modelId
stringlengths 5
139
| author
stringlengths 2
42
| last_modified
timestamp[us, tz=UTC]date 2020-02-15 11:33:14
2025-09-10 18:30:15
| downloads
int64 0
223M
| likes
int64 0
11.7k
| library_name
stringclasses 553
values | tags
listlengths 1
4.05k
| pipeline_tag
stringclasses 55
values | createdAt
timestamp[us, tz=UTC]date 2022-03-02 23:29:04
2025-09-10 18:29:50
| card
stringlengths 11
1.01M
|
|---|---|---|---|---|---|---|---|---|---|
TalesLF/a2c-PandaReachDense-v2
|
TalesLF
| 2023-07-09T19:40:12Z | 0 | 0 |
stable-baselines3
|
[
"stable-baselines3",
"PandaReachDense-v2",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-07-09T19:36:52Z |
---
library_name: stable-baselines3
tags:
- PandaReachDense-v2
- deep-reinforcement-learning
- reinforcement-learning
- stable-baselines3
model-index:
- name: A2C
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: PandaReachDense-v2
type: PandaReachDense-v2
metrics:
- type: mean_reward
value: -0.66 +/- 0.21
name: mean_reward
verified: false
---
# **A2C** Agent playing **PandaReachDense-v2**
This is a trained model of a **A2C** agent playing **PandaReachDense-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
...
```
|
hopkins/eng-ind-nng
|
hopkins
| 2023-07-09T19:35:41Z | 102 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"mbart",
"text2text-generation",
"translation",
"generated_from_trainer",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
translation
| 2023-07-09T19:18:00Z |
---
tags:
- translation
- generated_from_trainer
metrics:
- bleu
model-index:
- name: eng-ind-nng
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. -->
# eng-ind-nng
This model is a fine-tuned version of [facebook/mbart-large-50-many-to-many-mmt](https://huggingface.co/facebook/mbart-large-50-many-to-many-mmt) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.7826
- Bleu: 20.9168
## 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: 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: 3
- mixed_precision_training: Native AMP
### Training results
### Framework versions
- Transformers 4.26.1
- Pytorch 2.0.1+cu117
- Datasets 2.12.0
- Tokenizers 0.13.3
|
Jonathaniu/alpaca-bitcoin-tweets-sentiment
|
Jonathaniu
| 2023-07-09T19:32:27Z | 5 | 1 |
peft
|
[
"peft",
"region:us"
] | null | 2023-07-08T01:05:11Z |
---
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.dev0
|
dp15/poca-SoccerTwos
|
dp15
| 2023-07-09T19:28:19Z | 1 | 0 |
ml-agents
|
[
"ml-agents",
"tensorboard",
"onnx",
"SoccerTwos",
"deep-reinforcement-learning",
"reinforcement-learning",
"ML-Agents-SoccerTwos",
"region:us"
] |
reinforcement-learning
| 2023-07-09T17:03:18Z |
---
library_name: ml-agents
tags:
- SoccerTwos
- deep-reinforcement-learning
- reinforcement-learning
- ML-Agents-SoccerTwos
---
# **poca** Agent playing **SoccerTwos**
This is a trained model of a **poca** agent playing **SoccerTwos**
using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents).
## Usage (with ML-Agents)
The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/
We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub:
- A *short tutorial* where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your
browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction
- A *longer tutorial* to understand how works ML-Agents:
https://huggingface.co/learn/deep-rl-course/unit5/introduction
### Resume the training
```bash
mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume
```
### Watch your Agent play
You can watch your agent **playing directly in your browser**
1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity
2. Step 1: Find your model_id: dp15/poca-SoccerTwos
3. Step 2: Select your *.nn /*.onnx file
4. Click on Watch the agent play 👀
|
israel-avihail/rl_course_vizdoom_health_gathering_supreme
|
israel-avihail
| 2023-07-09T19:11:32Z | 0 | 0 |
sample-factory
|
[
"sample-factory",
"tensorboard",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-07-09T13:11:55Z |
---
library_name: sample-factory
tags:
- deep-reinforcement-learning
- reinforcement-learning
- sample-factory
model-index:
- name: APPO
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: doom_health_gathering_supreme
type: doom_health_gathering_supreme
metrics:
- type: mean_reward
value: 10.39 +/- 4.97
name: mean_reward
verified: false
---
A(n) **APPO** model trained on the **doom_health_gathering_supreme** environment.
This model was trained using Sample-Factory 2.0: https://github.com/alex-petrenko/sample-factory.
Documentation for how to use Sample-Factory can be found at https://www.samplefactory.dev/
## Downloading the model
After installing Sample-Factory, download the model with:
```
python -m sample_factory.huggingface.load_from_hub -r israel-avihail/rl_course_vizdoom_health_gathering_supreme
```
## Using the model
To run the model after download, use the `enjoy` script corresponding to this environment:
```
python -m .usr.local.lib.python3.10.dist-packages.ipykernel_launcher --algo=APPO --env=doom_health_gathering_supreme --train_dir=./train_dir --experiment=rl_course_vizdoom_health_gathering_supreme
```
You can also upload models to the Hugging Face Hub using the same script with the `--push_to_hub` flag.
See https://www.samplefactory.dev/10-huggingface/huggingface/ for more details
## Training with this model
To continue training with this model, use the `train` script corresponding to this environment:
```
python -m .usr.local.lib.python3.10.dist-packages.ipykernel_launcher --algo=APPO --env=doom_health_gathering_supreme --train_dir=./train_dir --experiment=rl_course_vizdoom_health_gathering_supreme --restart_behavior=resume --train_for_env_steps=10000000000
```
Note, you may have to adjust `--train_for_env_steps` to a suitably high number as the experiment will resume at the number of steps it concluded at.
|
hopkins/eng-deu-nng
|
hopkins
| 2023-07-09T19:05:53Z | 3 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"mbart",
"text2text-generation",
"translation",
"generated_from_trainer",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
translation
| 2023-07-09T18:47:31Z |
---
tags:
- translation
- generated_from_trainer
metrics:
- bleu
model-index:
- name: eng-deu-nng
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. -->
# eng-deu-nng
This model is a fine-tuned version of [facebook/mbart-large-50-many-to-many-mmt](https://huggingface.co/facebook/mbart-large-50-many-to-many-mmt) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.6582
- Bleu: 20.2230
## 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: 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: 3
- mixed_precision_training: Native AMP
### Training results
### Framework versions
- Transformers 4.26.1
- Pytorch 2.0.1+cu117
- Datasets 2.12.0
- Tokenizers 0.13.3
|
Bisht0538/gauravbisht
|
Bisht0538
| 2023-07-09T18:55:53Z | 183 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"rust",
"bart",
"text2text-generation",
"summarization",
"en",
"dataset:cnn_dailymail",
"arxiv:1910.13461",
"license:mit",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
summarization
| 2023-07-09T17:31:09Z |
---
language:
- en
tags:
- summarization
license: mit
thumbnail: https://huggingface.co/front/thumbnails/facebook.png
datasets:
- cnn_dailymail
model-index:
- name: facebook/bart-large-cnn
results:
- task:
type: summarization
name: Summarization
dataset:
name: cnn_dailymail
type: cnn_dailymail
config: 3.0.0
split: train
metrics:
- name: ROUGE-1
type: rouge
value: 42.9486
verified: true
- name: ROUGE-2
type: rouge
value: 20.8149
verified: true
- name: ROUGE-L
type: rouge
value: 30.6186
verified: true
- name: ROUGE-LSUM
type: rouge
value: 40.0376
verified: true
- name: loss
type: loss
value: 2.529000997543335
verified: true
- name: gen_len
type: gen_len
value: 78.5866
verified: true
---
# BART (large-sized model), fine-tuned on CNN Daily Mail
BART model pre-trained on English language, and fine-tuned on [CNN Daily Mail](https://huggingface.co/datasets/cnn_dailymail). It was introduced in the paper [BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension](https://arxiv.org/abs/1910.13461) by Lewis et al. and first released in [this repository (https://github.com/pytorch/fairseq/tree/master/examples/bart).
Disclaimer: The team releasing BART did not write a model card for this model so this model card has been written by the Hugging Face team.
## Model description
BART is a transformer encoder-encoder (seq2seq) model with a bidirectional (BERT-like) encoder and an autoregressive (GPT-like) decoder. BART is pre-trained by (1) corrupting text with an arbitrary noising function, and (2) learning a model to reconstruct the original text.
BART is particularly effective when fine-tuned for text generation (e.g. summarization, translation) but also works well for comprehension tasks (e.g. text classification, question answering). This particular checkpoint has been fine-tuned on CNN Daily Mail, a large collection of text-summary pairs.
## Intended uses & limitations
You can use this model for text summarization.
### How to use
Here is how to use this model with the [pipeline API](https://huggingface.co/transformers/main_classes/pipelines.html):
```python
from transformers import pipeline
summarizer = pipeline("summarization", model="facebook/bart-large-cnn")
ARTICLE = """ New York (CNN)When Liana Barrientos was 23 years old, she got married in Westchester County, New York.
A year later, she got married again in Westchester County, but to a different man and without divorcing her first husband.
Only 18 days after that marriage, she got hitched yet again. Then, Barrientos declared "I do" five more times, sometimes only within two weeks of each other.
In 2010, she married once more, this time in the Bronx. In an application for a marriage license, she stated it was her "first and only" marriage.
Barrientos, now 39, is facing two criminal counts of "offering a false instrument for filing in the first degree," referring to her false statements on the
2010 marriage license application, according to court documents.
Prosecutors said the marriages were part of an immigration scam.
On Friday, she pleaded not guilty at State Supreme Court in the Bronx, according to her attorney, Christopher Wright, who declined to comment further.
After leaving court, Barrientos was arrested and charged with theft of service and criminal trespass for allegedly sneaking into the New York subway through an emergency exit, said Detective
Annette Markowski, a police spokeswoman. In total, Barrientos has been married 10 times, with nine of her marriages occurring between 1999 and 2002.
All occurred either in Westchester County, Long Island, New Jersey or the Bronx. She is believed to still be married to four men, and at one time, she was married to eight men at once, prosecutors say.
Prosecutors said the immigration scam involved some of her husbands, who filed for permanent residence status shortly after the marriages.
Any divorces happened only after such filings were approved. It was unclear whether any of the men will be prosecuted.
The case was referred to the Bronx District Attorney\'s Office by Immigration and Customs Enforcement and the Department of Homeland Security\'s
Investigation Division. Seven of the men are from so-called "red-flagged" countries, including Egypt, Turkey, Georgia, Pakistan and Mali.
Her eighth husband, Rashid Rajput, was deported in 2006 to his native Pakistan after an investigation by the Joint Terrorism Task Force.
If convicted, Barrientos faces up to four years in prison. Her next court appearance is scheduled for May 18.
"""
print(summarizer(ARTICLE, max_length=130, min_length=30, do_sample=False))
>>> [{'summary_text': 'Liana Barrientos, 39, is charged with two counts of "offering a false instrument for filing in the first degree" In total, she has been married 10 times, with nine of her marriages occurring between 1999 and 2002. She is believed to still be married to four men.'}]
```
### BibTeX entry and citation info
```bibtex
@article{DBLP:journals/corr/abs-1910-13461,
author = {Mike Lewis and
Yinhan Liu and
Naman Goyal and
Marjan Ghazvininejad and
Abdelrahman Mohamed and
Omer Levy and
Veselin Stoyanov and
Luke Zettlemoyer},
title = {{BART:} Denoising Sequence-to-Sequence Pre-training for Natural Language
Generation, Translation, and Comprehension},
journal = {CoRR},
volume = {abs/1910.13461},
year = {2019},
url = {http://arxiv.org/abs/1910.13461},
eprinttype = {arXiv},
eprint = {1910.13461},
timestamp = {Thu, 31 Oct 2019 14:02:26 +0100},
biburl = {https://dblp.org/rec/journals/corr/abs-1910-13461.bib},
bibsource = {dblp computer science bibliography, https://dblp.org}
}
|
aryjessen/SkyHawk
|
aryjessen
| 2023-07-09T18:54:28Z | 0 | 0 | null |
[
"text-to-image",
"en",
"region:us"
] |
text-to-image
| 2023-07-09T18:49:43Z |
---
language:
- en
pipeline_tag: text-to-image
---
|
Weikang01/distilbert-base-uncased_emotion_ft_0416
|
Weikang01
| 2023-07-09T18:48:33Z | 105 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"distilbert",
"text-classification",
"generated_from_trainer",
"dataset:emotion",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2023-07-09T03:31:40Z |
---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- emotion
metrics:
- accuracy
- f1
- precision
model-index:
- name: distilbert-base-uncased_emotion_ft_0416
results:
- task:
name: Text Classification
type: text-classification
dataset:
name: emotion
type: emotion
config: split
split: validation
args: split
metrics:
- name: Accuracy
type: accuracy
value: 0.937
- name: F1
type: f1
value: 0.9371267820617502
- name: Precision
type: precision
value: 0.9127268366622657
---
<!-- 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_emotion_ft_0416
This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the emotion dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1487
- Accuracy: 0.937
- F1: 0.9371
- Precision: 0.9127
## 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: 4
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Precision |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|:---------:|
| 0.7892 | 1.0 | 250 | 0.2543 | 0.9235 | 0.9221 | 0.9172 |
| 0.2039 | 2.0 | 500 | 0.1742 | 0.9275 | 0.9276 | 0.9069 |
| 0.1371 | 3.0 | 750 | 0.1521 | 0.9375 | 0.9378 | 0.9104 |
| 0.1108 | 4.0 | 1000 | 0.1487 | 0.937 | 0.9371 | 0.9127 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
hsc748NLP/GujiBERT_jian_fan
|
hsc748NLP
| 2023-07-09T18:38:49Z | 107 | 1 |
transformers
|
[
"transformers",
"pytorch",
"bert",
"fill-mask",
"generated_from_trainer",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2023-07-09T17:03:35Z |
---
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: output
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. -->
# output
This model is a fine-tuned version of [/gemini/data-1/sikubert_vocabtxt](https://huggingface.co//gemini/data-1/sikubert_vocabtxt) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 2.2556
- Accuracy: 0.5514
## 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: 192
- eval_batch_size: 384
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5.0
### Training results
### Framework versions
- Transformers 4.29.2
- Pytorch 2.0.1+cu117
- Datasets 2.12.0
- Tokenizers 0.13.3
|
hsc748NLP/GujiGPT_jian
|
hsc748NLP
| 2023-07-09T18:38:06Z | 137 | 1 |
transformers
|
[
"transformers",
"pytorch",
"gpt2",
"text-generation",
"generated_from_trainer",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2023-07-09T17:15:19Z |
---
tags:
- generated_from_trainer
model-index:
- name: output
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. -->
# output
This model is a fine-tuned version of [/gemini/data-1/gpt2-chinese-cluecorpussmall](https://huggingface.co//gemini/data-1/gpt2-chinese-cluecorpussmall) 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: 64
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
### Training results
### Framework versions
- Transformers 4.29.2
- Pytorch 2.0.1+cu117
- Datasets 2.12.0
- Tokenizers 0.13.3
|
RogerB/roberta-base-finetuned-kintweetsE
|
RogerB
| 2023-07-09T18:28:41Z | 131 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"roberta",
"fill-mask",
"generated_from_trainer",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2023-07-09T18:13:57Z |
---
license: mit
tags:
- generated_from_trainer
model-index:
- name: roberta-base-finetuned-kintweetsE
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. -->
# roberta-base-finetuned-kintweetsE
This model is a fine-tuned version of [roberta-base](https://huggingface.co/roberta-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 2.6524
## 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: 10
- eval_batch_size: 10
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 3.1657 | 1.0 | 1000 | 2.8429 |
| 2.8541 | 2.0 | 2000 | 2.6654 |
| 2.7484 | 3.0 | 3000 | 2.6122 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
nikolai40/iam-trocr
|
nikolai40
| 2023-07-09T18:18:53Z | 46 | 0 |
transformers
|
[
"transformers",
"pytorch",
"vision-encoder-decoder",
"image-text-to-text",
"endpoints_compatible",
"region:us"
] |
image-text-to-text
| 2023-06-08T10:57:44Z |
---
metrics:
- trocr
- image-to-text
- CER
---
# TrOCR model fine-tuned on IAM dataset using augmentation (stretching and dilation)
Using [trocr-small-stage1](https://huggingface.co/microsoft/trocr-small-stage1) version
|
RogerB/distilbert-base-multilingual-cased-finetuned-kintweetsE
|
RogerB
| 2023-07-09T18:12:36Z | 124 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"distilbert",
"fill-mask",
"generated_from_trainer",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2023-07-09T17:56:03Z |
---
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: distilbert-base-multilingual-cased-finetuned-kintweetsE
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-multilingual-cased-finetuned-kintweetsE
This model is a fine-tuned version of [distilbert-base-multilingual-cased](https://huggingface.co/distilbert-base-multilingual-cased) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 3.1438
## 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: 10
- eval_batch_size: 10
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 3.7676 | 1.0 | 1000 | 3.3390 |
| 3.3493 | 2.0 | 2000 | 3.1638 |
| 3.2122 | 3.0 | 3000 | 3.1040 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
mrml/ppo-LunarLander-v2-1000000
|
mrml
| 2023-07-09T18:08:04Z | 0 | 0 |
stable-baselines3
|
[
"stable-baselines3",
"LunarLander-v2",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-07-09T18:07:46Z |
---
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: 263.31 +/- 15.23
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
...
```
|
mrizalf7/t5-small-finetuned-indosum-3
|
mrizalf7
| 2023-07-09T17:46:49Z | 105 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"t5",
"text2text-generation",
"generated_from_trainer",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2023-07-09T16:18:16Z |
---
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: t5-small-finetuned-indosum-3
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. -->
# t5-small-finetuned-indosum-3
This model is a fine-tuned version of [t5-small](https://huggingface.co/t5-small) on the None 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: 10
- eval_batch_size: 10
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 1
- mixed_precision_training: Native AMP
### Framework versions
- Transformers 4.28.0
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
vnktrmnb/bert-base-multilingual-cased-finetuned-tydiqa
|
vnktrmnb
| 2023-07-09T17:42:01Z | 61 | 0 |
transformers
|
[
"transformers",
"tf",
"tensorboard",
"bert",
"question-answering",
"generated_from_keras_callback",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] |
question-answering
| 2023-07-08T19:08:21Z |
---
license: apache-2.0
tags:
- generated_from_keras_callback
model-index:
- name: vnktrmnb/bert-base-multilingual-cased-finetuned-tydiqa
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. -->
# vnktrmnb/bert-base-multilingual-cased-finetuned-tydiqa
This model is a fine-tuned version of [bert-base-multilingual-cased](https://huggingface.co/bert-base-multilingual-cased) on an unknown dataset.
It achieves the following results on the evaluation set:
- Train Loss: 0.6211
- Train End Logits Accuracy: 0.8146
- Train Start Logits Accuracy: 0.8612
- Validation Loss: 0.4720
- Validation End Logits Accuracy: 0.8544
- Validation Start Logits Accuracy: 0.9103
- Epoch: 1
## 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': 'Adam', 'weight_decay': None, 'clipnorm': None, 'global_clipnorm': None, 'clipvalue': None, 'use_ema': False, 'ema_momentum': 0.99, 'ema_overwrite_frequency': None, 'jit_compile': True, 'is_legacy_optimizer': False, 'learning_rate': {'class_name': 'PolynomialDecay', 'config': {'initial_learning_rate': 2e-05, 'decay_steps': 836, 'end_learning_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}}, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False}
- training_precision: float32
### Training results
| Train Loss | Train End Logits Accuracy | Train Start Logits Accuracy | Validation Loss | Validation End Logits Accuracy | Validation Start Logits Accuracy | Epoch |
|:----------:|:-------------------------:|:---------------------------:|:---------------:|:------------------------------:|:--------------------------------:|:-----:|
| 0.9948 | 0.7232 | 0.7723 | 0.4799 | 0.8500 | 0.9029 | 0 |
| 0.6211 | 0.8146 | 0.8612 | 0.4720 | 0.8544 | 0.9103 | 1 |
### Framework versions
- Transformers 4.30.2
- TensorFlow 2.12.0
- Datasets 2.13.1
- Tokenizers 0.13.3
|
Ricky1981/Hjbsm
|
Ricky1981
| 2023-07-09T17:36:09Z | 0 | 0 | null |
[
"license:creativeml-openrail-m",
"region:us"
] | null | 2023-07-09T17:35:14Z |
---
license: creativeml-openrail-m
---
|
RiadxAvatar/rare-puppers
|
RiadxAvatar
| 2023-07-09T17:33:19Z | 215 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"vit",
"image-classification",
"huggingpics",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
image-classification
| 2023-07-09T17:33:12Z |
---
tags:
- image-classification
- pytorch
- huggingpics
metrics:
- accuracy
model-index:
- name: rare-puppers
results:
- task:
name: Image Classification
type: image-classification
metrics:
- name: Accuracy
type: accuracy
value: 0.8484848737716675
---
# rare-puppers
Autogenerated by HuggingPics🤗🖼️
Create your own image classifier for **anything** by running [the demo on Google Colab](https://colab.research.google.com/github/nateraw/huggingpics/blob/main/HuggingPics.ipynb).
Report any issues with the demo at the [github repo](https://github.com/nateraw/huggingpics).
