HF_HUB_ENABLE_HF_TRANSFER` to `1`:
```shell
mkdir Metis-0.4-GPTQ
HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download TheBloke/Metis-0.4-GPTQ --local-dir Metis-0.4-GPTQ --local-dir-use-symlinks False
```
Windows Command Line users: You can set the environment variable by running `set HF_HUB_ENABLE_HF_TRANSFER=1` before the download command.
</details>
### With `git` (**not** recommended)
To clone a specific branch with `git`, use a command like this:
```shell
git clone --single-branch --branch gptq-4bit-32g-actorder_True https://huggingface.co/TheBloke/Metis-0.4-GPTQ
```
Note that using Git with HF repos is strongly discouraged. It will be much slower than using `huggingface-hub`, and will use twice as much disk space as it has to store the model files twice (it stores every byte both in the intended target folder, and again in the `.git` folder as a blob.)
<!-- README_GPTQ.md-download-from-branches end -->
<!-- README_GPTQ.md-text-generation-webui start -->
## How to easily download and use this model in [text-generation-webui](https://github.com/oobabooga/text-generation-webui)
Please make sure you're using the latest version of [text-generation-webui](https://github.com/oobabooga/text-generation-webui).
It is strongly recommended to use the text-generation-webui one-click-installers unless you're sure you know how to make a manual install.
1. Click the **Model tab**.
2. Under **Download custom model or LoRA**, enter `TheBloke/Metis-0.4-GPTQ`.
- To download from a specific branch, enter for example `TheBloke/Metis-0.4-GPTQ:gptq-4bit-32g-actorder_True`
- see Provided Files above for the list of branches for each option.
3. Click **Download**.
4. The model will start downloading. Once it's finished it will say "Done".
5. In the top left, click the refresh icon next to **Model**.
6. In the **Model** dropdown, choose the model you just downloaded: `Metis-0.4-GPTQ`
7. The model will automatically load, and is now ready for use!
8. If you want any custom settings, set them and then click **Save settings for this model** followed by **Reload the Model** in the top right.
- Note that you do not need to and should not set manual GPTQ parameters any more. These are set automatically from the file `quantize_config.json`.
9. Once you're ready, click the **Text Generation** tab and enter a prompt to get started!
<!-- README_GPTQ.md-text-generation-webui end -->
<!-- README_GPTQ.md-use-from-tgi start -->
## Serving this model from Text Generation Inference (TGI)
It's recommended to use TGI version 1.1.0 or later. The official Docker container is: `ghcr.io/huggingface/text-generation-inference:1.1.0`
Example Docker parameters:
```shell
--model-id TheBloke/Metis-0.4-GPTQ --port 3000 --quantize gptq --max-input-length 3696 --max-total-tokens 4096 --max-batch-prefill-tokens 4096
```
Example Python code for interfacing with TGI (requires huggingface-hub 0.17.0 or later):
```shell
pip3 install huggingface-hub
```
```python
from huggingface_hub import InferenceClient
endpoint_url = "https://your-endpoint-url-here"
prompt = "Tell me about AI"
prompt_template=f'''<|system|>
{system_message}</s>
<|user|>
{prompt}</s>
<|assistant|>
'''
client = InferenceClient(endpoint_url)
response = client.text_generation(prompt,
max_new_tokens=128,
do_sample=True,
temperature=0.7,
top_p=0.95,
top_k=40,
repetition_penalty=1.1)
print(f"Model output: {response}")
```
<!-- README_GPTQ.md-use-from-tgi end -->
<!-- README_GPTQ.md-use-from-python start -->
## Python code example: inference from this GPTQ model
### Install the necessary packages
Requires: Transformers 4.33.0 or later, Optimum 1.12.0 or later, and AutoGPTQ 0.4.2 or later.
```shell
pip3 install --upgrade transformers optimum
# If using PyTorch 2.1 + CUDA 12.x:
pip3 install --upgrade auto-gptq
# or, if using PyTorch 2.1 + CUDA 11.x:
pip3 install --upgrade auto-gptq --extra-index-url https://huggingface.github.io/autogptq-index/whl/cu118/
```
If you are using PyTorch 2.0, you will need to install AutoGPTQ from source. Likewise if you have problems with the pre-built wheels, you should try building from source:
```shell
pip3 uninstall -y auto-gptq
git clone https://github.com/PanQiWei/AutoGPTQ
cd AutoGPTQ
git checkout v0.5.1
pip3 install .