## Example Images
#### corgi
#### samoyed
#### shiba inu
|
shauryakudiyal/fine-tuned-bart
|
shauryakudiyal
| 2023-07-09T17:16:22Z | 178 | 2 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"rust",
"bart",
"text2text-generation",
"summarization",
"en",
"dataset:cnn_dailymail",
"arxiv:1910.13461",
"license:mit",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
summarization
| 2023-01-22T20:39:04Z |
---
language:
- en
tags:
- summarization
license: mit
thumbnail: https://huggingface.co/front/thumbnails/facebook.png
datasets:
- cnn_dailymail
model-index:
- name: facebook/bart-large-cnn
results:
- task:
type: summarization
name: Summarization
dataset:
name: cnn_dailymail
type: cnn_dailymail
config: 3.0.0
split: train
metrics:
- name: ROUGE-1
type: rouge
value: 42.9486
verified: true
- name: ROUGE-2
type: rouge
value: 20.8149
verified: true
- name: ROUGE-L
type: rouge
value: 30.6186
verified: true
- name: ROUGE-LSUM
type: rouge
value: 40.0376
verified: true
- name: loss
type: loss
value: 2.529000997543335
verified: true
- name: gen_len
type: gen_len
value: 78.5866
verified: true
---
# BART (large-sized model), fine-tuned on CNN Daily Mail
BART model pre-trained on English language, and fine-tuned on [CNN Daily Mail](https://huggingface.co/datasets/cnn_dailymail). It was introduced in the paper [BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension](https://arxiv.org/abs/1910.13461) by Lewis et al. and first released in [this repository (https://github.com/pytorch/fairseq/tree/master/examples/bart).
Disclaimer: The team releasing BART did not write a model card for this model so this model card has been written by the Hugging Face team.
## Model description
BART is a transformer encoder-encoder (seq2seq) model with a bidirectional (BERT-like) encoder and an autoregressive (GPT-like) decoder. BART is pre-trained by (1) corrupting text with an arbitrary noising function, and (2) learning a model to reconstruct the original text.
BART is particularly effective when fine-tuned for text generation (e.g. summarization, translation) but also works well for comprehension tasks (e.g. text classification, question answering). This particular checkpoint has been fine-tuned on CNN Daily Mail, a large collection of text-summary pairs.
## Intended uses & limitations
You can use this model for text summarization.
### How to use
Here is how to use this model with the [pipeline API](https://huggingface.co/transformers/main_classes/pipelines.html):
```python
from transformers import pipeline
summarizer = pipeline("summarization", model="facebook/bart-large-cnn")
ARTICLE = """ New York (CNN)When Liana Barrientos was 23 years old, she got married in Westchester County, New York.
A year later, she got married again in Westchester County, but to a different man and without divorcing her first husband.
Only 18 days after that marriage, she got hitched yet again. Then, Barrientos declared "I do" five more times, sometimes only within two weeks of each other.
In 2010, she married once more, this time in the Bronx. In an application for a marriage license, she stated it was her "first and only" marriage.
Barrientos, now 39, is facing two criminal counts of "offering a false instrument for filing in the first degree," referring to her false statements on the
2010 marriage license application, according to court documents.
Prosecutors said the marriages were part of an immigration scam.
On Friday, she pleaded not guilty at State Supreme Court in the Bronx, according to her attorney, Christopher Wright, who declined to comment further.
After leaving court, Barrientos was arrested and charged with theft of service and criminal trespass for allegedly sneaking into the New York subway through an emergency exit, said Detective
Annette Markowski, a police spokeswoman. In total, Barrientos has been married 10 times, with nine of her marriages occurring between 1999 and 2002.
All occurred either in Westchester County, Long Island, New Jersey or the Bronx. She is believed to still be married to four men, and at one time, she was married to eight men at once, prosecutors say.
Prosecutors said the immigration scam involved some of her husbands, who filed for permanent residence status shortly after the marriages.
Any divorces happened only after such filings were approved. It was unclear whether any of the men will be prosecuted.
The case was referred to the Bronx District Attorney\'s Office by Immigration and Customs Enforcement and the Department of Homeland Security\'s
Investigation Division. Seven of the men are from so-called "red-flagged" countries, including Egypt, Turkey, Georgia, Pakistan and Mali.
Her eighth husband, Rashid Rajput, was deported in 2006 to his native Pakistan after an investigation by the Joint Terrorism Task Force.
If convicted, Barrientos faces up to four years in prison. Her next court appearance is scheduled for May 18.
"""
print(summarizer(ARTICLE, max_length=130, min_length=30, do_sample=False))
>>> [{'summary_text': 'Liana Barrientos, 39, is charged with two counts of "offering a false instrument for filing in the first degree" In total, she has been married 10 times, with nine of her marriages occurring between 1999 and 2002. She is believed to still be married to four men.'}]
```
### BibTeX entry and citation info
```bibtex
@article{DBLP:journals/corr/abs-1910-13461,
author = {Mike Lewis and
Yinhan Liu and
Naman Goyal and
Marjan Ghazvininejad and
Abdelrahman Mohamed and
Omer Levy and
Veselin Stoyanov and
Luke Zettlemoyer},
title = {{BART:} Denoising Sequence-to-Sequence Pre-training for Natural Language
Generation, Translation, and Comprehension},
journal = {CoRR},
volume = {abs/1910.13461},
year = {2019},
url = {http://arxiv.org/abs/1910.13461},
eprinttype = {arXiv},
eprint = {1910.13461},
timestamp = {Thu, 31 Oct 2019 14:02:26 +0100},
biburl = {https://dblp.org/rec/journals/corr/abs-1910-13461.bib},
bibsource = {dblp computer science bibliography, https://dblp.org}
}
|
hsc748NLP/GujiRoBERTa_fan
|
hsc748NLP
| 2023-07-09T17:14:45Z | 111 | 1 |
transformers
|
[
"transformers",
"pytorch",
"bert",
"fill-mask",
"generated_from_trainer",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2023-07-09T16:41:37Z |
---
tags:
- generated_from_trainer
model-index:
- name: output
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. -->
# output
This model is a fine-tuned version of [/gemini/data-1/sikuroberta_vocabtxt](https://huggingface.co//gemini/data-1/sikuroberta_vocabtxt) on the None 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: 192
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5.0
### Training results
### Framework versions
- Transformers 4.29.2
- Pytorch 2.0.1+cu117
- Datasets 2.12.0
- Tokenizers 0.13.3
|
turhancan97/yolov5-detect-trash-classification
|
turhancan97
| 2023-07-09T17:11:13Z | 0 | 2 | null |
[
"object-detection",
"computer-vision",
"yolov5",
"en",
"dataset:garythung/trashnet",
"dataset:Zesky665/TACO",
"dataset:detection-datasets/coco",
"license:mit",
"region:us"
] |
object-detection
| 2023-07-09T16:59:04Z |
---
license: mit
datasets:
- garythung/trashnet
- Zesky665/TACO
- detection-datasets/coco
language:
- en
tags:
- object-detection
- computer-vision
- yolov5
---
# Examples
<div align="center">
<img width="416" alt="turhancan97/yolov5-detect-trash-classification" src="https://huggingface.co/turhancan97/yolov5-detect-trash-classification/resolve/main/example1.jpg">
</div>
<div align="center">
<img width="416" alt="turhancan97/yolov5-detect-trash-classification" src="https://huggingface.co/turhancan97/yolov5-detect-trash-classification/resolve/main/example2.jpg">
</div>
<div align="center">
<img width="416" alt="turhancan97/yolov5-detect-trash-classification" src="https://huggingface.co/turhancan97/yolov5-detect-trash-classification/resolve/main/example3.jpg">
</div>
### How to use
- Install [yolov5](https://github.com/fcakyon/yolov5-pip):
```bash
pip install -U yolov5
```
- Load model and perform prediction:
```python
import yolov5
# load model
model = yolov5.load('turhancan97/yolov5-detect-trash-classification')
# set model parameters
model.conf = 0.25 # NMS confidence threshold
model.iou = 0.45 # NMS IoU threshold
model.agnostic = False # NMS class-agnostic
model.multi_label = False # NMS multiple labels per box
model.max_det = 1000 # maximum number of detections per image
# set image
img = 'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg'
# perform inference
results = model(img, size=416)
# inference with test time augmentation
results = model(img, augment=True)
# parse results
predictions = results.pred[0]
boxes = predictions[:, :4] # x1, y1, x2, y2
scores = predictions[:, 4]
categories = predictions[:, 5]
# show detection bounding boxes on image
results.show()
# save results into "results/" folder
results.save(save_dir='results/')
```
- Finetune the model on your custom dataset:
```bash
yolov5 train --data data.yaml --img 416 --batch 16 --weights turhancan97/yolov5-detect-trash-classification --epochs 10
```
|
guaguale/model_kthv_v1
|
guaguale
| 2023-07-09T17:11:11Z | 1 | 0 |
diffusers
|
[
"diffusers",
"tensorboard",
"stable-diffusion",
"stable-diffusion-diffusers",
"text-to-image",
"dreambooth",
"base_model:runwayml/stable-diffusion-v1-5",
"base_model:finetune:runwayml/stable-diffusion-v1-5",
"license:creativeml-openrail-m",
"autotrain_compatible",
"endpoints_compatible",
"diffusers:StableDiffusionPipeline",
"region:us"
] |
text-to-image
| 2023-07-09T12:03:40Z |
---
license: creativeml-openrail-m
base_model: runwayml/stable-diffusion-v1-5
instance_prompt: a male idol sks with blonde hair, wearing a black jacket and fringes on the sides of the jacket
tags:
- stable-diffusion
- stable-diffusion-diffusers
- text-to-image
- diffusers
- dreambooth
inference: true
---
# DreamBooth - guaguale/model_kthv_v1
This is a dreambooth model derived from runwayml/stable-diffusion-v1-5. The weights were trained on a male idol sks with blonde hair, wearing a black jacket and fringes on the sides of the jacket using [DreamBooth](https://dreambooth.github.io/).
You can find some example images in the following.
DreamBooth for the text encoder was enabled: False.
|
digiplay/LuckyStrikeMix0.2Realistic
|
digiplay
| 2023-07-09T17:07:13Z | 311 | 1 |
diffusers
|
[
"diffusers",
"safetensors",
"stable-diffusion",
"stable-diffusion-diffusers",
"text-to-image",
"license:other",
"autotrain_compatible",
"endpoints_compatible",
"diffusers:StableDiffusionPipeline",
"region:us"
] |
text-to-image
| 2023-07-08T13:03:32Z |
---
license: other
tags:
- stable-diffusion
- stable-diffusion-diffusers
- text-to-image
- diffusers
inference: true
---
Model info :
https://civitai.com/models/13034/lucky-strike-mix
***Note: please use "realistic" keywords to make some realistic results.***
Sample image I made thru huggingface's API:
```
realistic ,MCU,(masterpiece, best quality, ultra high res:1.3), 1girl, (abstract art:1.3), half demon,
```
|
SwampMan/Reinforce-1
|
SwampMan
| 2023-07-09T17:05:57Z | 0 | 0 | null |
[
"CartPole-v1",
"reinforce",
"reinforcement-learning",
"custom-implementation",
"deep-rl-class",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-07-09T17:05:46Z |
---
tags:
- CartPole-v1
- reinforce
- reinforcement-learning
- custom-implementation
- deep-rl-class
model-index:
- name: Reinforce-1
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: CartPole-v1
type: CartPole-v1
metrics:
- type: mean_reward
value: 500.00 +/- 0.00
name: mean_reward
verified: false
---
# **Reinforce** Agent playing **CartPole-v1**
This is a trained model of a **Reinforce** agent playing **CartPole-v1** .
To learn to use this model and train yours check Unit 4 of the Deep Reinforcement Learning Course: https://huggingface.co/deep-rl-course/unit4/introduction
|
RogerB/afriberta_small-finetuned-kintweetsD
|
RogerB
| 2023-07-09T17:05:25Z | 103 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"xlm-roberta",
"fill-mask",
"generated_from_trainer",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2023-07-09T16:55:42Z |
---
tags:
- generated_from_trainer
model-index:
- name: afriberta_small-finetuned-kintweetsD
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. -->
# afriberta_small-finetuned-kintweetsD
This model is a fine-tuned version of [castorini/afriberta_small](https://huggingface.co/castorini/afriberta_small) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 3.2332
## 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: 10
- eval_batch_size: 10
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 3.6431 | 1.0 | 900 | 3.3421 |
| 3.4111 | 2.0 | 1800 | 3.2661 |
| 3.3391 | 3.0 | 2700 | 3.2382 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
ericNguyen0132/roberta-large-Dep-second
|
ericNguyen0132
| 2023-07-09T16:54:58Z | 3 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"roberta",
"text-classification",
"generated_from_trainer",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2023-07-07T16:25:28Z |
---
tags:
- generated_from_trainer
metrics:
- accuracy
- f1
model-index:
- name: roberta-large-Dep-second
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. -->
# roberta-large-Dep-second
This model is a fine-tuned version of [rafalposwiata/deproberta-large-depression](https://huggingface.co/rafalposwiata/deproberta-large-depression) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.1600
- Accuracy: 0.8517
- F1: 0.9113
## 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: 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: 10
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|
| No log | 1.0 | 469 | 0.3551 | 0.86 | 0.9188 |
| 0.3676 | 2.0 | 938 | 0.4666 | 0.8617 | 0.9198 |
| 0.3042 | 3.0 | 1407 | 0.5818 | 0.86 | 0.9170 |
| 0.2651 | 4.0 | 1876 | 0.8291 | 0.865 | 0.9200 |
| 0.174 | 5.0 | 2345 | 0.8843 | 0.8567 | 0.9155 |
| 0.1363 | 6.0 | 2814 | 1.1669 | 0.8317 | 0.8968 |
| 0.075 | 7.0 | 3283 | 1.2803 | 0.8283 | 0.8952 |
| 0.0401 | 8.0 | 3752 | 1.0247 | 0.8617 | 0.9184 |
| 0.0301 | 9.0 | 4221 | 1.2848 | 0.83 | 0.8961 |
| 0.0281 | 10.0 | 4690 | 1.1600 | 0.8517 | 0.9113 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
RogerB/afriberta_base-finetuned-kintweetsD
|
RogerB
| 2023-07-09T16:51:31Z | 103 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"xlm-roberta",
"fill-mask",
"generated_from_trainer",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2023-07-09T16:38:47Z |
---
tags:
- generated_from_trainer
model-index:
- name: afriberta_base-finetuned-kintweetsD
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. -->
# afriberta_base-finetuned-kintweetsD
This model is a fine-tuned version of [castorini/afriberta_base](https://huggingface.co/castorini/afriberta_base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 3.0707
## 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: 10
- eval_batch_size: 10
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 3.4666 | 1.0 | 900 | 3.1772 |
| 3.2296 | 2.0 | 1800 | 3.1050 |
| 3.1467 | 3.0 | 2700 | 3.0831 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
aclodic/taxi-v3
|
aclodic
| 2023-07-09T16:49:05Z | 0 | 0 | null |
[
"Taxi-v3",
"q-learning",
"reinforcement-learning",
"custom-implementation",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-07-09T16:49:04Z |
---
tags:
- Taxi-v3
- q-learning
- reinforcement-learning
- custom-implementation
model-index:
- name: taxi-v3
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: Taxi-v3
type: Taxi-v3
metrics:
- type: mean_reward
value: 7.50 +/- 2.78
name: mean_reward
verified: false
---
# **Q-Learning** Agent playing1 **Taxi-v3**
This is a trained model of a **Q-Learning** agent playing **Taxi-v3** .
## Usage
```python
model = load_from_hub(repo_id="aclodic/taxi-v3", filename="q-learning.pkl")
# Don't forget to check if you need to add additional attributes (is_slippery=False etc)
env = gym.make(model["env_id"])
```
|
Hedayat-Abrishami/ppo-SnowballTarget
|
Hedayat-Abrishami
| 2023-07-09T16:37:35Z | 3 | 0 |
ml-agents
|
[
"ml-agents",
"tensorboard",
"onnx",
"SnowballTarget",
"deep-reinforcement-learning",
"reinforcement-learning",
"ML-Agents-SnowballTarget",
"region:us"
] |
reinforcement-learning
| 2023-07-09T16:37:33Z |
---
library_name: ml-agents
tags:
- SnowballTarget
- deep-reinforcement-learning
- reinforcement-learning
- ML-Agents-SnowballTarget
---
# **ppo** Agent playing **SnowballTarget**
This is a trained model of a **ppo** agent playing **SnowballTarget**
using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents).
## Usage (with ML-Agents)
The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/
We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub:
- A *short tutorial* where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your
browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction
- A *longer tutorial* to understand how works ML-Agents:
https://huggingface.co/learn/deep-rl-course/unit5/introduction
### Resume the training
```bash
mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume
```
### Watch your Agent play
You can watch your agent **playing directly in your browser**
1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity
2. Step 1: Find your model_id: Hedayat-Abrishami/ppo-SnowballTarget
3. Step 2: Select your *.nn /*.onnx file
4. Click on Watch the agent play 👀
|
etweedy/roberta-base-squad-v2
|
etweedy
| 2023-07-09T16:37:08Z | 132 | 0 |
transformers
|
[
"transformers",
"pytorch",
"safetensors",
"roberta",
"question-answering",
"en",
"dataset:squad_v2",
"arxiv:1907.11692",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] |
question-answering
| 2023-07-06T03:34:13Z |
---
datasets:
- squad_v2
language:
- en
license: apache-2.0
inference:
parameters:
handle_impossible_answer: true
---
# Model Card for etweedy/roberta-base-squad-v2
An instance of [roberta-base for QA](https://huggingface.co/roberta-base) which was fine-tuned for context-based question answering on the [SQuAD v2 dataset](https://huggingface.co/datasets/squad_v2), a dataset of English-language context-question-answer triples designed for extractive question answering training and benchmarking. Version 2 of SQuAD (Stanford Question Answering Dataset) contains the 100,000 examples from SQuAD Version 1.1, along with 50,000 additional "unanswerable" questions, i.e. questions whose answer cannot be found in the provided context.
The original RoBERTa (Robustly Optimized BERT Pretraining Approach) model was introduced in [this paper](https://arxiv.org/abs/1907.11692) and [this repository](https://github.com/facebookresearch/fairseq/tree/main/examples/roberta)
## Demonstration space
Try out inference on this model using [this app](https://huggingface.co/spaces/etweedy/roberta-squad-v2)
## Overview
**Pretrained model:** [roberta-base](https://huggingface.co/roberta-base)
**Language:** English
**Downstream-task:** Extractive QA
**Training data:** [SQuAD v2](https://huggingface.co/datasets/squad_v2) train split
**Eval data:** [SQuAD v2](https://huggingface.co/datasets/squad_v2) validation split
## How to Get Started with the Model
Initializing pipeline:
```python
from transformers import AutoModelForQuestionAnswering, AutoTokenizer, pipeline
repo_id = "etweedy/roberta-base-squad-v2"
QA_pipeline = pipeline(
task = 'question-answering',
model=repo_id,
tokenizer=repo_id,
handle_impossible_answer = True
)
```
Inference:
```python
input = {
'question': 'Who invented Twinkies?',
'context': 'Twinkies were invented on April 6, 1930, by Canadian-born baker James Alexander Dewar for the Continental Baking Company in Schiller Park, Illinois.'
}
response = QA_pipeline(**input)
```
### Training Hyperparameters
```
batch_size = 16
n_epochs = 3
learning_rate = 3e-5
base_LM_model = ["roberta-base"](https://huggingface.co/roberta-base)
max_seq_len = 384
stride=128
lr_schedule = LinearWarmup
warmup_proportion = 0.0
mixed_precision="fp16"
```
## Evaluation results
The model was evaluated on the validation split of [SQuAD v2](https://huggingface.co/datasets/squad_v2) and attained the following results:
```python
{"exact": 79.87029394424324,
"f1": 82.91251169582613,
"total": 11873,
"HasAns_exact": 77.93522267206478,
"HasAns_f1": 84.02838248389763,
"HasAns_total": 5928,
"NoAns_exact": 81.79983179142137,
"NoAns_f1": 81.79983179142137,
"NoAns_total": 5945}
```
**BibTeX base model citation:**
```bibtex
@article{DBLP:journals/corr/abs-1907-11692,
author = {Yinhan Liu and
Myle Ott and
Naman Goyal and
Jingfei Du and
Mandar Joshi and
Danqi Chen and
Omer Levy and
Mike Lewis and
Luke Zettlemoyer and
Veselin Stoyanov},
title = {RoBERTa: {A} Robustly Optimized {BERT} Pretraining Approach},
journal = {CoRR},
volume = {abs/1907.11692},
year = {2019},
url = {http://arxiv.org/abs/1907.11692},
archivePrefix = {arXiv},
eprint = {1907.11692},
timestamp = {Thu, 01 Aug 2019 08:59:33 +0200},
biburl = {https://dblp.org/rec/journals/corr/abs-1907-11692.bib},
bibsource = {dblp computer science bibliography, https://dblp.org}
}
```
|
RogerB/afriberta_large-finetuned-kintweetsD
|
RogerB
| 2023-07-09T16:36:04Z | 103 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"xlm-roberta",
"fill-mask",
"generated_from_trainer",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2023-07-09T16:19:51Z |
---
license: mit
tags:
- generated_from_trainer
model-index:
- name: afriberta_large-finetuned-kintweetsD
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. -->
# afriberta_large-finetuned-kintweetsD
This model is a fine-tuned version of [castorini/afriberta_large](https://huggingface.co/castorini/afriberta_large) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 3.0184
## 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: 10
- eval_batch_size: 10
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 3.417 | 1.0 | 900 | 3.1264 |
| 3.1701 | 2.0 | 1800 | 3.0456 |
| 3.0911 | 3.0 | 2700 | 3.0284 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
hugfacerhaha/Reinforce-heli
|
hugfacerhaha
| 2023-07-09T16:18:14Z | 0 | 0 | null |
[
"Pixelcopter-PLE-v0",
"reinforce",
"reinforcement-learning",
"custom-implementation",
"deep-rl-class",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-07-09T16:18:12Z |
---
tags:
- Pixelcopter-PLE-v0
- reinforce
- reinforcement-learning
- custom-implementation
- deep-rl-class
model-index:
- name: Reinforce-heli
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: Pixelcopter-PLE-v0
type: Pixelcopter-PLE-v0
metrics:
- type: mean_reward
value: 16.60 +/- 12.43
name: mean_reward
verified: false
---
# **Reinforce** Agent playing **Pixelcopter-PLE-v0**
This is a trained model of a **Reinforce** agent playing **Pixelcopter-PLE-v0** .
To learn to use this model and train yours check Unit 4 of the Deep Reinforcement Learning Course: https://huggingface.co/deep-rl-course/unit4/introduction
|
aclodic/q-FrozenLake-v1-4x4-noSlippery
|
aclodic
| 2023-07-09T16:13:58Z | 0 | 0 | null |
[
"FrozenLake-v1-4x4-no_slippery",
"q-learning",
"reinforcement-learning",
"custom-implementation",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-07-09T16:13:56Z |
---
tags:
- FrozenLake-v1-4x4-no_slippery
- q-learning
- reinforcement-learning
- custom-implementation
model-index:
- name: q-FrozenLake-v1-4x4-noSlippery
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: FrozenLake-v1-4x4-no_slippery
type: FrozenLake-v1-4x4-no_slippery
metrics:
- type: mean_reward
value: 1.00 +/- 0.00
name: mean_reward
verified: false
---
# **Q-Learning** Agent playing1 **FrozenLake-v1**
This is a trained model of a **Q-Learning** agent playing **FrozenLake-v1** .
## Usage
```python
model = load_from_hub(repo_id="aclodic/q-FrozenLake-v1-4x4-noSlippery", filename="q-learning.pkl")
# Don't forget to check if you need to add additional attributes (is_slippery=False etc)
env = gym.make(model["env_id"])
```
|
gautam1989/distilbert-base-uncased-finetuned-squad-d5716d28
|
gautam1989
| 2023-07-09T16:11:21Z | 107 | 0 |
transformers
|
[
"transformers",
"pytorch",
"distilbert",
"fill-mask",
"question-answering",
"en",
"dataset:squad",
"arxiv:1910.01108",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
question-answering
| 2023-07-09T15:57:51Z |
---
language:
- en
thumbnail: https://github.com/karanchahal/distiller/blob/master/distiller.jpg
tags:
- question-answering
license: apache-2.0
datasets:
- squad
metrics:
- squad
---
# DistilBERT with a second step of distillation
## Model description
This model replicates the "DistilBERT (D)" model from Table 2 of the [DistilBERT paper](https://arxiv.org/pdf/1910.01108.pdf). In this approach, a DistilBERT student is fine-tuned on SQuAD v1.1, but with a BERT model (also fine-tuned on SQuAD v1.1) acting as a teacher for a second step of task-specific distillation.
In this version, the following pre-trained models were used:
* Student: `distilbert-base-uncased`
* Teacher: `lewtun/bert-base-uncased-finetuned-squad-v1`
## Training data
This model was trained on the SQuAD v1.1 dataset which can be obtained from the `datasets` library as follows:
```python
from datasets import load_dataset
squad = load_dataset('squad')
```
## Training procedure
## Eval results
| | Exact Match | F1 |
|------------------|-------------|------|
| DistilBERT paper | 79.1 | 86.9 |
| Ours | 78.4 | 86.5 |
The scores were calculated using the `squad` metric from `datasets`.
### BibTeX entry and citation info
```bibtex
@misc{sanh2020distilbert,
title={DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter},
author={Victor Sanh and Lysandre Debut and Julien Chaumond and Thomas Wolf},
year={2020},
eprint={1910.01108},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
```
|
EleutherAI/pythia-70m-deduped
|
EleutherAI
| 2023-07-09T16:07:33Z | 122,668 | 25 |
transformers
|
[
"transformers",
"pytorch",
"safetensors",
"gpt_neox",
"text-generation",
"causal-lm",
"pythia",
"en",
"dataset:EleutherAI/the_pile_deduplicated",
"arxiv:2304.01373",
"arxiv:2101.00027",
"arxiv:2201.07311",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2023-02-13T16:01:41Z |
---
language:
- en
tags:
- pytorch
- causal-lm
- pythia
license: apache-2.0
datasets:
- EleutherAI/the_pile_deduplicated
---
The *Pythia Scaling Suite* is a collection of models developed to facilitate
interpretability research [(see paper)](https://arxiv.org/pdf/2304.01373.pdf).
It contains two sets of eight models of sizes
70M, 160M, 410M, 1B, 1.4B, 2.8B, 6.9B, and 12B. For each size, there are two
models: one trained on the Pile, and one trained on the Pile after the dataset
has been globally deduplicated. All 8 model sizes are trained on the exact
same data, in the exact same order. We also provide 154 intermediate
checkpoints per model, hosted on Hugging Face as branches.
The Pythia model suite was designed to promote scientific
research on large language models, especially interpretability research.
Despite not centering downstream performance as a design goal, we find the
models <a href="#evaluations">match or exceed</a> the performance of
similar and same-sized models, such as those in the OPT and GPT-Neo suites.
<details>
<summary style="font-weight:600">Details on previous early release and naming convention.</summary>
Previously, we released an early version of the Pythia suite to the public.
However, we decided to retrain the model suite to address a few hyperparameter
discrepancies. This model card <a href="#changelog">lists the changes</a>;
see appendix B in the Pythia paper for further discussion. We found no
difference in benchmark performance between the two Pythia versions.
The old models are
[still available](https://huggingface.co/models?other=pythia_v0), but we
suggest the retrained suite if you are just starting to use Pythia.<br>
**This is the current release.**
Please note that all models in the *Pythia* suite were renamed in January
2023. For clarity, a <a href="#naming-convention-and-parameter-count">table
comparing the old and new names</a> is provided in this model card, together
with exact parameter counts.
</details>
<br>
# Pythia-70M-deduped
## Model Details
- Developed by: [EleutherAI](http://eleuther.ai)
- Model type: Transformer-based Language Model
- Language: English
- Learn more: [Pythia's GitHub repository](https://github.com/EleutherAI/pythia)
for training procedure, config files, and details on how to use.
[See paper](https://arxiv.org/pdf/2304.01373.pdf) for more evals and implementation
details.
- Library: [GPT-NeoX](https://github.com/EleutherAI/gpt-neox)
- License: Apache 2.0
- Contact: to ask questions about this model, join the [EleutherAI
Discord](https://discord.gg/zBGx3azzUn), and post them in `#release-discussion`.
Please read the existing *Pythia* documentation before asking about it in the
EleutherAI Discord. For general correspondence: [contact@eleuther.
ai](mailto:contact@eleuther.ai).
<figure>
| Pythia model | Non-Embedding Params | Layers | Model Dim | Heads | Batch Size | Learning Rate | Equivalent Models |
| -----------: | -------------------: | :----: | :-------: | :---: | :--------: | :-------------------: | :--------------------: |
| 70M | 18,915,328 | 6 | 512 | 8 | 2M | 1.0 x 10<sup>-3</sup> | — |
| 160M | 85,056,000 | 12 | 768 | 12 | 2M | 6.0 x 10<sup>-4</sup> | GPT-Neo 125M, OPT-125M |
| 410M | 302,311,424 | 24 | 1024 | 16 | 2M | 3.0 x 10<sup>-4</sup> | OPT-350M |
| 1.0B | 805,736,448 | 16 | 2048 | 8 | 2M | 3.0 x 10<sup>-4</sup> | — |
| 1.4B | 1,208,602,624 | 24 | 2048 | 16 | 2M | 2.0 x 10<sup>-4</sup> | GPT-Neo 1.3B, OPT-1.3B |
| 2.8B | 2,517,652,480 | 32 | 2560 | 32 | 2M | 1.6 x 10<sup>-4</sup> | GPT-Neo 2.7B, OPT-2.7B |
| 6.9B | 6,444,163,072 | 32 | 4096 | 32 | 2M | 1.2 x 10<sup>-4</sup> | OPT-6.7B |
| 12B | 11,327,027,200 | 36 | 5120 | 40 | 2M | 1.2 x 10<sup>-4</sup> | — |
<figcaption>Engineering details for the <i>Pythia Suite</i>. Deduped and
non-deduped models of a given size have the same hyperparameters. “Equivalent”
models have <b>exactly</b> the same architecture, and the same number of
non-embedding parameters.</figcaption>
</figure>
## Uses and Limitations
### Intended Use
The primary intended use of Pythia is research on the behavior, functionality,
and limitations of large language models. This suite is intended to provide
a controlled setting for performing scientific experiments. We also provide
154 checkpoints per model: initial `step0`, 10 log-spaced checkpoints
`step{1,2,4...512}`, and 143 evenly-spaced checkpoints from `step1000` to
`step143000`. These checkpoints are hosted on Hugging Face as branches. Note
that branch `143000` corresponds exactly to the model checkpoint on the `main`
branch of each model.
You may also further fine-tune and adapt Pythia-70M-deduped for deployment,
as long as your use is in accordance with the Apache 2.0 license. Pythia
models work with the Hugging Face [Transformers
Library](https://huggingface.co/docs/transformers/index). If you decide to use
pre-trained Pythia-70M-deduped as a basis for your fine-tuned model, please
conduct your own risk and bias assessment.
### Out-of-scope use
The Pythia Suite is **not** intended for deployment. It is not a in itself
a product and cannot be used for human-facing interactions. For example,
the model may generate harmful or offensive text. Please evaluate the risks
associated with your particular use case.
Pythia models are English-language only, and are not suitable for translation
or generating text in other languages.
Pythia-70M-deduped has not been fine-tuned for downstream contexts in which
language models are commonly deployed, such as writing genre prose,
or commercial chatbots. This means Pythia-70M-deduped will **not**
respond to a given prompt the way a product like ChatGPT does. This is because,
unlike this model, ChatGPT was fine-tuned using methods such as Reinforcement
Learning from Human Feedback (RLHF) to better “follow” human instructions.
### Limitations and biases
The core functionality of a large language model is to take a string of text
and predict the next token. The token used by the model need not produce the
most “accurate” text. Never rely on Pythia-70M-deduped to produce factually accurate
output.
This model was trained on [the Pile](https://pile.eleuther.ai/), a dataset
known to contain profanity and texts that are lewd or otherwise offensive.
See [Section 6 of the Pile paper](https://arxiv.org/abs/2101.00027) for a
discussion of documented biases with regards to gender, religion, and race.
Pythia-70M-deduped may produce socially unacceptable or undesirable text, *even if*
the prompt itself does not include anything explicitly offensive.
If you plan on using text generated through, for example, the Hosted Inference
API, we recommend having a human curate the outputs of this language model
before presenting it to other people. Please inform your audience that the
text was generated by Pythia-70M-deduped.
### Quickstart
Pythia models can be loaded and used via the following code, demonstrated here
for the third `pythia-70m-deduped` checkpoint:
```python
from transformers import GPTNeoXForCausalLM, AutoTokenizer
model = GPTNeoXForCausalLM.from_pretrained(
"EleutherAI/pythia-70m-deduped",
revision="step3000",
cache_dir="./pythia-70m-deduped/step3000",
)
tokenizer = AutoTokenizer.from_pretrained(
"EleutherAI/pythia-70m-deduped",
revision="step3000",
cache_dir="./pythia-70m-deduped/step3000",
)
inputs = tokenizer("Hello, I am", return_tensors="pt")
tokens = model.generate(**inputs)
tokenizer.decode(tokens[0])
```
Revision/branch `step143000` corresponds exactly to the model checkpoint on
the `main` branch of each model.<br>
For more information on how to use all Pythia models, see [documentation on
GitHub](https://github.com/EleutherAI/pythia).
## Training
### Training data
Pythia-70M-deduped was trained on the Pile **after the dataset has been globally
deduplicated**.<br>
[The Pile](https://pile.eleuther.ai/) is a 825GiB general-purpose dataset in
English. It was created by EleutherAI specifically for training large language
models. It contains texts from 22 diverse sources, roughly broken down into
five categories: academic writing (e.g. arXiv), internet (e.g. CommonCrawl),
prose (e.g. Project Gutenberg), dialogue (e.g. YouTube subtitles), and
miscellaneous (e.g. GitHub, Enron Emails). See [the Pile
paper](https://arxiv.org/abs/2101.00027) for a breakdown of all data sources,
methodology, and a discussion of ethical implications. Consult [the
datasheet](https://arxiv.org/abs/2201.07311) for more detailed documentation
about the Pile and its component datasets. The Pile can be downloaded from
the [official website](https://pile.eleuther.ai/), or from a [community
mirror](https://the-eye.eu/public/AI/pile/).
### Training procedure
All models were trained on the exact same data, in the exact same order. Each
model saw 299,892,736,000 tokens during training, and 143 checkpoints for each
model are saved every 2,097,152,000 tokens, spaced evenly throughout training,
from `step1000` to `step143000` (which is the same as `main`). In addition, we
also provide frequent early checkpoints: `step0` and `step{1,2,4...512}`.
This corresponds to training for just under 1 epoch on the Pile for
non-deduplicated models, and about 1.5 epochs on the deduplicated Pile.
All *Pythia* models trained for 143000 steps at a batch size
of 2M (2,097,152 tokens).<br>
See [GitHub](https://github.com/EleutherAI/pythia) for more details on training
procedure, including [how to reproduce
it](https://github.com/EleutherAI/pythia/blob/main/README.md#reproducing-training).<br>
Pythia uses the same tokenizer as [GPT-NeoX-
20B](https://huggingface.co/EleutherAI/gpt-neox-20b).
## Evaluations
All 16 *Pythia* models were evaluated using the [LM Evaluation
Harness](https://github.com/EleutherAI/lm-evaluation-harness). You can access
the results by model and step at `results/json/*` in the [GitHub
repository](https://github.com/EleutherAI/pythia/tree/main/results/json/).<br>
Expand the sections below to see plots of evaluation results for all
Pythia and Pythia-deduped models compared with OPT and BLOOM.
<details>
<summary>LAMBADA – OpenAI</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/lambada_openai_v1.png" style="width:auto"/>
</details>
<details>
<summary>Physical Interaction: Question Answering (PIQA)</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/piqa_v1.png" style="width:auto"/>
</details>
<details>
<summary>WinoGrande</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/winogrande_v1.png" style="width:auto"/>
</details>
<details>
<summary>AI2 Reasoning Challenge—Easy Set</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/arc_easy_v1.png" style="width:auto"/>
</details>
<details>
<summary>SciQ</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/sciq_v1.png" style="width:auto"/>
</details>
## Changelog
This section compares differences between previously released
[Pythia v0](https://huggingface.co/models?other=pythia_v0) and the current
models. See Appendix B of the Pythia paper for further discussion of these
changes and the motivation behind them. We found that retraining Pythia had no
impact on benchmark performance.
- All model sizes are now trained with uniform batch size of 2M tokens.
Previously, the models of size 160M, 410M, and 1.4B parameters were trained
with batch sizes of 4M tokens.
- We added checkpoints at initialization (step 0) and steps {1,2,4,8,16,32,64,
128,256,512} in addition to every 1000 training steps.
- Flash Attention was used in the new retrained suite.
- We remedied a minor inconsistency that existed in the original suite: all
models of size 2.8B parameters or smaller had a learning rate (LR) schedule
which decayed to a minimum LR of 10% the starting LR rate, but the 6.9B and
12B models all used an LR schedule which decayed to a minimum LR of 0. In
the redone training runs, we rectified this inconsistency: all models now were
trained with LR decaying to a minimum of 0.1× their maximum LR.
### Naming convention and parameter count
*Pythia* models were renamed in January 2023. It is possible that the old
naming convention still persists in some documentation by accident. The
current naming convention (70M, 160M, etc.) is based on total parameter count.
<figure style="width:32em">
| current Pythia suffix | old suffix | total params | non-embedding params |
| --------------------: | ---------: | -------------: | -------------------: |
| 70M | 19M | 70,426,624 | 18,915,328 |
| 160M | 125M | 162,322,944 | 85,056,000 |
| 410M | 350M | 405,334,016 | 302,311,424 |
| 1B | 800M | 1,011,781,632 | 805,736,448 |
| 1.4B | 1.3B | 1,414,647,808 | 1,208,602,624 |
| 2.8B | 2.7B | 2,775,208,960 | 2,517,652,480 |
| 6.9B | 6.7B | 6,857,302,016 | 6,444,163,072 |
| 12B | 13B | 11,846,072,320 | 11,327,027,200 |
</figure>
|
EleutherAI/pythia-2.8b-deduped
|
EleutherAI
| 2023-07-09T16:06:37Z | 11,958 | 14 |
transformers
|
[
"transformers",
"pytorch",
"safetensors",
"gpt_neox",
"text-generation",
"causal-lm",
"pythia",
"en",
"dataset:EleutherAI/the_pile_deduplicated",
"arxiv:2304.01373",
"arxiv:2101.00027",
"arxiv:2201.07311",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2023-02-10T22:26:20Z |
---
language:
- en
tags:
- pytorch
- causal-lm
- pythia
license: apache-2.0
datasets:
- EleutherAI/the_pile_deduplicated
---
The *Pythia Scaling Suite* is a collection of models developed to facilitate
interpretability research [(see paper)](https://arxiv.org/pdf/2304.01373.pdf).
It contains two sets of eight models of sizes
70M, 160M, 410M, 1B, 1.4B, 2.8B, 6.9B, and 12B. For each size, there are two
models: one trained on the Pile, and one trained on the Pile after the dataset
has been globally deduplicated. All 8 model sizes are trained on the exact
same data, in the exact same order. We also provide 154 intermediate
checkpoints per model, hosted on Hugging Face as branches.
The Pythia model suite was designed to promote scientific
research on large language models, especially interpretability research.
Despite not centering downstream performance as a design goal, we find the
models <a href="#evaluations">match or exceed</a> the performance of
similar and same-sized models, such as those in the OPT and GPT-Neo suites.
<details>
<summary style="font-weight:600">Details on previous early release and naming convention.</summary>
Previously, we released an early version of the Pythia suite to the public.
However, we decided to retrain the model suite to address a few hyperparameter
discrepancies. This model card <a href="#changelog">lists the changes</a>;
see appendix B in the Pythia paper for further discussion. We found no
difference in benchmark performance between the two Pythia versions.
The old models are
[still available](https://huggingface.co/models?other=pythia_v0), but we
suggest the retrained suite if you are just starting to use Pythia.<br>
**This is the current release.**
Please note that all models in the *Pythia* suite were renamed in January
2023. For clarity, a <a href="#naming-convention-and-parameter-count">table
comparing the old and new names</a> is provided in this model card, together
with exact parameter counts.
</details>
<br>
# Pythia-2.8B-deduped
## Model Details
- Developed by: [EleutherAI](http://eleuther.ai)
- Model type: Transformer-based Language Model
- Language: English
- Learn more: [Pythia's GitHub repository](https://github.com/EleutherAI/pythia)
for training procedure, config files, and details on how to use.
[See paper](https://arxiv.org/pdf/2304.01373.pdf) for more evals and implementation
details.
- Library: [GPT-NeoX](https://github.com/EleutherAI/gpt-neox)
- License: Apache 2.0
- Contact: to ask questions about this model, join the [EleutherAI
Discord](https://discord.gg/zBGx3azzUn), and post them in `#release-discussion`.
Please read the existing *Pythia* documentation before asking about it in the
EleutherAI Discord. For general correspondence: [contact@eleuther.
ai](mailto:contact@eleuther.ai).
<figure>
| Pythia model | Non-Embedding Params | Layers | Model Dim | Heads | Batch Size | Learning Rate | Equivalent Models |
| -----------: | -------------------: | :----: | :-------: | :---: | :--------: | :-------------------: | :--------------------: |
| 70M | 18,915,328 | 6 | 512 | 8 | 2M | 1.0 x 10<sup>-3</sup> | — |
| 160M | 85,056,000 | 12 | 768 | 12 | 2M | 6.0 x 10<sup>-4</sup> | GPT-Neo 125M, OPT-125M |
| 410M | 302,311,424 | 24 | 1024 | 16 | 2M | 3.0 x 10<sup>-4</sup> | OPT-350M |
| 1.0B | 805,736,448 | 16 | 2048 | 8 | 2M | 3.0 x 10<sup>-4</sup> | — |
| 1.4B | 1,208,602,624 | 24 | 2048 | 16 | 2M | 2.0 x 10<sup>-4</sup> | GPT-Neo 1.3B, OPT-1.3B |
| 2.8B | 2,517,652,480 | 32 | 2560 | 32 | 2M | 1.6 x 10<sup>-4</sup> | GPT-Neo 2.7B, OPT-2.7B |
| 6.9B | 6,444,163,072 | 32 | 4096 | 32 | 2M | 1.2 x 10<sup>-4</sup> | OPT-6.7B |
| 12B | 11,327,027,200 | 36 | 5120 | 40 | 2M | 1.2 x 10<sup>-4</sup> | — |
<figcaption>Engineering details for the <i>Pythia Suite</i>. Deduped and
non-deduped models of a given size have the same hyperparameters. “Equivalent”
models have <b>exactly</b> the same architecture, and the same number of
non-embedding parameters.</figcaption>
</figure>
## Uses and Limitations
### Intended Use
The primary intended use of Pythia is research on the behavior, functionality,
and limitations of large language models. This suite is intended to provide
a controlled setting for performing scientific experiments. We also provide
154 checkpoints per model: initial `step0`, 10 log-spaced checkpoints
`step{1,2,4...512}`, and 143 evenly-spaced checkpoints from `step1000` to
`step143000`. These checkpoints are hosted on Hugging Face as branches. Note
that branch `143000` corresponds exactly to the model checkpoint on the `main`
branch of each model.