```
### Example Python code
```python
from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
model_name_or_path = "TheBloke/Metis-0.4-GPTQ"
# To use a different branch, change revision
# For example: revision="gptq-4bit-32g-actorder_True"
model = AutoModelForCausalLM.from_pretrained(model_name_or_path,
device_map="auto",
trust_remote_code=False,
revision="main")
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
prompt = "Write a story about llamas"
system_message = "You are a story writing assistant"
prompt_template=f'''<|system|>
{system_message}</s>
<|user|>
{prompt}</s>
<|assistant|>
'''
print("\n\n*** Generate:")
input_ids = tokenizer(prompt_template, return_tensors='pt').input_ids.cuda()
output = model.generate(inputs=input_ids, temperature=0.7, do_sample=True, top_p=0.95, top_k=40, max_new_tokens=512)
print(tokenizer.decode(output[0]))
# Inference can also be done using transformers' pipeline
print("*** Pipeline:")
pipe = pipeline(
"text-generation",
model=model,
tokenizer=tokenizer,
max_new_tokens=512,
do_sample=True,
temperature=0.7,
top_p=0.95,
top_k=40,
repetition_penalty=1.1
)
print(pipe(prompt_template)[0]['generated_text'])
```
<!-- README_GPTQ.md-use-from-python end -->
<!-- README_GPTQ.md-compatibility start -->
## Compatibility
The files provided are tested to work with Transformers. For non-Mistral models, AutoGPTQ can also be used directly.
[ExLlama](https://github.com/turboderp/exllama) is compatible with Llama architecture models (including Mistral, Yi, DeepSeek, SOLAR, etc) in 4-bit. Please see the Provided Files table above for per-file compatibility.
For a list of clients/servers, please see "Known compatible clients / servers", above.
<!-- README_GPTQ.md-compatibility end -->
<!-- footer start -->
<!-- 200823 -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute
Thanks to the [chirper.ai](https://chirper.ai) team!
Thanks to Clay from [gpus.llm-utils.org](llm-utils)!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Aemon Algiz.
**Patreon special mentions**: Michael Levine, 阿明, Trailburnt, Nikolai Manek, John Detwiler, Randy H, Will Dee, Sebastain Graf, NimbleBox.ai, Eugene Pentland, Emad Mostaque, Ai Maven, Jim Angel, Jeff Scroggin, Michael Davis, Manuel Alberto Morcote, Stephen Murray, Robert, Justin Joy, Luke @flexchar, Brandon Frisco, Elijah Stavena, S_X, Dan Guido, Undi ., Komninos Chatzipapas, Shadi, theTransient, Lone Striker, Raven Klaugh, jjj, Cap'n Zoog, Michel-Marie MAUDET (LINAGORA), Matthew Berman, David, Fen Risland, Omer Bin Jawed, Luke Pendergrass, Kalila, OG, Erik Bjäreholt, Rooh Singh, Joseph William Delisle, Dan Lewis, TL, John Villwock, AzureBlack, Brad, Pedro Madruga, Caitlyn Gatomon, K, jinyuan sun, Mano Prime, Alex, Jeffrey Morgan, Alicia Loh, Illia Dulskyi, Chadd, transmissions 11, fincy, Rainer Wilmers, ReadyPlayerEmma, knownsqashed, Mandus, biorpg, Deo Leter, Brandon Phillips, SuperWojo, Sean Connelly, Iucharbius, Jack West, Harry Royden McLaughlin, Nicholas, terasurfer, Vitor Caleffi, Duane Dunston, Johann-Peter Hartmann, David Ziegler, Olakabola, Ken Nordquist, Trenton Dambrowitz, Tom X Nguyen, Vadim, Ajan Kanaga, Leonard Tan, Clay Pascal, Alexandros Triantafyllidis, JM33133, Xule, vamX, ya boyyy, subjectnull, Talal Aujan, Alps Aficionado, wassieverse, Ari Malik, James Bentley, Woland, Spencer Kim, Michael Dempsey, Fred von Graf, Elle, zynix, William Richards, Stanislav Ovsiannikov, Edmond Seymore, Jonathan Leane, Martin Kemka, usrbinkat, Enrico Ros
Thank you to all my generous patrons and donaters!
And thank you again to a16z for their generous grant.
<!-- footer end -->
# Original model card: Mihai's Metis 0.4
This is a merge between Metis-0.3 and Metis-0.1 having Metis-0.1 as base.
It was done using [mergekit](https://github.com/cg123/mergekit).
It works well with long system prompts.
It isn't generic in a sense that it shouldn't be used for story telling, for example, but only for reasoning and text comprehension.
This model is trained on a private dataset. The high GSM8K score is **NOT** because of the MetaMath dataset.
# Prompt Format:
```
<|system|>
{system_message} </s>
<|user|>
{prompt} </s>
<|assistant|>
```
Merge config:
```yaml
slices:
- sources:
- model: Mihaiii/Metis-0.3
layer_range: [0, 32]
- model: Mihaiii/Metis-0.1
layer_range: [0, 32]
merge_method: slerp
base_model: Mihaiii/Metis-0.1
parameters:
t:
- filter: self_attn
value: [0, 0.5, 0.3, 0.7, 1]
- filter: mlp
value: [1, 0.5, 0.7, 0.3, 0]
- value: 0.5 # fallback for rest of tensors
dtype: bfloat16
```
|
erfanzar/LinguaMatic-Tiny-GGUF
|
erfanzar
| 2023-12-19T17:29:02Z | 71 | 0 | null |
[
"gguf",
"code",
"text-generation",
"en",
"dataset:erfanzar/UltraChat-Mixin",
"license:mit",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2023-12-19T12:18:36Z |
---
license: mit
datasets:
- erfanzar/UltraChat-Mixin
language:
- en
pipeline_tag: text-generation
tags:
- code
---
# LinguaMatic
LinguaMatic is an advanced AI model designed to handle a wide range of Natural Language Processing (NLP) tasks. With its powerful capabilities, LinguaMatic can assist with tasks such as text classification, sentiment analysis, language translation, question answering, and much more.