You may also further fine-tune and adapt Pythia-2.8B-deduped for deployment,
as long as your use is in accordance with the Apache 2.0 license. Pythia
models work with the Hugging Face [Transformers
Library](https://huggingface.co/docs/transformers/index). If you decide to use
pre-trained Pythia-2.8B-deduped as a basis for your fine-tuned model, please
conduct your own risk and bias assessment.
### Out-of-scope use
The Pythia Suite is **not** intended for deployment. It is not a in itself
a product and cannot be used for human-facing interactions. For example,
the model may generate harmful or offensive text. Please evaluate the risks
associated with your particular use case.
Pythia models are English-language only, and are not suitable for translation
or generating text in other languages.
Pythia-2.8B-deduped has not been fine-tuned for downstream contexts in which
language models are commonly deployed, such as writing genre prose,
or commercial chatbots. This means Pythia-2.8B-deduped will **not**
respond to a given prompt the way a product like ChatGPT does. This is because,
unlike this model, ChatGPT was fine-tuned using methods such as Reinforcement
Learning from Human Feedback (RLHF) to better “follow” human instructions.
### Limitations and biases
The core functionality of a large language model is to take a string of text
and predict the next token. The token used by the model need not produce the
most “accurate” text. Never rely on Pythia-2.8B-deduped to produce factually accurate
output.
This model was trained on [the Pile](https://pile.eleuther.ai/), a dataset
known to contain profanity and texts that are lewd or otherwise offensive.
See [Section 6 of the Pile paper](https://arxiv.org/abs/2101.00027) for a
discussion of documented biases with regards to gender, religion, and race.
Pythia-2.8B-deduped may produce socially unacceptable or undesirable text, *even if*
the prompt itself does not include anything explicitly offensive.
If you plan on using text generated through, for example, the Hosted Inference
API, we recommend having a human curate the outputs of this language model
before presenting it to other people. Please inform your audience that the
text was generated by Pythia-2.8B-deduped.
### Quickstart
Pythia models can be loaded and used via the following code, demonstrated here
for the third `pythia-70m-deduped` checkpoint:
```python
from transformers import GPTNeoXForCausalLM, AutoTokenizer
model = GPTNeoXForCausalLM.from_pretrained(
"EleutherAI/pythia-70m-deduped",
revision="step3000",
cache_dir="./pythia-70m-deduped/step3000",
)
tokenizer = AutoTokenizer.from_pretrained(
"EleutherAI/pythia-70m-deduped",
revision="step3000",
cache_dir="./pythia-70m-deduped/step3000",
)
inputs = tokenizer("Hello, I am", return_tensors="pt")
tokens = model.generate(**inputs)
tokenizer.decode(tokens[0])
```
Revision/branch `step143000` corresponds exactly to the model checkpoint on
the `main` branch of each model.<br>
For more information on how to use all Pythia models, see [documentation on
GitHub](https://github.com/EleutherAI/pythia).
## Training
### Training data
Pythia-2.8B-deduped was trained on the Pile **after the dataset has been globally
deduplicated**.<br>
[The Pile](https://pile.eleuther.ai/) is a 825GiB general-purpose dataset in
English. It was created by EleutherAI specifically for training large language
models. It contains texts from 22 diverse sources, roughly broken down into
five categories: academic writing (e.g. arXiv), internet (e.g. CommonCrawl),
prose (e.g. Project Gutenberg), dialogue (e.g. YouTube subtitles), and
miscellaneous (e.g. GitHub, Enron Emails). See [the Pile
paper](https://arxiv.org/abs/2101.00027) for a breakdown of all data sources,
methodology, and a discussion of ethical implications. Consult [the
datasheet](https://arxiv.org/abs/2201.07311) for more detailed documentation
about the Pile and its component datasets. The Pile can be downloaded from
the [official website](https://pile.eleuther.ai/), or from a [community
mirror](https://the-eye.eu/public/AI/pile/).
### Training procedure
All models were trained on the exact same data, in the exact same order. Each
model saw 299,892,736,000 tokens during training, and 143 checkpoints for each
model are saved every 2,097,152,000 tokens, spaced evenly throughout training,
from `step1000` to `step143000` (which is the same as `main`). In addition, we
also provide frequent early checkpoints: `step0` and `step{1,2,4...512}`.
This corresponds to training for just under 1 epoch on the Pile for
non-deduplicated models, and about 1.5 epochs on the deduplicated Pile.
All *Pythia* models trained for 143000 steps at a batch size
of 2M (2,097,152 tokens).<br>
See [GitHub](https://github.com/EleutherAI/pythia) for more details on training
procedure, including [how to reproduce
it](https://github.com/EleutherAI/pythia/blob/main/README.md#reproducing-training).<br>
Pythia uses the same tokenizer as [GPT-NeoX-
20B](https://huggingface.co/EleutherAI/gpt-neox-20b).
## Evaluations
All 16 *Pythia* models were evaluated using the [LM Evaluation
Harness](https://github.com/EleutherAI/lm-evaluation-harness). You can access
the results by model and step at `results/json/*` in the [GitHub
repository](https://github.com/EleutherAI/pythia/tree/main/results/json/).<br>
Expand the sections below to see plots of evaluation results for all
Pythia and Pythia-deduped models compared with OPT and BLOOM.
<details>
<summary>LAMBADA – OpenAI</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/lambada_openai_v1.png" style="width:auto"/>
</details>
<details>
<summary>Physical Interaction: Question Answering (PIQA)</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/piqa_v1.png" style="width:auto"/>
</details>
<details>
<summary>WinoGrande</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/winogrande_v1.png" style="width:auto"/>
</details>
<details>
<summary>AI2 Reasoning Challenge—Easy Set</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/arc_easy_v1.png" style="width:auto"/>
</details>
<details>
<summary>SciQ</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/sciq_v1.png" style="width:auto"/>
</details>
## Changelog
This section compares differences between previously released
[Pythia v0](https://huggingface.co/models?other=pythia_v0) and the current
models. See Appendix B of the Pythia paper for further discussion of these
changes and the motivation behind them. We found that retraining Pythia had no
impact on benchmark performance.
- All model sizes are now trained with uniform batch size of 2M tokens.
Previously, the models of size 160M, 410M, and 1.4B parameters were trained
with batch sizes of 4M tokens.
- We added checkpoints at initialization (step 0) and steps {1,2,4,8,16,32,64,
128,256,512} in addition to every 1000 training steps.
- Flash Attention was used in the new retrained suite.
- We remedied a minor inconsistency that existed in the original suite: all
models of size 2.8B parameters or smaller had a learning rate (LR) schedule
which decayed to a minimum LR of 10% the starting LR rate, but the 6.9B and
12B models all used an LR schedule which decayed to a minimum LR of 0. In
the redone training runs, we rectified this inconsistency: all models now were
trained with LR decaying to a minimum of 0.1× their maximum LR.
### Naming convention and parameter count
*Pythia* models were renamed in January 2023. It is possible that the old
naming convention still persists in some documentation by accident. The
current naming convention (70M, 160M, etc.) is based on total parameter count.
<figure style="width:32em">
| current Pythia suffix | old suffix | total params | non-embedding params |
| --------------------: | ---------: | -------------: | -------------------: |
| 70M | 19M | 70,426,624 | 18,915,328 |
| 160M | 125M | 162,322,944 | 85,056,000 |
| 410M | 350M | 405,334,016 | 302,311,424 |
| 1B | 800M | 1,011,781,632 | 805,736,448 |
| 1.4B | 1.3B | 1,414,647,808 | 1,208,602,624 |
| 2.8B | 2.7B | 2,775,208,960 | 2,517,652,480 |
| 6.9B | 6.7B | 6,857,302,016 | 6,444,163,072 |
| 12B | 13B | 11,846,072,320 | 11,327,027,200 |
</figure>
|
EleutherAI/pythia-160m-deduped
|
EleutherAI
| 2023-07-09T16:04:57Z | 43,492 | 3 |
transformers
|
[
"transformers",
"pytorch",
"safetensors",
"gpt_neox",
"text-generation",
"causal-lm",
"pythia",
"en",
"dataset:EleutherAI/the_pile_deduplicated",
"arxiv:2304.01373",
"arxiv:2101.00027",
"arxiv:2201.07311",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2023-02-08T21:50:19Z |
---
language:
- en
tags:
- pytorch
- causal-lm
- pythia
license: apache-2.0
datasets:
- EleutherAI/the_pile_deduplicated
---
The *Pythia Scaling Suite* is a collection of models developed to facilitate
interpretability research [(see paper)](https://arxiv.org/pdf/2304.01373.pdf).
It contains two sets of eight models of sizes
70M, 160M, 410M, 1B, 1.4B, 2.8B, 6.9B, and 12B. For each size, there are two
models: one trained on the Pile, and one trained on the Pile after the dataset
has been globally deduplicated. All 8 model sizes are trained on the exact
same data, in the exact same order. We also provide 154 intermediate
checkpoints per model, hosted on Hugging Face as branches.
The Pythia model suite was designed to promote scientific
research on large language models, especially interpretability research.
Despite not centering downstream performance as a design goal, we find the
models <a href="#evaluations">match or exceed</a> the performance of
similar and same-sized models, such as those in the OPT and GPT-Neo suites.
<details>
<summary style="font-weight:600">Details on previous early release and naming convention.</summary>
Previously, we released an early version of the Pythia suite to the public.
However, we decided to retrain the model suite to address a few hyperparameter
discrepancies. This model card <a href="#changelog">lists the changes</a>;
see appendix B in the Pythia paper for further discussion. We found no
difference in benchmark performance between the two Pythia versions.
The old models are
[still available](https://huggingface.co/models?other=pythia_v0), but we
suggest the retrained suite if you are just starting to use Pythia.<br>
**This is the current release.**
Please note that all models in the *Pythia* suite were renamed in January
2023. For clarity, a <a href="#naming-convention-and-parameter-count">table
comparing the old and new names</a> is provided in this model card, together
with exact parameter counts.
</details>
<br>
# Pythia-160M-deduped
## Model Details
- Developed by: [EleutherAI](http://eleuther.ai)
- Model type: Transformer-based Language Model
- Language: English
- Learn more: [Pythia's GitHub repository](https://github.com/EleutherAI/pythia)
for training procedure, config files, and details on how to use.
[See paper](https://arxiv.org/pdf/2304.01373.pdf) for more evals and implementation
details.
- Library: [GPT-NeoX](https://github.com/EleutherAI/gpt-neox)
- License: Apache 2.0
- Contact: to ask questions about this model, join the [EleutherAI
Discord](https://discord.gg/zBGx3azzUn), and post them in `#release-discussion`.
Please read the existing *Pythia* documentation before asking about it in the
EleutherAI Discord. For general correspondence: [contact@eleuther.
ai](mailto:contact@eleuther.ai).
<figure>
| Pythia model | Non-Embedding Params | Layers | Model Dim | Heads | Batch Size | Learning Rate | Equivalent Models |
| -----------: | -------------------: | :----: | :-------: | :---: | :--------: | :-------------------: | :--------------------: |
| 70M | 18,915,328 | 6 | 512 | 8 | 2M | 1.0 x 10<sup>-3</sup> | — |
| 160M | 85,056,000 | 12 | 768 | 12 | 2M | 6.0 x 10<sup>-4</sup> | GPT-Neo 125M, OPT-125M |
| 410M | 302,311,424 | 24 | 1024 | 16 | 2M | 3.0 x 10<sup>-4</sup> | OPT-350M |
| 1.0B | 805,736,448 | 16 | 2048 | 8 | 2M | 3.0 x 10<sup>-4</sup> | — |
| 1.4B | 1,208,602,624 | 24 | 2048 | 16 | 2M | 2.0 x 10<sup>-4</sup> | GPT-Neo 1.3B, OPT-1.3B |
| 2.8B | 2,517,652,480 | 32 | 2560 | 32 | 2M | 1.6 x 10<sup>-4</sup> | GPT-Neo 2.7B, OPT-2.7B |
| 6.9B | 6,444,163,072 | 32 | 4096 | 32 | 2M | 1.2 x 10<sup>-4</sup> | OPT-6.7B |
| 12B | 11,327,027,200 | 36 | 5120 | 40 | 2M | 1.2 x 10<sup>-4</sup> | — |
<figcaption>Engineering details for the <i>Pythia Suite</i>. Deduped and
non-deduped models of a given size have the same hyperparameters. “Equivalent”
models have <b>exactly</b> the same architecture, and the same number of
non-embedding parameters.</figcaption>
</figure>
## Uses and Limitations
### Intended Use
The primary intended use of Pythia is research on the behavior, functionality,
and limitations of large language models. This suite is intended to provide
a controlled setting for performing scientific experiments. We also provide
154 checkpoints per model: initial `step0`, 10 log-spaced checkpoints
`step{1,2,4...512}`, and 143 evenly-spaced checkpoints from `step1000` to
`step143000`. These checkpoints are hosted on Hugging Face as branches. Note
that branch `143000` corresponds exactly to the model checkpoint on the `main`
branch of each model.
You may also further fine-tune and adapt Pythia-160M-deduped for deployment,
as long as your use is in accordance with the Apache 2.0 license. Pythia
models work with the Hugging Face [Transformers
Library](https://huggingface.co/docs/transformers/index). If you decide to use
pre-trained Pythia-160M-deduped as a basis for your fine-tuned model, please
conduct your own risk and bias assessment.
### Out-of-scope use
The Pythia Suite is **not** intended for deployment. It is not a in itself
a product and cannot be used for human-facing interactions. For example,
the model may generate harmful or offensive text. Please evaluate the risks
associated with your particular use case.
Pythia models are English-language only, and are not suitable for translation
or generating text in other languages.
Pythia-160M-deduped has not been fine-tuned for downstream contexts in which
language models are commonly deployed, such as writing genre prose,
or commercial chatbots. This means Pythia-160M-deduped will **not**
respond to a given prompt the way a product like ChatGPT does. This is because,
unlike this model, ChatGPT was fine-tuned using methods such as Reinforcement
Learning from Human Feedback (RLHF) to better “follow” human instructions.
### Limitations and biases
The core functionality of a large language model is to take a string of text
and predict the next token. The token used by the model need not produce the
most “accurate” text. Never rely on Pythia-160M-deduped to produce factually accurate
output.
This model was trained on [the Pile](https://pile.eleuther.ai/), a dataset
known to contain profanity and texts that are lewd or otherwise offensive.
See [Section 6 of the Pile paper](https://arxiv.org/abs/2101.00027) for a
discussion of documented biases with regards to gender, religion, and race.
Pythia-160M-deduped may produce socially unacceptable or undesirable text, *even if*
the prompt itself does not include anything explicitly offensive.
If you plan on using text generated through, for example, the Hosted Inference
API, we recommend having a human curate the outputs of this language model
before presenting it to other people. Please inform your audience that the
text was generated by Pythia-160M-deduped.
### Quickstart
Pythia models can be loaded and used via the following code, demonstrated here
for the third `pythia-70m-deduped` checkpoint:
```python
from transformers import GPTNeoXForCausalLM, AutoTokenizer
model = GPTNeoXForCausalLM.from_pretrained(
"EleutherAI/pythia-70m-deduped",
revision="step3000",
cache_dir="./pythia-70m-deduped/step3000",
)
tokenizer = AutoTokenizer.from_pretrained(
"EleutherAI/pythia-70m-deduped",
revision="step3000",
cache_dir="./pythia-70m-deduped/step3000",
)
inputs = tokenizer("Hello, I am", return_tensors="pt")
tokens = model.generate(**inputs)
tokenizer.decode(tokens[0])
```
Revision/branch `step143000` corresponds exactly to the model checkpoint on
the `main` branch of each model.<br>
For more information on how to use all Pythia models, see [documentation on
GitHub](https://github.com/EleutherAI/pythia).
## Training
### Training data
Pythia-160M-deduped was trained on the Pile **after the dataset has been globally
deduplicated**.<br>
[The Pile](https://pile.eleuther.ai/) is a 825GiB general-purpose dataset in
English. It was created by EleutherAI specifically for training large language
models. It contains texts from 22 diverse sources, roughly broken down into
five categories: academic writing (e.g. arXiv), internet (e.g. CommonCrawl),
prose (e.g. Project Gutenberg), dialogue (e.g. YouTube subtitles), and
miscellaneous (e.g. GitHub, Enron Emails). See [the Pile
paper](https://arxiv.org/abs/2101.00027) for a breakdown of all data sources,
methodology, and a discussion of ethical implications. Consult [the
datasheet](https://arxiv.org/abs/2201.07311) for more detailed documentation
about the Pile and its component datasets. The Pile can be downloaded from
the [official website](https://pile.eleuther.ai/), or from a [community
mirror](https://the-eye.eu/public/AI/pile/).
### Training procedure
All models were trained on the exact same data, in the exact same order. Each
model saw 299,892,736,000 tokens during training, and 143 checkpoints for each
model are saved every 2,097,152,000 tokens, spaced evenly throughout training,
from `step1000` to `step143000` (which is the same as `main`). In addition, we
also provide frequent early checkpoints: `step0` and `step{1,2,4...512}`.
This corresponds to training for just under 1 epoch on the Pile for
non-deduplicated models, and about 1.5 epochs on the deduplicated Pile.
All *Pythia* models trained for 143000 steps at a batch size
of 2M (2,097,152 tokens).<br>
See [GitHub](https://github.com/EleutherAI/pythia) for more details on training
procedure, including [how to reproduce
it](https://github.com/EleutherAI/pythia/blob/main/README.md#reproducing-training).<br>
Pythia uses the same tokenizer as [GPT-NeoX-
20B](https://huggingface.co/EleutherAI/gpt-neox-20b).
## Evaluations
All 16 *Pythia* models were evaluated using the [LM Evaluation
Harness](https://github.com/EleutherAI/lm-evaluation-harness). You can access
the results by model and step at `results/json/*` in the [GitHub
repository](https://github.com/EleutherAI/pythia/tree/main/results/json/).<br>
Expand the sections below to see plots of evaluation results for all
Pythia and Pythia-deduped models compared with OPT and BLOOM.
<details>
<summary>LAMBADA – OpenAI</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/lambada_openai_v1.png" style="width:auto"/>
</details>
<details>
<summary>Physical Interaction: Question Answering (PIQA)</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/piqa_v1.png" style="width:auto"/>
</details>
<details>
<summary>WinoGrande</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/winogrande_v1.png" style="width:auto"/>
</details>
<details>
<summary>AI2 Reasoning Challenge—Easy Set</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/arc_easy_v1.png" style="width:auto"/>
</details>
<details>
<summary>SciQ</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/sciq_v1.png" style="width:auto"/>
</details>
## Changelog
This section compares differences between previously released
[Pythia v0](https://huggingface.co/models?other=pythia_v0) and the current
models. See Appendix B of the Pythia paper for further discussion of these
changes and the motivation behind them. We found that retraining Pythia had no
impact on benchmark performance.
- All model sizes are now trained with uniform batch size of 2M tokens.
Previously, the models of size 160M, 410M, and 1.4B parameters were trained
with batch sizes of 4M tokens.
- We added checkpoints at initialization (step 0) and steps {1,2,4,8,16,32,64,
128,256,512} in addition to every 1000 training steps.
- Flash Attention was used in the new retrained suite.
- We remedied a minor inconsistency that existed in the original suite: all
models of size 2.8B parameters or smaller had a learning rate (LR) schedule
which decayed to a minimum LR of 10% the starting LR rate, but the 6.9B and
12B models all used an LR schedule which decayed to a minimum LR of 0. In
the redone training runs, we rectified this inconsistency: all models now were
trained with LR decaying to a minimum of 0.1× their maximum LR.
### Naming convention and parameter count
*Pythia* models were renamed in January 2023. It is possible that the old
naming convention still persists in some documentation by accident. The
current naming convention (70M, 160M, etc.) is based on total parameter count.
<figure style="width:32em">
| current Pythia suffix | old suffix | total params | non-embedding params |
| --------------------: | ---------: | -------------: | -------------------: |
| 70M | 19M | 70,426,624 | 18,915,328 |
| 160M | 125M | 162,322,944 | 85,056,000 |
| 410M | 350M | 405,334,016 | 302,311,424 |
| 1B | 800M | 1,011,781,632 | 805,736,448 |
| 1.4B | 1.3B | 1,414,647,808 | 1,208,602,624 |
| 2.8B | 2.7B | 2,775,208,960 | 2,517,652,480 |
| 6.9B | 6.7B | 6,857,302,016 | 6,444,163,072 |
| 12B | 13B | 11,846,072,320 | 11,327,027,200 |
</figure>
|
EleutherAI/pythia-160m-v0
|
EleutherAI
| 2023-07-09T16:03:26Z | 11,182 | 8 |
transformers
|
[
"transformers",
"pytorch",
"safetensors",
"gpt_neox",
"text-generation",
"causal-lm",
"pythia",
"pythia_v0",
"en",
"dataset:the_pile",
"arxiv:2101.00027",
"arxiv:2201.07311",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2022-10-16T17:40:11Z |
---
language:
- en
tags:
- pytorch
- causal-lm
- pythia
- pythia_v0
license: apache-2.0
datasets:
- the_pile
---
The *Pythia Scaling Suite* is a collection of models developed to facilitate
interpretability research. It contains two sets of eight models of sizes
70M, 160M, 410M, 1B, 1.4B, 2.8B, 6.9B, and 12B. For each size, there are two
models: one trained on the Pile, and one trained on the Pile after the dataset
has been globally deduplicated. All 8 model sizes are trained on the exact
same data, in the exact same order. All Pythia models are available
[on Hugging Face](https://huggingface.co/models?other=pythia).
The Pythia model suite was deliberately designed to promote scientific
research on large language models, especially interpretability research.
Despite not centering downstream performance as a design goal, we find the
models <a href="#evaluations">match or exceed</a> the performance of
similar and same-sized models, such as those in the OPT and GPT-Neo suites.