## Last Update
* Nov 19 - Now AI works better with system prompting and you can achive better result with giving model better system prompts
* Nov 20 - Fixing Questioning responses
## EasyDel
The model is finetuned Using a custom version of UltraChat on TPU-v4 POD using [EasyDel](https://github.com/erfanzar/EasyDeL)
## Prompting Method
LinguaMatic utilizes the llama2 prompting method to generate responses. This method, named after the friendly and intelligent llama, enhances the model's ability to engage in meaningful conversations. The `prompt_model` function provided below demonstrates how the llama2 prompting method is implemented:
```python
def prompt_model(message: str, chat_history,
system_prompt: str) -> str:
do_strip = False
texts = [f'<s>[INST] <<SYS>>\n{system_prompt}\n<</SYS>>\n\n']
for user_input, response in chat_history:
user_input = user_input.strip() if do_strip else user_input
do_strip = True
texts.append(f'{user_input} [/INST] {response.strip()} </s><s>[INST] ')
message = message.strip() if do_strip else message
texts.append(f'{message} [/INST]')
return ''.join(texts)
```
The `prompt_model` function takes a `message` as input, along with the `chat_history` and `system_prompt`. It generates a formatted text that includes the system prompt, user inputs, and the current message. This approach allows LinguaMatic to maintain context and provide more coherent and context-aware responses.
## Contributing
We welcome contributions to enhance LinguaMatic's capabilities and improve its performance. If you encounter any issues or have suggestions for improvement, please feel free to submit a pull request or open an issue on [EasyDel](https://github.com/erfanzar/EasyDeL) GitHub repository.
|
digiplay/RunDiffusionFX2.5D_v1_diffusers
|
digiplay
| 2023-12-19T17:24:42Z | 795 | 10 |
diffusers
|
[
"diffusers",
"safetensors",
"stable-diffusion",
"stable-diffusion-diffusers",
"text-to-image",
"license:other",
"autotrain_compatible",
"endpoints_compatible",
"diffusers:StableDiffusionPipeline",
"region:us"
] |
text-to-image
| 2023-06-03T22:33:29Z |
---
license: other
tags:
- stable-diffusion
- stable-diffusion-diffusers
- text-to-image
- diffusers
inference: true
---
Model info： https://civitai.com/models/82981/rundiffusion-fx-25d
Sample images I made：
|
Subsets and Splits
Filtered Qwen2.5 Distill Models
Identifies specific configurations of models by filtering cards that contain 'distill', 'qwen2.5', '7b' while excluding certain base models and incorrect model ID patterns, uncovering unique model variants.
Filtered Model Cards Count
Finds the count of entries with specific card details that include 'distill', 'qwen2.5', '7b' but exclude certain base models, revealing valuable insights about the dataset's content distribution.
Filtered Distill Qwen 7B Models
Filters for specific card entries containing 'distill', 'qwen', and '7b', excluding certain strings and patterns, to identify relevant model configurations.
Filtered Qwen-7b Model Cards
The query performs a detailed filtering based on specific keywords and excludes certain entries, which could be useful for identifying a specific subset of cards but does not provide deeper insights or trends.
Filtered Qwen 7B Model Cards
The query filters for specific terms related to "distilled" or "distill", "qwen", and "7b" in the 'card' column but excludes certain base models, providing a limited set of entries for further inspection.
Qwen 7B Distilled Models
The query provides a basic filtering of records to find specific card names that include keywords related to distilled Qwen 7b models, excluding a particular base model, which gives limited insight but helps in focusing on relevant entries.
Qwen 7B Distilled Model Cards
The query filters data based on specific keywords in the modelId and card fields, providing limited insight primarily useful for locating specific entries rather than revealing broad patterns or trends.
Qwen 7B Distilled Models
Finds all entries containing the terms 'distilled', 'qwen', and '7b' in a case-insensitive manner, providing a filtered set of records but without deeper analysis.
Distilled Qwen 7B Models
The query filters for specific model IDs containing 'distilled', 'qwen', and '7b', providing a basic retrieval of relevant entries but without deeper analysis or insight.
Filtered Model Cards with Distill Qwen2.
Filters and retrieves records containing specific keywords in the card description while excluding certain phrases, providing a basic count of relevant entries.
Filtered Model Cards with Distill Qwen 7
The query filters specific variations of card descriptions containing 'distill', 'qwen', and '7b' while excluding a particular base model, providing limited but specific data retrieval.
Distill Qwen 7B Model Cards
The query filters and retrieves rows where the 'card' column contains specific keywords ('distill', 'qwen', and '7b'), providing a basic filter result that can help in identifying specific entries.