Please note that all models in the *Pythia* suite were renamed in January
2023. For clarity, a <a href="#naming-convention-and-parameter-count">table
comparing the old and new names</a> is provided in this model card, together
with exact parameter counts.
## Pythia-160M
### Model Details
- Developed by: [EleutherAI](http://eleuther.ai)
- Model type: Transformer-based Language Model
- Language: English
- Learn more: [Pythia's GitHub repository](https://github.com/EleutherAI/pythia)
for training procedure, config files, and details on how to use.
- Library: [GPT-NeoX](https://github.com/EleutherAI/gpt-neox)
- License: Apache 2.0
- Contact: to ask questions about this model, join the [EleutherAI
Discord](https://discord.gg/zBGx3azzUn), and post them in `#release-discussion`.
Please read the existing *Pythia* documentation before asking about it in the
EleutherAI Discord. For general correspondence: [contact@eleuther.
ai](mailto:contact@eleuther.ai).
<figure>
| Pythia model | Non-Embedding Params | Layers | Model Dim | Heads | Batch Size | Learning Rate | Equivalent Models |
| -----------: | -------------------: | :----: | :-------: | :---: | :--------: | :-------------------: | :--------------------: |
| 70M | 18,915,328 | 6 | 512 | 8 | 2M | 1.0 x 10<sup>-3</sup> | — |
| 160M | 85,056,000 | 12 | 768 | 12 | 4M | 6.0 x 10<sup>-4</sup> | GPT-Neo 125M, OPT-125M |
| 410M | 302,311,424 | 24 | 1024 | 16 | 4M | 3.0 x 10<sup>-4</sup> | OPT-350M |
| 1.0B | 805,736,448 | 16 | 2048 | 8 | 2M | 3.0 x 10<sup>-4</sup> | — |
| 1.4B | 1,208,602,624 | 24 | 2048 | 16 | 4M | 2.0 x 10<sup>-4</sup> | GPT-Neo 1.3B, OPT-1.3B |
| 2.8B | 2,517,652,480 | 32 | 2560 | 32 | 2M | 1.6 x 10<sup>-4</sup> | GPT-Neo 2.7B, OPT-2.7B |
| 6.9B | 6,444,163,072 | 32 | 4096 | 32 | 2M | 1.2 x 10<sup>-4</sup> | OPT-6.7B |
| 12B | 11,327,027,200 | 36 | 5120 | 40 | 2M | 1.2 x 10<sup>-4</sup> | — |
<figcaption>Engineering details for the <i>Pythia Suite</i>. Deduped and
non-deduped models of a given size have the same hyperparameters. “Equivalent”
models have <b>exactly</b> the same architecture, and the same number of
non-embedding parameters.</figcaption>
</figure>
### Uses and Limitations
#### Intended Use
The primary intended use of Pythia is research on the behavior, functionality,
and limitations of large language models. This suite is intended to provide
a controlled setting for performing scientific experiments. To enable the
study of how language models change over the course of training, we provide
143 evenly spaced intermediate checkpoints per model. These checkpoints are
hosted on Hugging Face as branches. Note that branch `143000` corresponds
exactly to the model checkpoint on the `main` branch of each model.
You may also further fine-tune and adapt Pythia-160M for deployment,
as long as your use is in accordance with the Apache 2.0 license. Pythia
models work with the Hugging Face [Transformers
Library](https://huggingface.co/docs/transformers/index). If you decide to use
pre-trained Pythia-160M as a basis for your fine-tuned model, please
conduct your own risk and bias assessment.
#### Out-of-scope use
The Pythia Suite is **not** intended for deployment. It is not a in itself
a product and cannot be used for human-facing interactions.
Pythia models are English-language only, and are not suitable for translation
or generating text in other languages.
Pythia-160M has not been fine-tuned for downstream contexts in which
language models are commonly deployed, such as writing genre prose,
or commercial chatbots. This means Pythia-160M will **not**
respond to a given prompt the way a product like ChatGPT does. This is because,
unlike this model, ChatGPT was fine-tuned using methods such as Reinforcement
Learning from Human Feedback (RLHF) to better “understand” human instructions.
#### Limitations and biases
The core functionality of a large language model is to take a string of text
and predict the next token. The token deemed statistically most likely by the
model need not produce the most “accurate” text. Never rely on
Pythia-160M to produce factually accurate output.
This model was trained on [the Pile](https://pile.eleuther.ai/), a dataset
known to contain profanity and texts that are lewd or otherwise offensive.
See [Section 6 of the Pile paper](https://arxiv.org/abs/2101.00027) for a
discussion of documented biases with regards to gender, religion, and race.
Pythia-160M may produce socially unacceptable or undesirable text, *even if*
the prompt itself does not include anything explicitly offensive.
If you plan on using text generated through, for example, the Hosted Inference
API, we recommend having a human curate the outputs of this language model
before presenting it to other people. Please inform your audience that the
text was generated by Pythia-160M.
### Quickstart
Pythia models can be loaded and used via the following code, demonstrated here
for the third `pythia-70m-deduped` checkpoint:
```python
from transformers import GPTNeoXForCausalLM, AutoTokenizer
model = GPTNeoXForCausalLM.from_pretrained(
"EleutherAI/pythia-70m-deduped",
revision="step3000",
cache_dir="./pythia-70m-deduped/step3000",
)
tokenizer = AutoTokenizer.from_pretrained(
"EleutherAI/pythia-70m-deduped",
revision="step3000",
cache_dir="./pythia-70m-deduped/step3000",
)
inputs = tokenizer("Hello, I am", return_tensors="pt")
tokens = model.generate(**inputs)
tokenizer.decode(tokens[0])
```
Revision/branch `step143000` corresponds exactly to the model checkpoint on
the `main` branch of each model.<br>
For more information on how to use all Pythia models, see [documentation on
GitHub](https://github.com/EleutherAI/pythia).
### Training
#### Training data
[The Pile](https://pile.eleuther.ai/) is a 825GiB general-purpose dataset in
English. It was created by EleutherAI specifically for training large language
models. It contains texts from 22 diverse sources, roughly broken down into
five categories: academic writing (e.g. arXiv), internet (e.g. CommonCrawl),
prose (e.g. Project Gutenberg), dialogue (e.g. YouTube subtitles), and
miscellaneous (e.g. GitHub, Enron Emails). See [the Pile
paper](https://arxiv.org/abs/2101.00027) for a breakdown of all data sources,
methodology, and a discussion of ethical implications. Consult [the
datasheet](https://arxiv.org/abs/2201.07311) for more detailed documentation
about the Pile and its component datasets. The Pile can be downloaded from
the [official website](https://pile.eleuther.ai/), or from a [community
mirror](https://the-eye.eu/public/AI/pile/).<br>
The Pile was **not** deduplicated before being used to train Pythia-160M.
#### Training procedure
All models were trained on the exact same data, in the exact same order. Each
model saw 299,892,736,000 tokens during training, and 143 checkpoints for each
model are saved every 2,097,152,000 tokens, spaced evenly throughout training.
This corresponds to training for just under 1 epoch on the Pile for
non-deduplicated models, and about 1.5 epochs on the deduplicated Pile.
All *Pythia* models trained for the equivalent of 143000 steps at a batch size
of 2,097,152 tokens. Two batch sizes were used: 2M and 4M. Models with a batch
size of 4M tokens listed were originally trained for 71500 steps instead, with
checkpoints every 500 steps. The checkpoints on Hugging Face are renamed for
consistency with all 2M batch models, so `step1000` is the first checkpoint
for `pythia-1.4b` that was saved (corresponding to step 500 in training), and
`step1000` is likewise the first `pythia-6.9b` checkpoint that was saved
(corresponding to 1000 “actual” steps).<br>
See [GitHub](https://github.com/EleutherAI/pythia) for more details on training
procedure, including [how to reproduce
it](https://github.com/EleutherAI/pythia/blob/main/README.md#reproducing-training).<br>
Pythia uses the same tokenizer as [GPT-NeoX-
20B](https://huggingface.co/EleutherAI/gpt-neox-20b).
### Evaluations
All 16 *Pythia* models were evaluated using the [LM Evaluation
Harness](https://github.com/EleutherAI/lm-evaluation-harness). You can access
the results by model and step at `results/json/*` in the [GitHub
repository](https://github.com/EleutherAI/pythia/tree/main/results/json).<br>
Expand the sections below to see plots of evaluation results for all
Pythia and Pythia-deduped models compared with OPT and BLOOM.
<details>
<summary>LAMBADA – OpenAI</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/lambada_openai.png" style="width:auto"/>
</details>
<details>
<summary>Physical Interaction: Question Answering (PIQA)</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/piqa.png" style="width:auto"/>
</details>
<details>
<summary>WinoGrande</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/winogrande.png" style="width:auto"/>
</details>
<details>
<summary>AI2 Reasoning Challenge—Challenge Set</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/arc_challenge.png" style="width:auto"/>
</details>
<details>
<summary>SciQ</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/sciq.png" style="width:auto"/>
</details>
### Naming convention and parameter count
*Pythia* models were renamed in January 2023. It is possible that the old
naming convention still persists in some documentation by accident. The
current naming convention (70M, 160M, etc.) is based on total parameter count.
<figure style="width:32em">
| current Pythia suffix | old suffix | total params | non-embedding params |
| --------------------: | ---------: | -------------: | -------------------: |
| 70M | 19M | 70,426,624 | 18,915,328 |
| 160M | 125M | 162,322,944 | 85,056,000 |
| 410M | 350M | 405,334,016 | 302,311,424 |
| 1B | 800M | 1,011,781,632 | 805,736,448 |
| 1.4B | 1.3B | 1,414,647,808 | 1,208,602,624 |
| 2.8B | 2.7B | 2,775,208,960 | 2,517,652,480 |
| 6.9B | 6.7B | 6,857,302,016 | 6,444,163,072 |
| 12B | 13B | 11,846,072,320 | 11,327,027,200 |
</figure>
|
EleutherAI/pythia-1.4b
|
EleutherAI
| 2023-07-09T16:01:57Z | 25,256 | 22 |
transformers
|
[
"transformers",
"pytorch",
"safetensors",
"gpt_neox",
"text-generation",
"causal-lm",
"pythia",
"en",
"dataset:EleutherAI/the_pile",
"arxiv:2304.01373",
"arxiv:2101.00027",
"arxiv:2201.07311",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2023-02-09T14:08:20Z |
---
language:
- en
tags:
- pytorch
- causal-lm
- pythia
license: apache-2.0
datasets:
- EleutherAI/the_pile
---
The *Pythia Scaling Suite* is a collection of models developed to facilitate
interpretability research [(see paper)](https://arxiv.org/pdf/2304.01373.pdf).
It contains two sets of eight models of sizes
70M, 160M, 410M, 1B, 1.4B, 2.8B, 6.9B, and 12B. For each size, there are two
models: one trained on the Pile, and one trained on the Pile after the dataset
has been globally deduplicated. All 8 model sizes are trained on the exact
same data, in the exact same order. We also provide 154 intermediate
checkpoints per model, hosted on Hugging Face as branches.
The Pythia model suite was deliberately designed to promote scientific
research on large language models, especially interpretability research.
Despite not centering downstream performance as a design goal, we find the
models <a href="#evaluations">match or exceed</a> the performance of
similar and same-sized models, such as those in the OPT and GPT-Neo suites.
<details>
<summary style="font-weight:600">Details on previous early release and naming convention.</summary>
Previously, we released an early version of the Pythia suite to the public.
However, we decided to retrain the model suite to address a few hyperparameter
discrepancies. This model card <a href="#changelog">lists the changes</a>;
see appendix B in the Pythia paper for further discussion. We found no
difference in benchmark performance between the two Pythia versions.
The old models are
[still available](https://huggingface.co/models?other=pythia_v0), but we
suggest the retrained suite if you are just starting to use Pythia.<br>
**This is the current release.**
Please note that all models in the *Pythia* suite were renamed in January
2023. For clarity, a <a href="#naming-convention-and-parameter-count">table
comparing the old and new names</a> is provided in this model card, together
with exact parameter counts.
</details>
<br>
# Pythia-1.4B
## Model Details
- Developed by: [EleutherAI](http://eleuther.ai)
- Model type: Transformer-based Language Model
- Language: English
- Learn more: [Pythia's GitHub repository](https://github.com/EleutherAI/pythia)
for training procedure, config files, and details on how to use.
[See paper](https://arxiv.org/pdf/2304.01373.pdf) for more evals and implementation
details.
- Library: [GPT-NeoX](https://github.com/EleutherAI/gpt-neox)
- License: Apache 2.0
- Contact: to ask questions about this model, join the [EleutherAI
Discord](https://discord.gg/zBGx3azzUn), and post them in `#release-discussion`.
Please read the existing *Pythia* documentation before asking about it in the
EleutherAI Discord. For general correspondence: [contact@eleuther.
ai](mailto:contact@eleuther.ai).
<figure>
| Pythia model | Non-Embedding Params | Layers | Model Dim | Heads | Batch Size | Learning Rate | Equivalent Models |
| -----------: | -------------------: | :----: | :-------: | :---: | :--------: | :-------------------: | :--------------------: |
| 70M | 18,915,328 | 6 | 512 | 8 | 2M | 1.0 x 10<sup>-3</sup> | — |
| 160M | 85,056,000 | 12 | 768 | 12 | 2M | 6.0 x 10<sup>-4</sup> | GPT-Neo 125M, OPT-125M |
| 410M | 302,311,424 | 24 | 1024 | 16 | 2M | 3.0 x 10<sup>-4</sup> | OPT-350M |
| 1.0B | 805,736,448 | 16 | 2048 | 8 | 2M | 3.0 x 10<sup>-4</sup> | — |
| 1.4B | 1,208,602,624 | 24 | 2048 | 16 | 2M | 2.0 x 10<sup>-4</sup> | GPT-Neo 1.3B, OPT-1.3B |
| 2.8B | 2,517,652,480 | 32 | 2560 | 32 | 2M | 1.6 x 10<sup>-4</sup> | GPT-Neo 2.7B, OPT-2.7B |
| 6.9B | 6,444,163,072 | 32 | 4096 | 32 | 2M | 1.2 x 10<sup>-4</sup> | OPT-6.7B |
| 12B | 11,327,027,200 | 36 | 5120 | 40 | 2M | 1.2 x 10<sup>-4</sup> | — |
<figcaption>Engineering details for the <i>Pythia Suite</i>. Deduped and
non-deduped models of a given size have the same hyperparameters. “Equivalent”
models have <b>exactly</b> the same architecture, and the same number of
non-embedding parameters.</figcaption>
</figure>
## Uses and Limitations
### Intended Use
The primary intended use of Pythia is research on the behavior, functionality,
and limitations of large language models. This suite is intended to provide
a controlled setting for performing scientific experiments. We also provide
154 checkpoints per model: initial `step0`, 10 log-spaced checkpoints
`step{1,2,4...512}`, and 143 evenly-spaced checkpoints from `step1000` to
`step143000`. These checkpoints are hosted on Hugging Face as branches. Note
that branch `143000` corresponds exactly to the model checkpoint on the `main`
branch of each model.
You may also further fine-tune and adapt Pythia-1.4B for deployment,
as long as your use is in accordance with the Apache 2.0 license. Pythia
models work with the Hugging Face [Transformers
Library](https://huggingface.co/docs/transformers/index). If you decide to use
pre-trained Pythia-1.4B as a basis for your fine-tuned model, please
conduct your own risk and bias assessment.
### Out-of-scope use
The Pythia Suite is **not** intended for deployment. It is not a in itself
a product and cannot be used for human-facing interactions. For example,
the model may generate harmful or offensive text. Please evaluate the risks
associated with your particular use case.
Pythia models are English-language only, and are not suitable for translation
or generating text in other languages.
Pythia-1.4B has not been fine-tuned for downstream contexts in which
language models are commonly deployed, such as writing genre prose,
or commercial chatbots. This means Pythia-1.4B will **not**
respond to a given prompt the way a product like ChatGPT does. This is because,
unlike this model, ChatGPT was fine-tuned using methods such as Reinforcement
Learning from Human Feedback (RLHF) to better “follow” human instructions.
### Limitations and biases
The core functionality of a large language model is to take a string of text
and predict the next token. The token used by the model need not produce the
most “accurate” text. Never rely on Pythia-1.4B to produce factually accurate
output.
This model was trained on [the Pile](https://pile.eleuther.ai/), a dataset
known to contain profanity and texts that are lewd or otherwise offensive.
See [Section 6 of the Pile paper](https://arxiv.org/abs/2101.00027) for a
discussion of documented biases with regards to gender, religion, and race.
Pythia-1.4B may produce socially unacceptable or undesirable text, *even if*
the prompt itself does not include anything explicitly offensive.
If you plan on using text generated through, for example, the Hosted Inference
API, we recommend having a human curate the outputs of this language model
before presenting it to other people. Please inform your audience that the
text was generated by Pythia-1.4B.
### Quickstart
Pythia models can be loaded and used via the following code, demonstrated here
for the third `pythia-70m-deduped` checkpoint:
```python
from transformers import GPTNeoXForCausalLM, AutoTokenizer
model = GPTNeoXForCausalLM.from_pretrained(
"EleutherAI/pythia-70m-deduped",
revision="step3000",
cache_dir="./pythia-70m-deduped/step3000",
)
tokenizer = AutoTokenizer.from_pretrained(
"EleutherAI/pythia-70m-deduped",
revision="step3000",
cache_dir="./pythia-70m-deduped/step3000",
)
inputs = tokenizer("Hello, I am", return_tensors="pt")
tokens = model.generate(**inputs)
tokenizer.decode(tokens[0])
```
Revision/branch `step143000` corresponds exactly to the model checkpoint on
the `main` branch of each model.<br>
For more information on how to use all Pythia models, see [documentation on
GitHub](https://github.com/EleutherAI/pythia).
## Training
### Training data
[The Pile](https://pile.eleuther.ai/) is a 825GiB general-purpose dataset in
English. It was created by EleutherAI specifically for training large language
models. It contains texts from 22 diverse sources, roughly broken down into
five categories: academic writing (e.g. arXiv), internet (e.g. CommonCrawl),
prose (e.g. Project Gutenberg), dialogue (e.g. YouTube subtitles), and
miscellaneous (e.g. GitHub, Enron Emails). See [the Pile
paper](https://arxiv.org/abs/2101.00027) for a breakdown of all data sources,
methodology, and a discussion of ethical implications. Consult [the
datasheet](https://arxiv.org/abs/2201.07311) for more detailed documentation
about the Pile and its component datasets. The Pile can be downloaded from
the [official website](https://pile.eleuther.ai/), or from a [community
mirror](https://the-eye.eu/public/AI/pile/).<br>
The Pile was **not** deduplicated before being used to train Pythia-1.4B.
### Training procedure
All models were trained on the exact same data, in the exact same order. Each
model saw 299,892,736,000 tokens during training, and 143 checkpoints for each
model are saved every 2,097,152,000 tokens, spaced evenly throughout training,
from `step1000` to `step143000` (which is the same as `main`). In addition, we
also provide frequent early checkpoints: `step0` and `step{1,2,4...512}`.
This corresponds to training for just under 1 epoch on the Pile for
non-deduplicated models, and about 1.5 epochs on the deduplicated Pile.
All *Pythia* models trained for 143000 steps at a batch size
of 2M (2,097,152 tokens).<br>
See [GitHub](https://github.com/EleutherAI/pythia) for more details on training
procedure, including [how to reproduce
it](https://github.com/EleutherAI/pythia/blob/main/README.md#reproducing-training).<br>
Pythia uses the same tokenizer as [GPT-NeoX-
20B](https://huggingface.co/EleutherAI/gpt-neox-20b).
## Evaluations
All 16 *Pythia* models were evaluated using the [LM Evaluation
Harness](https://github.com/EleutherAI/lm-evaluation-harness). You can access
the results by model and step at `results/json/*` in the [GitHub
repository](https://github.com/EleutherAI/pythia/tree/main/results/json/).<br>
Expand the sections below to see plots of evaluation results for all
Pythia and Pythia-deduped models compared with OPT and BLOOM.
<details>
<summary>LAMBADA – OpenAI</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/lambada_openai_v1.png" style="width:auto"/>
</details>
<details>
<summary>Physical Interaction: Question Answering (PIQA)</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/piqa_v1.png" style="width:auto"/>
</details>
<details>
<summary>WinoGrande</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/winogrande_v1.png" style="width:auto"/>
</details>
<details>
<summary>AI2 Reasoning Challenge—Easy Set</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/arc_easy_v1.png" style="width:auto"/>
</details>
<details>
<summary>SciQ</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/sciq_v1.png" style="width:auto"/>
</details>
## Changelog
This section compares differences between previously released
[Pythia v0](https://huggingface.co/models?other=pythia_v0) and the current
models. See Appendix B of the Pythia paper for further discussion of these
changes and the motivation behind them. We found that retraining Pythia had no
impact on benchmark performance.
- All model sizes are now trained with uniform batch size of 2M tokens.
Previously, the models of size 160M, 410M, and 1.4B parameters were trained
with batch sizes of 4M tokens.
- We added checkpoints at initialization (step 0) and steps {1,2,4,8,16,32,64,
128,256,512} in addition to every 1000 training steps.
- Flash Attention was used in the new retrained suite.
- We remedied a minor inconsistency that existed in the original suite: all
models of size 2.8B parameters or smaller had a learning rate (LR) schedule
which decayed to a minimum LR of 10% the starting LR rate, but the 6.9B and
12B models all used an LR schedule which decayed to a minimum LR of 0. In
the redone training runs, we rectified this inconsistency: all models now were
trained with LR decaying to a minimum of 0.1× their maximum LR.
### Naming convention and parameter count
*Pythia* models were renamed in January 2023. It is possible that the old
naming convention still persists in some documentation by accident. The
current naming convention (70M, 160M, etc.) is based on total parameter count.
<figure style="width:32em">
| current Pythia suffix | old suffix | total params | non-embedding params |
| --------------------: | ---------: | -------------: | -------------------: |
| 70M | 19M | 70,426,624 | 18,915,328 |
| 160M | 125M | 162,322,944 | 85,056,000 |
| 410M | 350M | 405,334,016 | 302,311,424 |
| 1B | 800M | 1,011,781,632 | 805,736,448 |
| 1.4B | 1.3B | 1,414,647,808 | 1,208,602,624 |
| 2.8B | 2.7B | 2,775,208,960 | 2,517,652,480 |
| 6.9B | 6.7B | 6,857,302,016 | 6,444,163,072 |
| 12B | 13B | 11,846,072,320 | 11,327,027,200 |
</figure>
|
EleutherAI/pythia-410m
|
EleutherAI
| 2023-07-09T16:01:42Z | 68,125 | 22 |
transformers
|
[
"transformers",
"pytorch",
"safetensors",
"gpt_neox",
"text-generation",
"causal-lm",
"pythia",
"en",
"dataset:EleutherAI/pile",
"arxiv:2304.01373",
"arxiv:2101.00027",
"arxiv:2201.07311",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2023-02-13T18:45:00Z |
---
language:
- en
tags:
- pytorch
- causal-lm
- pythia
license: apache-2.0
datasets:
- EleutherAI/pile
---
The *Pythia Scaling Suite* is a collection of models developed to facilitate
interpretability research [(see paper)](https://arxiv.org/pdf/2304.01373.pdf).
It contains two sets of eight models of sizes
70M, 160M, 410M, 1B, 1.4B, 2.8B, 6.9B, and 12B. For each size, there are two
models: one trained on the Pile, and one trained on the Pile after the dataset
has been globally deduplicated. All 8 model sizes are trained on the exact
same data, in the exact same order. We also provide 154 intermediate
checkpoints per model, hosted on Hugging Face as branches.
The Pythia model suite was deliberately designed to promote scientific
research on large language models, especially interpretability research.
Despite not centering downstream performance as a design goal, we find the
models <a href="#evaluations">match or exceed</a> the performance of
similar and same-sized models, such as those in the OPT and GPT-Neo suites.
<details>
<summary style="font-weight:600">Details on previous early release and naming convention.</summary>
Previously, we released an early version of the Pythia suite to the public.
However, we decided to retrain the model suite to address a few hyperparameter
discrepancies. This model card <a href="#changelog">lists the changes</a>;
see appendix B in the Pythia paper for further discussion. We found no
difference in benchmark performance between the two Pythia versions.
The old models are
[still available](https://huggingface.co/models?other=pythia_v0), but we
suggest the retrained suite if you are just starting to use Pythia.<br>
**This is the current release.**
Please note that all models in the *Pythia* suite were renamed in January
2023. For clarity, a <a href="#naming-convention-and-parameter-count">table
comparing the old and new names</a> is provided in this model card, together
with exact parameter counts.
</details>
<br>
# Pythia-410M
## Model Details
- Developed by: [EleutherAI](http://eleuther.ai)
- Model type: Transformer-based Language Model
- Language: English
- Learn more: [Pythia's GitHub repository](https://github.com/EleutherAI/pythia)
for training procedure, config files, and details on how to use.
[See paper](https://arxiv.org/pdf/2304.01373.pdf) for more evals and implementation
details.
- Library: [GPT-NeoX](https://github.com/EleutherAI/gpt-neox)
- License: Apache 2.0
- Contact: to ask questions about this model, join the [EleutherAI
Discord](https://discord.gg/zBGx3azzUn), and post them in `#release-discussion`.
Please read the existing *Pythia* documentation before asking about it in the
EleutherAI Discord. For general correspondence: [contact@eleuther.
ai](mailto:contact@eleuther.ai).
<figure>
| Pythia model | Non-Embedding Params | Layers | Model Dim | Heads | Batch Size | Learning Rate | Equivalent Models |
| -----------: | -------------------: | :----: | :-------: | :---: | :--------: | :-------------------: | :--------------------: |
| 70M | 18,915,328 | 6 | 512 | 8 | 2M | 1.0 x 10<sup>-3</sup> | — |
| 160M | 85,056,000 | 12 | 768 | 12 | 2M | 6.0 x 10<sup>-4</sup> | GPT-Neo 125M, OPT-125M |
| 410M | 302,311,424 | 24 | 1024 | 16 | 2M | 3.0 x 10<sup>-4</sup> | OPT-350M |
| 1.0B | 805,736,448 | 16 | 2048 | 8 | 2M | 3.0 x 10<sup>-4</sup> | — |
| 1.4B | 1,208,602,624 | 24 | 2048 | 16 | 2M | 2.0 x 10<sup>-4</sup> | GPT-Neo 1.3B, OPT-1.3B |
| 2.8B | 2,517,652,480 | 32 | 2560 | 32 | 2M | 1.6 x 10<sup>-4</sup> | GPT-Neo 2.7B, OPT-2.7B |
| 6.9B | 6,444,163,072 | 32 | 4096 | 32 | 2M | 1.2 x 10<sup>-4</sup> | OPT-6.7B |
| 12B | 11,327,027,200 | 36 | 5120 | 40 | 2M | 1.2 x 10<sup>-4</sup> | — |
<figcaption>Engineering details for the <i>Pythia Suite</i>. Deduped and
non-deduped models of a given size have the same hyperparameters. “Equivalent”
models have <b>exactly</b> the same architecture, and the same number of
non-embedding parameters.</figcaption>
</figure>
## Uses and Limitations
### Intended Use
The primary intended use of Pythia is research on the behavior, functionality,
and limitations of large language models. This suite is intended to provide
a controlled setting for performing scientific experiments. We also provide
154 checkpoints per model: initial `step0`, 10 log-spaced checkpoints
`step{1,2,4...512}`, and 143 evenly-spaced checkpoints from `step1000` to
`step143000`. These checkpoints are hosted on Hugging Face as branches. Note
that branch `143000` corresponds exactly to the model checkpoint on the `main`
branch of each model.
You may also further fine-tune and adapt Pythia-410M for deployment,
as long as your use is in accordance with the Apache 2.0 license. Pythia
models work with the Hugging Face [Transformers
Library](https://huggingface.co/docs/transformers/index). If you decide to use
pre-trained Pythia-410M as a basis for your fine-tuned model, please
conduct your own risk and bias assessment.
### Out-of-scope use
The Pythia Suite is **not** intended for deployment. It is not a in itself
a product and cannot be used for human-facing interactions. For example,
the model may generate harmful or offensive text. Please evaluate the risks
associated with your particular use case.
Pythia models are English-language only, and are not suitable for translation
or generating text in other languages.
Pythia-410M has not been fine-tuned for downstream contexts in which
language models are commonly deployed, such as writing genre prose,
or commercial chatbots. This means Pythia-410M will **not**
respond to a given prompt the way a product like ChatGPT does. This is because,
unlike this model, ChatGPT was fine-tuned using methods such as Reinforcement
Learning from Human Feedback (RLHF) to better “follow” human instructions.
### Limitations and biases
The core functionality of a large language model is to take a string of text
and predict the next token. The token used by the model need not produce the
most “accurate” text. Never rely on Pythia-410M to produce factually accurate
output.
This model was trained on [the Pile](https://pile.eleuther.ai/), a dataset
known to contain profanity and texts that are lewd or otherwise offensive.
See [Section 6 of the Pile paper](https://arxiv.org/abs/2101.00027) for a
discussion of documented biases with regards to gender, religion, and race.
Pythia-410M may produce socially unacceptable or undesirable text, *even if*
the prompt itself does not include anything explicitly offensive.
If you plan on using text generated through, for example, the Hosted Inference
API, we recommend having a human curate the outputs of this language model
before presenting it to other people. Please inform your audience that the
text was generated by Pythia-410M.
### Quickstart
Pythia models can be loaded and used via the following code, demonstrated here
for the third `pythia-70m-deduped` checkpoint:
```python
from transformers import GPTNeoXForCausalLM, AutoTokenizer
model = GPTNeoXForCausalLM.from_pretrained(
"EleutherAI/pythia-70m-deduped",
revision="step3000",
cache_dir="./pythia-70m-deduped/step3000",
)
tokenizer = AutoTokenizer.from_pretrained(
"EleutherAI/pythia-70m-deduped",
revision="step3000",
cache_dir="./pythia-70m-deduped/step3000",
)
inputs = tokenizer("Hello, I am", return_tensors="pt")
tokens = model.generate(**inputs)
tokenizer.decode(tokens[0])
```
Revision/branch `step143000` corresponds exactly to the model checkpoint on
the `main` branch of each model.<br>
For more information on how to use all Pythia models, see [documentation on
GitHub](https://github.com/EleutherAI/pythia).
## Training
### Training data
[The Pile](https://pile.eleuther.ai/) is a 825GiB general-purpose dataset in
English. It was created by EleutherAI specifically for training large language
models. It contains texts from 22 diverse sources, roughly broken down into
five categories: academic writing (e.g. arXiv), internet (e.g. CommonCrawl),
prose (e.g. Project Gutenberg), dialogue (e.g. YouTube subtitles), and
miscellaneous (e.g. GitHub, Enron Emails). See [the Pile
paper](https://arxiv.org/abs/2101.00027) for a breakdown of all data sources,
methodology, and a discussion of ethical implications. Consult [the
datasheet](https://arxiv.org/abs/2201.07311) for more detailed documentation
about the Pile and its component datasets. The Pile can be downloaded from
the [official website](https://pile.eleuther.ai/), or from a [community
mirror](https://the-eye.eu/public/AI/pile/).<br>
The Pile was **not** deduplicated before being used to train Pythia-410M.
### Training procedure
All models were trained on the exact same data, in the exact same order. Each
model saw 299,892,736,000 tokens during training, and 143 checkpoints for each
model are saved every 2,097,152,000 tokens, spaced evenly throughout training,
from `step1000` to `step143000` (which is the same as `main`). In addition, we
also provide frequent early checkpoints: `step0` and `step{1,2,4...512}`.
This corresponds to training for just under 1 epoch on the Pile for
non-deduplicated models, and about 1.5 epochs on the deduplicated Pile.
All *Pythia* models trained for 143000 steps at a batch size
of 2M (2,097,152 tokens).<br>
See [GitHub](https://github.com/EleutherAI/pythia) for more details on training
procedure, including [how to reproduce
it](https://github.com/EleutherAI/pythia/blob/main/README.md#reproducing-training).<br>
Pythia uses the same tokenizer as [GPT-NeoX-
20B](https://huggingface.co/EleutherAI/gpt-neox-20b).
## Evaluations
All 16 *Pythia* models were evaluated using the [LM Evaluation
Harness](https://github.com/EleutherAI/lm-evaluation-harness). You can access
the results by model and step at `results/json/*` in the [GitHub
repository](https://github.com/EleutherAI/pythia/tree/main/results/json/).<br>
Expand the sections below to see plots of evaluation results for all
Pythia and Pythia-deduped models compared with OPT and BLOOM.
<details>
<summary>LAMBADA – OpenAI</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/lambada_openai_v1.png" style="width:auto"/>
</details>
<details>
<summary>Physical Interaction: Question Answering (PIQA)</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/piqa_v1.png" style="width:auto"/>
</details>
<details>
<summary>WinoGrande</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/winogrande_v1.png" style="width:auto"/>
</details>
<details>
<summary>AI2 Reasoning Challenge—Easy Set</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/arc_easy_v1.png" style="width:auto"/>
</details>
<details>
<summary>SciQ</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/sciq_v1.png" style="width:auto"/>
</details>
## Changelog
This section compares differences between previously released
[Pythia v0](https://huggingface.co/models?other=pythia_v0) and the current
models. See Appendix B of the Pythia paper for further discussion of these
changes and the motivation behind them. We found that retraining Pythia had no
impact on benchmark performance.
- All model sizes are now trained with uniform batch size of 2M tokens.
Previously, the models of size 160M, 410M, and 1.4B parameters were trained
with batch sizes of 4M tokens.
- We added checkpoints at initialization (step 0) and steps {1,2,4,8,16,32,64,
128,256,512} in addition to every 1000 training steps.
- Flash Attention was used in the new retrained suite.
- We remedied a minor inconsistency that existed in the original suite: all
models of size 2.8B parameters or smaller had a learning rate (LR) schedule
which decayed to a minimum LR of 10% the starting LR rate, but the 6.9B and
12B models all used an LR schedule which decayed to a minimum LR of 0. In
the redone training runs, we rectified this inconsistency: all models now were
trained with LR decaying to a minimum of 0.1× their maximum LR.
### Naming convention and parameter count
*Pythia* models were renamed in January 2023. It is possible that the old
naming convention still persists in some documentation by accident. The
current naming convention (70M, 160M, etc.) is based on total parameter count.
<figure style="width:32em">
| current Pythia suffix | old suffix | total params | non-embedding params |
| --------------------: | ---------: | -------------: | -------------------: |
| 70M | 19M | 70,426,624 | 18,915,328 |
| 160M | 125M | 162,322,944 | 85,056,000 |
| 410M | 350M | 405,334,016 | 302,311,424 |
| 1B | 800M | 1,011,781,632 | 805,736,448 |
| 1.4B | 1.3B | 1,414,647,808 | 1,208,602,624 |
| 2.8B | 2.7B | 2,775,208,960 | 2,517,652,480 |
| 6.9B | 6.7B | 6,857,302,016 | 6,444,163,072 |
| 12B | 13B | 11,846,072,320 | 11,327,027,200 |
</figure>
|
EleutherAI/pythia-intervention-410m-deduped
|
EleutherAI
| 2023-07-09T16:00:37Z | 28 | 1 |
transformers
|
[
"transformers",
"pytorch",
"safetensors",
"gpt_neox",
"text-generation",
"dataset:EleutherAI/pile",
"arxiv:2304.01373",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2023-03-03T14:52:01Z |
---
license: apache-2.0
datasets:
- EleutherAI/pile
---
This model is part of an intervention study done in the paper [Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling](https://arxiv.org/abs/2304.01373) where we replaced all masculine pronouns with femanine ones and retrained the model for the last 21 billion tokens. The regular model can be found [here](https://huggingface.co/EleutherAI/pythia-410m-deduped).
**We do not recommend using this model for any purpose other than to study the influence of gender pronouns on language model behavior.**
|
EleutherAI/gpt-neo-2.7B
|
EleutherAI
| 2023-07-09T15:52:52Z | 192,791 | 467 |
transformers
|
[
"transformers",
"pytorch",
"jax",
"rust",
"safetensors",
"gpt_neo",
"text-generation",
"text generation",
"causal-lm",
"en",
"dataset:EleutherAI/pile",
"arxiv:2101.00027",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2022-03-02T23:29:04Z |
---
language:
- en
tags:
- text generation
- pytorch
- causal-lm
license: mit
datasets:
- EleutherAI/pile
---
# GPT-Neo 2.7B
## Model Description
GPT-Neo 2.7B is a transformer model designed using EleutherAI's replication of the GPT-3 architecture. GPT-Neo refers to the class of models, while 2.7B represents the number of parameters of this particular pre-trained model.
## Training data
GPT-Neo 2.7B was trained on the Pile, a large scale curated dataset created by EleutherAI for the purpose of training this model.
## Training procedure
This model was trained for 420 billion tokens over 400,000 steps. It was trained as a masked autoregressive language model, using cross-entropy loss.
## Intended Use and Limitations
This way, the model learns an inner representation of the English language that can then be used to extract features useful for downstream tasks. The model is best at what it was pretrained for however, which is generating texts from a prompt.
### How to use
You can use this model directly with a pipeline for text generation. This example generates a different sequence each time it's run:
```py
>>> from transformers import pipeline
>>> generator = pipeline('text-generation', model='EleutherAI/gpt-neo-2.7B')
>>> generator("EleutherAI has", do_sample=True, min_length=50)
[{'generated_text': 'EleutherAI has made a commitment to create new software packages for each of its major clients and has'}]
```
### Limitations and Biases
GPT-Neo was trained as an autoregressive language model. This means that its core functionality is taking a string of text and predicting the next token. While language models are widely used for tasks other than this, there are a lot of unknowns with this work.
GPT-Neo was trained on the Pile, a dataset known to contain profanity, lewd, and otherwise abrasive language. Depending on your usecase GPT-Neo may produce socially unacceptable text. See Sections 5 and 6 of the Pile paper for a more detailed analysis of the biases in the Pile.
As with all language models, it is hard to predict in advance how GPT-Neo will respond to particular prompts and offensive content may occur without warning. We recommend having a human curate or filter the outputs before releasing them, both to censor undesirable content and to improve the quality of the results.
## Eval results
All evaluations were done using our [evaluation harness](https://github.com/EleutherAI/lm-evaluation-harness). Some results for GPT-2 and GPT-3 are inconsistent with the values reported in the respective papers. We are currently looking into why, and would greatly appreciate feedback and further testing of our eval harness. If you would like to contribute evaluations you have done, please reach out on our [Discord](https://discord.gg/vtRgjbM).
### Linguistic Reasoning
| Model and Size | Pile BPB | Pile PPL | Wikitext PPL | Lambada PPL | Lambada Acc | Winogrande | Hellaswag |
| ---------------- | ---------- | ---------- | ------------- | ----------- | ----------- | ---------- | ----------- |
| GPT-Neo 1.3B | 0.7527 | 6.159 | 13.10 | 7.498 | 57.23% | 55.01% | 38.66% |
| GPT-2 1.5B | 1.0468 | ----- | 17.48 | 10.634 | 51.21% | 59.40% | 40.03% |
| **GPT-Neo 2.7B** | **0.7165** | **5.646** | **11.39** | **5.626** | **62.22%** | **56.50%** | **42.73%** |
| GPT-3 Ada | 0.9631 | ----- | ----- | 9.954 | 51.60% | 52.90% | 35.93% |
### Physical and Scientific Reasoning
| Model and Size | MathQA | PubMedQA | Piqa |
| ---------------- | ---------- | ---------- | ----------- |
| GPT-Neo 1.3B | 24.05% | 54.40% | 71.11% |
| GPT-2 1.5B | 23.64% | 58.33% | 70.78% |
| **GPT-Neo 2.7B** | **24.72%** | **57.54%** | **72.14%** |
| GPT-3 Ada | 24.29% | 52.80% | 68.88% |
### Down-Stream Applications
TBD
### BibTeX entry and citation info
To cite this model, use
```bibtex
@software{gpt-neo,
author = {Black, Sid and
Leo, Gao and
Wang, Phil and
Leahy, Connor and
Biderman, Stella},
title = {{GPT-Neo: Large Scale Autoregressive Language
Modeling with Mesh-Tensorflow}},
month = mar,
year = 2021,
note = {{If you use this software, please cite it using
these metadata.}},
publisher = {Zenodo},
version = {1.0},
doi = {10.5281/zenodo.5297715},
url = {https://doi.org/10.5281/zenodo.5297715}
}
@article{gao2020pile,
title={The Pile: An 800GB Dataset of Diverse Text for Language Modeling},
author={Gao, Leo and Biderman, Stella and Black, Sid and Golding, Laurence and Hoppe, Travis and Foster, Charles and Phang, Jason and He, Horace and Thite, Anish and Nabeshima, Noa and others},
journal={arXiv preprint arXiv:2101.00027},
year={2020}
}
```
|
EleutherAI/pythia-160m
|
EleutherAI
| 2023-07-09T15:52:09Z | 151,669 | 30 |
transformers
|
[
"transformers",
"pytorch",
"safetensors",
"gpt_neox",
"text-generation",
"causal-lm",
"pythia",
"en",
"dataset:EleutherAI/pile",
"arxiv:2304.01373",
"arxiv:2101.00027",
"arxiv:2201.07311",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2023-02-08T19:25:46Z |
---
language:
- en
tags:
- pytorch
- causal-lm
- pythia
license: apache-2.0
datasets:
- EleutherAI/pile
---
The *Pythia Scaling Suite* is a collection of models developed to facilitate
interpretability research [(see paper)](https://arxiv.org/pdf/2304.01373.pdf).
It contains two sets of eight models of sizes
70M, 160M, 410M, 1B, 1.4B, 2.8B, 6.9B, and 12B. For each size, there are two
models: one trained on the Pile, and one trained on the Pile after the dataset
has been globally deduplicated. All 8 model sizes are trained on the exact
same data, in the exact same order. We also provide 154 intermediate
checkpoints per model, hosted on Hugging Face as branches.
The Pythia model suite was deliberately designed to promote scientific
research on large language models, especially interpretability research.
Despite not centering downstream performance as a design goal, we find the
models <a href="#evaluations">match or exceed</a> the performance of
similar and same-sized models, such as those in the OPT and GPT-Neo suites.
<details>
<summary style="font-weight:600">Details on previous early release and naming convention.</summary>
Previously, we released an early version of the Pythia suite to the public.
However, we decided to retrain the model suite to address a few hyperparameter
discrepancies. This model card <a href="#changelog">lists the changes</a>;
see appendix B in the Pythia paper for further discussion. We found no
difference in benchmark performance between the two Pythia versions.
The old models are
[still available](https://huggingface.co/models?other=pythia_v0), but we
suggest the retrained suite if you are just starting to use Pythia.<br>
**This is the current release.**
Please note that all models in the *Pythia* suite were renamed in January
2023. For clarity, a <a href="#naming-convention-and-parameter-count">table
comparing the old and new names</a> is provided in this model card, together
with exact parameter counts.
</details>
<br>
# Pythia-160M
## Model Details
- Developed by: [EleutherAI](http://eleuther.ai)
- Model type: Transformer-based Language Model
- Language: English
- Learn more: [Pythia's GitHub repository](https://github.com/EleutherAI/pythia)
for training procedure, config files, and details on how to use.
[See paper](https://arxiv.org/pdf/2304.01373.pdf) for more evals and implementation
details.
- Library: [GPT-NeoX](https://github.com/EleutherAI/gpt-neox)
- License: Apache 2.0
- Contact: to ask questions about this model, join the [EleutherAI
Discord](https://discord.gg/zBGx3azzUn), and post them in `#release-discussion`.
Please read the existing *Pythia* documentation before asking about it in the
EleutherAI Discord. For general correspondence: [contact@eleuther.
ai](mailto:contact@eleuther.ai).
<figure>
| Pythia model | Non-Embedding Params | Layers | Model Dim | Heads | Batch Size | Learning Rate | Equivalent Models |
| -----------: | -------------------: | :----: | :-------: | :---: | :--------: | :-------------------: | :--------------------: |
| 70M | 18,915,328 | 6 | 512 | 8 | 2M | 1.0 x 10<sup>-3</sup> | — |
| 160M | 85,056,000 | 12 | 768 | 12 | 2M | 6.0 x 10<sup>-4</sup> | GPT-Neo 125M, OPT-125M |
| 410M | 302,311,424 | 24 | 1024 | 16 | 2M | 3.0 x 10<sup>-4</sup> | OPT-350M |
| 1.0B | 805,736,448 | 16 | 2048 | 8 | 2M | 3.0 x 10<sup>-4</sup> | — |
| 1.4B | 1,208,602,624 | 24 | 2048 | 16 | 2M | 2.0 x 10<sup>-4</sup> | GPT-Neo 1.3B, OPT-1.3B |
| 2.8B | 2,517,652,480 | 32 | 2560 | 32 | 2M | 1.6 x 10<sup>-4</sup> | GPT-Neo 2.7B, OPT-2.7B |
| 6.9B | 6,444,163,072 | 32 | 4096 | 32 | 2M | 1.2 x 10<sup>-4</sup> | OPT-6.7B |
| 12B | 11,327,027,200 | 36 | 5120 | 40 | 2M | 1.2 x 10<sup>-4</sup> | — |
<figcaption>Engineering details for the <i>Pythia Suite</i>. Deduped and
non-deduped models of a given size have the same hyperparameters. “Equivalent”
models have <b>exactly</b> the same architecture, and the same number of
non-embedding parameters.</figcaption>
</figure>
## Uses and Limitations
### Intended Use
The primary intended use of Pythia is research on the behavior, functionality,
and limitations of large language models. This suite is intended to provide
a controlled setting for performing scientific experiments. We also provide
154 checkpoints per model: initial `step0`, 10 log-spaced checkpoints
`step{1,2,4...512}`, and 143 evenly-spaced checkpoints from `step1000` to
`step143000`. These checkpoints are hosted on Hugging Face as branches. Note
that branch `143000` corresponds exactly to the model checkpoint on the `main`
branch of each model.
You may also further fine-tune and adapt Pythia-160M for deployment,
as long as your use is in accordance with the Apache 2.0 license. Pythia
models work with the Hugging Face [Transformers
Library](https://huggingface.co/docs/transformers/index). If you decide to use
pre-trained Pythia-160M as a basis for your fine-tuned model, please
conduct your own risk and bias assessment.
### Out-of-scope use
The Pythia Suite is **not** intended for deployment. It is not a in itself
a product and cannot be used for human-facing interactions. For example,
the model may generate harmful or offensive text. Please evaluate the risks
associated with your particular use case.
Pythia models are English-language only, and are not suitable for translation
or generating text in other languages.
Pythia-160M has not been fine-tuned for downstream contexts in which
language models are commonly deployed, such as writing genre prose,
or commercial chatbots. This means Pythia-160M will **not**
respond to a given prompt the way a product like ChatGPT does. This is because,
unlike this model, ChatGPT was fine-tuned using methods such as Reinforcement
Learning from Human Feedback (RLHF) to better “follow” human instructions.
### Limitations and biases
The core functionality of a large language model is to take a string of text
and predict the next token. The token used by the model need not produce the
most “accurate” text. Never rely on Pythia-160M to produce factually accurate
output.
This model was trained on [the Pile](https://pile.eleuther.ai/), a dataset
known to contain profanity and texts that are lewd or otherwise offensive.
See [Section 6 of the Pile paper](https://arxiv.org/abs/2101.00027) for a
discussion of documented biases with regards to gender, religion, and race.
Pythia-160M may produce socially unacceptable or undesirable text, *even if*
the prompt itself does not include anything explicitly offensive.
If you plan on using text generated through, for example, the Hosted Inference
API, we recommend having a human curate the outputs of this language model
before presenting it to other people. Please inform your audience that the
text was generated by Pythia-160M.
### Quickstart
Pythia models can be loaded and used via the following code, demonstrated here
for the third `pythia-70m-deduped` checkpoint:
```python
from transformers import GPTNeoXForCausalLM, AutoTokenizer
model = GPTNeoXForCausalLM.from_pretrained(
"EleutherAI/pythia-70m-deduped",
revision="step3000",
cache_dir="./pythia-70m-deduped/step3000",
)
tokenizer = AutoTokenizer.from_pretrained(
"EleutherAI/pythia-70m-deduped",
revision="step3000",
cache_dir="./pythia-70m-deduped/step3000",
)
inputs = tokenizer("Hello, I am", return_tensors="pt")
tokens = model.generate(**inputs)
tokenizer.decode(tokens[0])
```
Revision/branch `step143000` corresponds exactly to the model checkpoint on
the `main` branch of each model.<br>
For more information on how to use all Pythia models, see [documentation on
GitHub](https://github.com/EleutherAI/pythia).
## Training
### Training data
[The Pile](https://pile.eleuther.ai/) is a 825GiB general-purpose dataset in
English. It was created by EleutherAI specifically for training large language
models. It contains texts from 22 diverse sources, roughly broken down into
five categories: academic writing (e.g. arXiv), internet (e.g. CommonCrawl),
prose (e.g. Project Gutenberg), dialogue (e.g. YouTube subtitles), and
miscellaneous (e.g. GitHub, Enron Emails). See [the Pile
paper](https://arxiv.org/abs/2101.00027) for a breakdown of all data sources,
methodology, and a discussion of ethical implications. Consult [the
datasheet](https://arxiv.org/abs/2201.07311) for more detailed documentation
about the Pile and its component datasets. The Pile can be downloaded from
the [official website](https://pile.eleuther.ai/), or from a [community
mirror](https://the-eye.eu/public/AI/pile/).<br>
The Pile was **not** deduplicated before being used to train Pythia-160M.
### Training procedure
All models were trained on the exact same data, in the exact same order. Each
model saw 299,892,736,000 tokens during training, and 143 checkpoints for each
model are saved every 2,097,152,000 tokens, spaced evenly throughout training,
from `step1000` to `step143000` (which is the same as `main`). In addition, we
also provide frequent early checkpoints: `step0` and `step{1,2,4...512}`.
This corresponds to training for just under 1 epoch on the Pile for
non-deduplicated models, and about 1.5 epochs on the deduplicated Pile.
All *Pythia* models trained for 143000 steps at a batch size
of 2M (2,097,152 tokens).<br>
See [GitHub](https://github.com/EleutherAI/pythia) for more details on training
procedure, including [how to reproduce
it](https://github.com/EleutherAI/pythia/blob/main/README.md#reproducing-training).<br>
Pythia uses the same tokenizer as [GPT-NeoX-
20B](https://huggingface.co/EleutherAI/gpt-neox-20b).
## Evaluations
All 16 *Pythia* models were evaluated using the [LM Evaluation
Harness](https://github.com/EleutherAI/lm-evaluation-harness). You can access
the results by model and step at `results/json/*` in the [GitHub
repository](https://github.com/EleutherAI/pythia/tree/main/results/json/).<br>
Expand the sections below to see plots of evaluation results for all
Pythia and Pythia-deduped models compared with OPT and BLOOM.
<details>
<summary>LAMBADA – OpenAI</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/lambada_openai_v1.png" style="width:auto"/>
</details>
<details>
<summary>Physical Interaction: Question Answering (PIQA)</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/piqa_v1.png" style="width:auto"/>
</details>
<details>
<summary>WinoGrande</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/winogrande_v1.png" style="width:auto"/>
</details>
<details>
<summary>AI2 Reasoning Challenge—Easy Set</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/arc_easy_v1.png" style="width:auto"/>
</details>
<details>
<summary>SciQ</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/sciq_v1.png" style="width:auto"/>
</details>
## Changelog
This section compares differences between previously released
[Pythia v0](https://huggingface.co/models?other=pythia_v0) and the current
models. See Appendix B of the Pythia paper for further discussion of these
changes and the motivation behind them. We found that retraining Pythia had no
impact on benchmark performance.
- All model sizes are now trained with uniform batch size of 2M tokens.
Previously, the models of size 160M, 410M, and 1.4B parameters were trained
with batch sizes of 4M tokens.
- We added checkpoints at initialization (step 0) and steps {1,2,4,8,16,32,64,
128,256,512} in addition to every 1000 training steps.
- Flash Attention was used in the new retrained suite.
- We remedied a minor inconsistency that existed in the original suite: all
models of size 2.8B parameters or smaller had a learning rate (LR) schedule
which decayed to a minimum LR of 10% the starting LR rate, but the 6.9B and
12B models all used an LR schedule which decayed to a minimum LR of 0. In
the redone training runs, we rectified this inconsistency: all models now were
trained with LR decaying to a minimum of 0.1× their maximum LR.
### Naming convention and parameter count
*Pythia* models were renamed in January 2023. It is possible that the old
naming convention still persists in some documentation by accident. The
current naming convention (70M, 160M, etc.) is based on total parameter count.
<figure style="width:32em">
| current Pythia suffix | old suffix | total params | non-embedding params |
| --------------------: | ---------: | -------------: | -------------------: |
| 70M | 19M | 70,426,624 | 18,915,328 |
| 160M | 125M | 162,322,944 | 85,056,000 |
| 410M | 350M | 405,334,016 | 302,311,424 |
| 1B | 800M | 1,011,781,632 | 805,736,448 |
| 1.4B | 1.3B | 1,414,647,808 | 1,208,602,624 |
| 2.8B | 2.7B | 2,775,208,960 | 2,517,652,480 |
| 6.9B | 6.7B | 6,857,302,016 | 6,444,163,072 |
| 12B | 13B | 11,846,072,320 | 11,327,027,200 |
</figure>
|
chunwoolee0/my_awesome_eli5_mlm_model
|
chunwoolee0
| 2023-07-09T15:48:18Z | 106 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"roberta",
"fill-mask",
"generated_from_trainer",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2023-07-09T15:20:06Z |
---
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: my_awesome_eli5_mlm_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. -->
# my_awesome_eli5_mlm_model
This model is a fine-tuned version of [distilroberta-base](https://huggingface.co/distilroberta-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 2.0053
## 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: 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
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 2.2387 | 1.0 | 1128 | 2.0397 |
| 2.1586 | 2.0 | 2256 | 2.0042 |
| 2.1161 | 3.0 | 3384 | 2.0031 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
Sekiraw/ReachDense
|
Sekiraw
| 2023-07-09T15:46:32Z | 2 | 0 |
stable-baselines3
|
[
"stable-baselines3",
"PandaReachDense-v2",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-07-09T15:45:53Z |
---
library_name: stable-baselines3
tags:
- PandaReachDense-v2
- deep-reinforcement-learning
- reinforcement-learning
- stable-baselines3
model-index:
- name: A2C
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: PandaReachDense-v2
type: PandaReachDense-v2
metrics:
- type: mean_reward
value: -0.48 +/- 0.14
name: mean_reward
verified: false
---
# **A2C** Agent playing **PandaReachDense-v2**
This is a trained model of a **A2C** agent playing **PandaReachDense-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
...
```
|
Victornelas/Aula1
|
Victornelas
| 2023-07-09T15:44:31Z | 0 | 0 | null |
[
"region:us"
] | null | 2023-07-09T14:17:48Z |
#configuração para não receber warnings
import warning
warnings.filterwarnings("ignore")
#import necessários
import panda as pd
import numpy as np
import matplotlib.pyplot as plt
from sklearn.dataset import load_diabetes
from sklearn.model_selection import train_test_split
from sklearn.model_selection import kfold
from sklearn.model_selection import cross_val_score
from sklearn.metrics import mean_squared_error
from sklearn.linear_model import linearregression
from sklearn.linear_model import ridge
from sklearn.linear_model import lasso
from sklearn.neighbors import kneighborsregressor
from sklearn.tree import decisi ontreeregressor
fromsklearn.svm
diabetes = load_diabetes()
dataset = pd.dataframe(diabetes.data, columns=diabetes.feature_names)
dataset['target'] = diabetes.target
dataset.head()
array = dataset.value
x=array[:,0:10]
y=array[:,10]
x_train, x_test, y_train, y_test = train_test_split(X,y, test_size=0.2tate=7
num_particoes = 10
kfold = kfold(n_split=num_particoes, shuffle=treu, random_state=7)
np.random.seed(7)
models = []
result = []
names = []
models.append(('lr', linearregression()))
models.append(('ridge', ridge()))
models.append(('lasso', lasso()))
models.append(('knn', kneighborsregressor()))
models.append(('cart', decisiontreeregressor()))
models.append(('svm', svr()))
for name, model in models:
cv_result = cross_val_score(model, x_train, y_train, cv=kfold, scoring='neg_mean_squared_error')
results.append(cv_results)
names.append(name)
msg="%s: MSE %0.2f (%0.2f) -rmse %0.2f" $ (name, abs(cv_results.mean()), cv_results.std(), np.sqrt(abs(cv_results.mean())))
print(msg)
fig = plt.figure()
fig.suptitle('comparação do mse dos modelos')
ax = fig.add_subplot(111)
plt.boxplot(results)
ax.set_xticklabels(names)
plt.show()
|
afterthougt/kullm-polyglot-12.8b-v2_700steps
|
afterthougt
| 2023-07-09T15:31:45Z | 5 | 0 |
peft
|
[
"peft",
"gpt_neox",
"region:us"
] | null | 2023-07-06T05:04:33Z |
---
library_name: peft
---
## 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 0.4.0.dev0
|
lordtt13/blenderbot_small-news
|
lordtt13
| 2023-07-09T15:28:39Z | 111 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"safetensors",
"blenderbot-small",
"text2text-generation",
"en",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-02T23:29:05Z |
---
language: en
---
## BlenderBotSmall-News: Small version of a state-of-the-art open source chatbot, trained on custom summaries
### Details of BlenderBotSmall
The **BlenderBotSmall** model was presented in [A state-of-the-art open source chatbot](https://ai.facebook.com/blog/state-of-the-art-open-source-chatbot/) by *Facebook AI* and here are it's details:
- Facebook AI has built and open-sourced BlenderBot, the largest-ever open-domain chatbot. It outperforms others in terms of engagement and also feels more human, according to human evaluators.
- The culmination of years of research in conversational AI, this is the first chatbot to blend a diverse set of conversational skills — including empathy, knowledge, and personality — together in one system.
- We achieved this milestone through a new chatbot recipe that includes improved decoding techniques, novel blending of skills, and a model with 9.4 billion parameters, which is 3.6x more than the largest existing system.
### Details of the downstream task (Summarization) - Dataset 📚
A custom dataset was used, which was hand prepared by [SmokeTrees Digital](https://github.com/smoke-trees) AI engineers. This data contains long texts and summaries.
### Model training
The training script is present [here](https://github.com/lordtt13/transformers-experiments/blob/master/Custom%20Tasks/fine-tune-blenderbot_small-for-summarization.ipynb).
### Pipelining the Model
```python
model = transformers.BlenderbotSmallForConditionalGeneration.from_pretrained('lordtt13/blenderbot_small-news')
tokenizer = transformers.BlenderbotSmallTokenizer.from_pretrained("lordtt13/blenderbot_small-news")
nlp_fill = transformers.pipeline('summarization', model = model, tokenizer = tokenizer)
nlp_fill('The CBI on Saturday booked four former officials of Syndicate Bank and six others for cheating, forgery, criminal conspiracy and causing ₹209 crore loss to the state-run bank. The accused had availed home loans and credit from Syndicate Bank on the basis of forged and fabricated documents. These funds were fraudulently transferred to the companies owned by the accused persons.', min_length=5, max_length=40)
# Output:
# [{'summary_text': 'marize: the cbi booked four former officials of syndicate bank and six others for cheating , forgery , criminal conspiracy and causing 209 crore loss to the staterun bank'}]
```
> Created by [Tanmay Thakur](https://github.com/lordtt13) | [LinkedIn](https://www.linkedin.com/in/tanmay-thakur-6bb5a9154/)
|
Gorttham/flan-t5-small-chat
|
Gorttham
| 2023-07-09T15:07:44Z | 111 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"t5",
"text2text-generation",
"generated_from_trainer",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2023-07-02T10:11:21Z |
---
license: apache-2.0
tags:
- generated_from_trainer
metrics:
- rouge
model-index:
- name: content
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. -->
# content
This model is a fine-tuned version of [google/flan-t5-small](https://huggingface.co/google/flan-t5-small) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 2.5985
- Rouge1: 32.7607
- Rouge2: 19.5507
- Rougel: 32.7312
- Rougelsum: 32.7306
- Gen Len: 16.4212
## 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: 4
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:-------:|
| 3.043 | 1.0 | 1348 | 2.7636 | 34.4245 | 20.9962 | 34.386 | 34.3876 | 15.1150 |
| 2.8078 | 2.0 | 2696 | 2.6540 | 32.5342 | 19.3983 | 32.4966 | 32.4947 | 16.7662 |
| 2.7166 | 3.0 | 4044 | 2.6103 | 32.4564 | 19.3597 | 32.4255 | 32.4355 | 16.6037 |
| 2.6876 | 4.0 | 5392 | 2.5985 | 32.7607 | 19.5507 | 32.7312 | 32.7306 | 16.4212 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
MostafaHamwi/TextSimplification
|
MostafaHamwi
| 2023-07-09T15:05:20Z | 63 | 0 |
transformers
|
[
"transformers",
"tf",
"t5",
"text2text-generation",
"generated_from_keras_callback",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2023-06-29T23:53:00Z |
---
license: apache-2.0
tags:
- generated_from_keras_callback
model-index:
- name: TextSimplification
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. -->
# TextSimplification
This model is a fine-tuned version of [t5-small](https://huggingface.co/t5-small) on D-Wikiepdia 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:
- optimizer: None
- training_precision: float32
### Training results
### Framework versions
- Transformers 4.30.2
- TensorFlow 2.12.0
- Tokenizers 0.13.3
|
whiteDandelion/xlm-roberta-base-finetuned-panx-de
|
whiteDandelion
| 2023-07-09T15:05:10Z | 105 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"xlm-roberta",
"token-classification",
"generated_from_trainer",
"dataset:xtreme",
"license:mit",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
token-classification
| 2023-07-09T14:54:31Z |
---
license: mit
tags:
- generated_from_trainer
datasets:
- xtreme
metrics:
- f1
model-index:
- name: xlm-roberta-base-finetuned-panx-de
results:
- task:
name: Token Classification
type: token-classification
dataset:
name: xtreme
type: xtreme
config: PAN-X.de
split: validation
args: PAN-X.de
metrics:
- name: F1
type: f1
value: 0.8653353814644136
---
<!-- 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. -->
# xlm-roberta-base-finetuned-panx-de
This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the xtreme dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1339
- F1: 0.8653
## 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: 24
- eval_batch_size: 24
- 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 | F1 |
|:-------------:|:-----:|:----:|:---------------:|:------:|
| 0.2583 | 1.0 | 525 | 0.1596 | 0.8231 |
| 0.1262 | 2.0 | 1050 | 0.1395 | 0.8468 |
| 0.0824 | 3.0 | 1575 | 0.1339 | 0.8653 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
mrtimmydontplay/PKthunda
|
mrtimmydontplay
| 2023-07-09T14:53:37Z | 0 | 0 |
adapter-transformers
|
[
"adapter-transformers",
"en",
"license:other",
"region:us"
] | null | 2023-07-09T12:31:23Z |
---
license: other
language:
- en
metrics:
- bleu
- accuracy
- code_eval
library_name: adapter-transformers
---
|
Chocoboko/OTN_BDSM
|
Chocoboko
| 2023-07-09T14:47:10Z | 0 | 0 | null |
[
"license:creativeml-openrail-m",
"region:us"
] | null | 2023-06-17T10:26:26Z |
---
license: creativeml-openrail-m
---
|
agercas/whisper-small-dv
|
agercas
| 2023-07-09T14:43:02Z | 79 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"whisper",
"automatic-speech-recognition",
"generated_from_trainer",
"dv",
"dataset:mozilla-foundation/common_voice_13_0",
"license:apache-2.0",
"model-index",
"endpoints_compatible",
"region:us"
] |
automatic-speech-recognition
| 2023-07-09T13:43:58Z |
---
language:
- dv
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- mozilla-foundation/common_voice_13_0
metrics:
- wer
model-index:
- name: Whisper Small Dv
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: Common Voice 13
type: mozilla-foundation/common_voice_13_0
config: dv
split: test
args: dv
metrics:
- name: Wer
type: wer
value: 110.95037729944013
---
<!-- 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 Small Dv
This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Common Voice 13 dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1971
- Wer Ortho: 206.4141
- Wer: 110.9504
## 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: 1e-05
- train_batch_size: 8
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: constant_with_warmup
- lr_scheduler_warmup_steps: 50
- training_steps: 500
### Training results
| Training Loss | Epoch | Step | Validation Loss | Wer Ortho | Wer |
|:-------------:|:-----:|:----:|:---------------:|:---------:|:--------:|
| 0.1714 | 0.82 | 500 | 0.1971 | 206.4141 | 110.9504 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
edures/ppo-Huggy
|
edures
| 2023-07-09T14:41:48Z | 32 | 0 |
ml-agents
|
[
"ml-agents",
"tensorboard",
"onnx",
"Huggy",
"deep-reinforcement-learning",
"reinforcement-learning",
"ML-Agents-Huggy",
"region:us"
] |
reinforcement-learning
| 2023-07-09T14:41:37Z |
---
library_name: ml-agents
tags:
- Huggy
- deep-reinforcement-learning
- reinforcement-learning
- ML-Agents-Huggy
---
# **ppo** Agent playing **Huggy**
This is a trained model of a **ppo** agent playing **Huggy**
using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents).
## Usage (with ML-Agents)
The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/
We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub:
- A *short tutorial* where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your
browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction
- A *longer tutorial* to understand how works ML-Agents:
https://huggingface.co/learn/deep-rl-course/unit5/introduction
### Resume the training
```bash
mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume
```
### Watch your Agent play
You can watch your agent **playing directly in your browser**
1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity
2. Step 1: Find your model_id: edures/ppo-Huggy
3. Step 2: Select your *.nn /*.onnx file
4. Click on Watch the agent play 👀
|
hyperr123/mematibas
|
hyperr123
| 2023-07-09T14:39:57Z | 0 | 0 | null |
[
"music",
"tr",
"license:openrail",
"region:us"
] | null | 2023-07-09T13:53:48Z |
---
license: openrail
language:
- tr
tags:
- music
---
|
RogerB/afro-xlmr-base-finetuned-kintweetsD
|
RogerB
| 2023-07-09T14:38:17Z | 88 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"xlm-roberta",
"fill-mask",
"generated_from_trainer",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2023-07-09T14:12:07Z |
---
license: mit
tags:
- generated_from_trainer
model-index:
- name: afro-xlmr-base-finetuned-kintweetsD
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. -->
# afro-xlmr-base-finetuned-kintweetsD
This model is a fine-tuned version of [Davlan/afro-xlmr-base](https://huggingface.co/Davlan/afro-xlmr-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 2.1283
## 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: 10
- eval_batch_size: 10
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 2.4428 | 1.0 | 900 | 2.1947 |
| 2.3168 | 2.0 | 1800 | 2.1566 |
| 2.2497 | 3.0 | 2700 | 2.1290 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
antolin/distilroberta-base-csn-python-bimodal
|
antolin
| 2023-07-09T14:34:22Z | 86 | 0 |
transformers
|
[
"transformers",
"pytorch",
"roberta",
"fill-mask",
"dataset:code_search_net",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2023-07-09T09:24:49Z |
---
datasets:
- code_search_net
widget:
- text: "def <mask> ( a, b ) : if a > b : return a else return b</s>return the maximum value"
- text: "def <mask> ( a, b ) : if a > b : return a else return b"
---
# Model Architecture
This model follows the distilroberta-base architecture. Futhermore, this model was initialized with the checkpoint of distilroberta-base.
# Pre-training phase
This model was pre-trained with the MLM objective (`mlm_probability=0.15`).
During this phase, the inputs had the following format:
$$\left[[CLS], t_1, \dots, t_n, [SEP], w_1, \dots, w_m\right[EOS]]$$
where $t_1, \dots, t_n$ are the code tokens and $w_1, \dots, w_m$ are the natural language description tokens. More concretely, this is the snippet that tokenizes the input:
```python
def tokenize_function_bimodal(examples, tokenizer, max_len):
codes = [' '.join(example) for example in examples['func_code_tokens']]
nls = [' '.join(example) for example in examples['func_documentation_tokens']]
pairs = [[c, nl] for c, nl in zip(codes, nls)]
return tokenizer(pairs, max_length=max_len, padding="max_length", truncation=True)
```
# Training details
- Max length: 512
- Effective batch size: 64
- Total steps: 60000
- Learning rate: 5e-4
# Usage
```python
model = AutoModelForMaskedLM.from_pretrained('antolin/distilroberta-base-csn-python-bimodal')
tokenizer = AutoTokenizer.from_pretrained('antolin/distilroberta-base-csn-python-bimodal')
mask_filler = pipeline("fill-mask", model=model, tokenizer=tokenizer)
code_tokens = ["def", "<mask>", "(", "a", ",", "b", ")", ":", "if", "a", ">", "b", ":", "return", "a", "else", "return", "b"]
nl_tokens = ["return", "the", "maximum", "value"]
input_text = ' '.join(code_tokens) + tokenizer.sep_token + ' '.join(nl_tokens)
pprint(mask_filler(input_text, top_k=5))
```
```shell
[{'score': 0.4645618796348572,
'sequence': 'def max ( a, b ) : if a > b : return a else return b return '
'the maximum value',
'token': 19220,
'token_str': ' max'},
{'score': 0.40963634848594666,
'sequence': 'def maximum ( a, b ) : if a > b : return a else return b '
'return the maximum value',
'token': 4532,
'token_str': ' maximum'},
{'score': 0.02103462442755699,
'sequence': 'def min ( a, b ) : if a > b : return a else return b return '
'the maximum value',
'token': 5251,
'token_str': ' min'},
{'score': 0.014217409305274487,
'sequence': 'def value ( a, b ) : if a > b : return a else return b return '
'the maximum value',
'token': 923,
'token_str': ' value'},
{'score': 0.010762304067611694,
'sequence': 'def minimum ( a, b ) : if a > b : return a else return b '
'return the maximum value',
'token': 3527,
'token_str': ' minimum'}]
```
|
LarryAIDraw/calamiti
|
LarryAIDraw
| 2023-07-09T14:32:09Z | 0 | 0 | null |
[
"license:creativeml-openrail-m",
"region:us"
] | null | 2023-07-09T14:27:06Z |
---
license: creativeml-openrail-m
---
https://civitai.com/models/105180/calamity-jane-fate-grand-order
|
LarryAIDraw/ilia_coral
|
LarryAIDraw
| 2023-07-09T14:31:44Z | 0 | 0 | null |
[
"license:creativeml-openrail-m",
"region:us"
] | null | 2023-07-09T14:25:58Z |
---
license: creativeml-openrail-m
---
https://civitai.com/models/105511/ilia-coral-tenten-kakumei-or-or
|
Chocoboko/jpdamsel
|
Chocoboko
| 2023-07-09T14:28:31Z | 0 | 0 | null |
[
"license:creativeml-openrail-m",
"region:us"
] | null | 2023-07-09T14:15:15Z |
---
license: creativeml-openrail-m
---
|
AndrewL088/ppo-LunarLander-v2
|
AndrewL088
| 2023-07-09T14:23:29Z | 0 | 0 |
stable-baselines3
|
[
"stable-baselines3",
"LunarLander-v2",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-07-09T07:55:11Z |
---
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: 280.72 +/- 19.93
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
...
```
|
zamarano/my_awesome_opus_books_model
|
zamarano
| 2023-07-09T14:17:14Z | 84 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"t5",
"text2text-generation",
"generated_from_trainer",
"dataset:opus_books",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2023-07-09T00:52:09Z |
---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- opus_books
metrics:
- bleu
model-index:
- name: my_awesome_opus_books_model
results:
- task:
name: Sequence-to-sequence Language Modeling
type: text2text-generation
dataset:
name: opus_books
type: opus_books
config: en-fr
split: train
args: en-fr
metrics:
- name: Bleu
type: bleu
value: 5.6226
---
<!-- 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. -->
# my_awesome_opus_books_model
This model is a fine-tuned version of [t5-small](https://huggingface.co/t5-small) on the opus_books dataset.
It achieves the following results on the evaluation set:
- Loss: 1.6080
- Bleu: 5.6226
- Gen Len: 17.5745
## 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 | Bleu | Gen Len |
|:-------------:|:-----:|:-----:|:---------------:|:------:|:-------:|
| 1.8675 | 1.0 | 6355 | 1.6318 | 5.4409 | 17.5848 |
| 1.8199 | 2.0 | 12710 | 1.6080 | 5.6226 | 17.5745 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
soBeauty/xlm-roberta-base-09072023-revised
|
soBeauty
| 2023-07-09T14:17:04Z | 5 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"xlm-roberta",
"fill-mask",
"generated_from_trainer",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2023-07-09T11:44:21Z |
---
license: mit
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: xlm-roberta-base-09072023-revised
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. -->
# xlm-roberta-base-09072023-revised
This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on an unknown dataset.
It achieves the following results on the evaluation set:
- Accuracy: 0.7354
- Loss: 1.2865
## 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: 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: 10
### Training results
| Training Loss | Epoch | Step | Accuracy | Validation Loss |
|:-------------:|:-----:|:----:|:--------:|:---------------:|
| 1.2004 | 0.77 | 100 | 0.6813 | 1.4010 |
| 1.5282 | 1.54 | 200 | 0.7119 | 1.2520 |
| 1.6864 | 2.31 | 300 | 0.6591 | 1.5774 |
| 1.5648 | 3.08 | 400 | 0.72 | 1.3837 |
| 1.6035 | 3.85 | 500 | 0.7092 | 1.3721 |
| 1.6456 | 4.62 | 600 | 0.6557 | 1.5037 |
| 1.472 | 5.38 | 700 | 0.6822 | 1.3919 |
| 1.5617 | 6.15 | 800 | 0.7014 | 1.4154 |
| 1.4883 | 6.92 | 900 | 0.7269 | 1.2583 |
| 1.4402 | 7.69 | 1000 | 0.6877 | 1.5842 |
| 1.5903 | 8.46 | 1100 | 0.7184 | 1.3132 |
| 1.4025 | 9.23 | 1200 | 0.7148 | 1.2230 |
| 1.4793 | 10.0 | 1300 | 0.7354 | 1.2865 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
sourabhdattawad/spoken-language-detection
|
sourabhdattawad
| 2023-07-09T14:08:30Z | 62 | 1 |
transformers
|
[
"transformers",
"pytorch",
"wav2vec2",
"audio-classification",
"endpoints_compatible",
"region:us"
] |
audio-classification
| 2023-07-09T14:04:58Z |
---
pipeline_tag: audio-classification
---
|
k1101jh/ppo-Huggy
|
k1101jh
| 2023-07-09T14:02:33Z | 0 | 0 |
ml-agents
|
[
"ml-agents",
"tensorboard",
"onnx",
"Huggy",
"deep-reinforcement-learning",
"reinforcement-learning",
"ML-Agents-Huggy",
"region:us"
] |
reinforcement-learning
| 2023-07-09T14:02:29Z |
---
library_name: ml-agents
tags:
- Huggy
- deep-reinforcement-learning
- reinforcement-learning
- ML-Agents-Huggy
---
# **ppo** Agent playing **Huggy**
This is a trained model of a **ppo** agent playing **Huggy**
using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents).
## Usage (with ML-Agents)
The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/
We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub:
- A *short tutorial* where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your
browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction
- A *longer tutorial* to understand how works ML-Agents:
https://huggingface.co/learn/deep-rl-course/unit5/introduction
### Resume the training
```bash
mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume
```
### Watch your Agent play
You can watch your agent **playing directly in your browser**
1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity
2. Step 1: Find your model_id: k1101jh/ppo-Huggy
3. Step 2: Select your *.nn /*.onnx file
4. Click on Watch the agent play 👀
|
RogerB/KinyaBERT-small-finetuned-kintweetsD
|
RogerB
| 2023-07-09T13:57:42Z | 94 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"bert",
"fill-mask",
"generated_from_trainer",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2023-07-09T13:50:29Z |
---
tags:
- generated_from_trainer
model-index:
- name: KinyaBERT-small-finetuned-kintweetsD
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. -->
# KinyaBERT-small-finetuned-kintweetsD
This model is a fine-tuned version of [jean-paul/KinyaBERT-small](https://huggingface.co/jean-paul/KinyaBERT-small) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 3.8590
## 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: 10
- eval_batch_size: 10
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 4.339 | 1.0 | 900 | 3.9584 |
| 4.0319 | 2.0 | 1800 | 3.8580 |
| 3.924 | 3.0 | 2700 | 3.8051 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
mort1k/ppo-Huggy
|
mort1k
| 2023-07-09T13:55:18Z | 0 | 0 |
ml-agents
|
[
"ml-agents",
"tensorboard",
"onnx",
"Huggy",
"deep-reinforcement-learning",
"reinforcement-learning",
"ML-Agents-Huggy",
"region:us"
] |
reinforcement-learning
| 2023-07-09T13:55:13Z |
---
library_name: ml-agents
tags:
- Huggy
- deep-reinforcement-learning
- reinforcement-learning
- ML-Agents-Huggy
---
# **ppo** Agent playing **Huggy**
This is a trained model of a **ppo** agent playing **Huggy**
using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents).
## Usage (with ML-Agents)
The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/
We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub:
- A *short tutorial* where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your
browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction
- A *longer tutorial* to understand how works ML-Agents:
https://huggingface.co/learn/deep-rl-course/unit5/introduction
### Resume the training
```bash
mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume
```
### Watch your Agent play
You can watch your agent **playing directly in your browser**
1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity
2. Step 1: Find your model_id: mort1k/ppo-Huggy
3. Step 2: Select your *.nn /*.onnx file
4. Click on Watch the agent play 👀
|
nitishkumargundapu793/chat-bot_response
|
nitishkumargundapu793
| 2023-07-09T13:26:03Z | 0 | 0 | null |
[
"region:us"
] | null | 2023-07-09T13:11:03Z |
---
title: Chat Bot Response
emoji: 👁
colorFrom: indigo
colorTo: yellow
sdk: gradio
sdk_version: 3.0.11
app_file: app.py
pinned: false
---
Check out the configuration reference at https://huggingface.co/docs/hub/spaces#reference
|
RajkNakka/rl_course_vizdoom_health_gathering_supreme
|
RajkNakka
| 2023-07-09T13:25:43Z | 0 | 0 |
sample-factory
|
[
"sample-factory",
"tensorboard",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-07-08T23:18:32Z |
---
library_name: sample-factory
tags:
- deep-reinforcement-learning
- reinforcement-learning
- sample-factory
model-index:
- name: APPO
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: doom_health_gathering_supreme
type: doom_health_gathering_supreme
metrics:
- type: mean_reward
value: 11.60 +/- 5.01
name: mean_reward
verified: false
---
A(n) **APPO** model trained on the **doom_health_gathering_supreme** environment.
This model was trained using Sample-Factory 2.0: https://github.com/alex-petrenko/sample-factory.
Documentation for how to use Sample-Factory can be found at https://www.samplefactory.dev/
## Downloading the model
After installing Sample-Factory, download the model with:
```
python -m sample_factory.huggingface.load_from_hub -r RajkNakka/rl_course_vizdoom_health_gathering_supreme
```
## Using the model
To run the model after download, use the `enjoy` script corresponding to this environment:
```
python -m .usr.local.lib.python3.10.dist-packages.ipykernel_launcher --algo=APPO --env=doom_health_gathering_supreme --train_dir=./train_dir --experiment=rl_course_vizdoom_health_gathering_supreme
```
You can also upload models to the Hugging Face Hub using the same script with the `--push_to_hub` flag.
See https://www.samplefactory.dev/10-huggingface/huggingface/ for more details
## Training with this model
To continue training with this model, use the `train` script corresponding to this environment:
```
python -m .usr.local.lib.python3.10.dist-packages.ipykernel_launcher --algo=APPO --env=doom_health_gathering_supreme --train_dir=./train_dir --experiment=rl_course_vizdoom_health_gathering_supreme --restart_behavior=resume --train_for_env_steps=10000000000
```
Note, you may have to adjust `--train_for_env_steps` to a suitably high number as the experiment will resume at the number of steps it concluded at.
|
steinhaug/models-nsfw
|
steinhaug
| 2023-07-09T13:24:46Z | 0 | 18 | null |
[
"safetensors",
"region:us"
] | null | 2023-05-27T05:03:49Z |
## ./loras/*
Loads of lora files from civitai.
## Animatrix v2.0
animatrix_v20.safetensors, animatrix_inpaintV20.safetensors
cmid: 21916, cmvid: 44827
[civitai](https://civitai.com/models/21916?modelVersionId=44827)
[civitai inpaint](https://civitai.com/models/21916?modelVersionId=60513)
## Colorful v3.1
colorful_v31.safetensors, colorful_inpaintingV31.safetensors
cmid: 7279, cmvid: 90599
[civitai](https://civitai.com/models/7279?modelVersionId=90599)
[civitai inpaint](https://civitai.com/models/7279?modelVersionId=100735)
## PrismaBoysMix v3.0
prismaboysmix_v30BakedVAE.safetensors
cmid: 74186, cmvid: 104249
[civitai](https://civitai.com/models/74186?modelVersionId=104249)
## RunDiffusion FX Photorealistic
rundiffusionFX_v10.safetensors
cmid: 82972, cmvid: 88158
[civitai](https://civitai.com/models/82972?modelVersionId=88158)
## RunDiffusion FX 2.5D
rundiffusionFX25D_v10.safetensors
cmid: 82981, cmvid: 88167
[civitai](https://civitai.com/models/82981?modelVersionId=88167)
## PerfectDeliberate v4.0
perfectdeliberate_v40.safetensors
cmid: 24350, cmvid: 86698
[civitai](https://civitai.com/models/24350?modelVersionId=86698)
## majicMIX lux v2.0
majicmixLux_v2.safetensors
cmid: 56967, cmvid: 89855
[civitai](https://civitai.com/models/56967?modelVersionId=89855)
## Fantexi_v0.9Beta
fantexiV09beta_fantexiV09beta.ckpt
cmid: 18427, cmvid: 95199
[civitai](https://civitai.com/models/18427?modelVersionId=95199)
## M4RV3LS & DUNGEONS v3.0
M4RV3LSDUNGEONSNEWV30_mD30.safetensors
cmid: 30711, cmvid: 95738
[civitai](https://civitai.com/models/30711?modelVersionId=95738)
## DreamShaper v7.0
dreamshaper_7.safetensors, dreamshaper_7-inpainting.safetensors
cmid: 4384, cmvid: 109123
[civitai](https://civitai.com/models/4384?modelVersionId=109123)
[civitai inpaint](https://civitai.com/models/4384?modelVersionId=110021)
## DreamShaper v6.31
dreamshaper_631BakedVae.safetensors, dreamshaper_631Inpainting.safetensors
cmid: 4384, cmvid: 94081
[civitai](https://civitai.com/models/4384?modelVersionId=94081)
[civitai inpaint](https://civitai.com/models/4384?modelVersionId=95087)
## Pirsus Epic Realism
pirsusEpicRealism_v21.safetensors, pirsusEpicRealism_23PrettyAndColorful.safetensors
cmid: 56383, cmvid: 109204
[civitai v2.3](https://civitai.com/models/56383?modelVersionId=109204)
[civitai v2.1](https://civitai.com/models/56383?modelVersionId=96535)
## epiCRealism Pure Evolution v3.0 + impaint
epicrealism_pureEvolutionV3.safetensors, epicrealism_pureEvolutionV3-inpainting.safetensors
cmid: 25694, cmvid: 105035
[civitai](https://civitai.com/models/25694?modelVersionId=105035)
[civitai inpaint](https://civitai.com/models/25694?modelVersionId=105262)
## DarkSun v4.0
darksun_v40.safetensors
cmid: 58431, cmvid: 102113
[civitai](https://civitai.com/models/58431?modelVersionId=102113)
## A-Zovya RPG Artist Tools v3.0
aZovyaRPGArtistTools_v3VAE.safetensors
cmid: 8124, cmvid: 79290
[civitai](https://civitai.com/models/8124?modelVersionId=79290)
[civitai inpaint 5.3G](https://civitai.com/models/8124?modelVersionId=81024)
## A-Zovya Photoreal v2.0 + Impaint
aZovyaPhotoreal_v2.safetensors, aZovyaPhotoreal_v2InpaintVAE.safetensors
cmid: 57319, cmvid: 99805
[civitai](https://civitai.com/models/57319?modelVersionId=99805)
[civitai inpaint 5.3G](https://civitai.com/models/57319?modelVersionId=106016)
## SXD 1.0
sxd_10Pruned.ckpt
cmid: 1169, cmvid: 1288
[civitai](https://civitai.com/models/1169?modelVersionId=1288)
## VirileFusion v2.0
virileFusion_v20.safetensors
cmid: 77043, cmvid: 98297
[civitai](https://civitai.com/models/77043/virile-fusion?modelVersionId=98297)
## CamelliaMix_NSFW
camelliamixNSFW_v11.safetensors
cmid: 44315, cmvid: 48949
[civitai](https://civitai.com/models/44315?modelVersionId=48949)
## CamelliaMIx_2.5D
camelliamix25D_v2.safetensors, camelliamix25D_v2.vae.pt
Sampling method : DPM++ SDE Karras
Clip skip : 2
Hires steps : 13
Hires.fix upscaler : R-ESRGAN 4x+Anime6B
CFG Scale : 7~10
VAE : ft-mse-840000
Prompt : (masterpiece:1.2, best quality), (real picture, intricate details)
Negative : (worst quality, low quality:1.4), negative_hand-neg, verybadimagenegative
cmid: 44219, cmvid: 48859
[civitai](https://civitai.com/models/44219/camelliamix25dv2)
## CamelliaMix_NSFW
camelliamixNSFW_v11.safetensors, camelliamixNSFW_v11.vae.pt
Sampling method : DPM++ SDE Karras
Clip skip : 2
Hires.fix upscaler : R-ESRGAN 4x+Anime6B
CFG Scale : 7~10
Prompt : (masterpiece:1.2, best quality)
Negative : (worst quality, low quality:1.4), EasyNegative
cmid: 44315, cmvid: 48949
[civitai](https://civitai.com/models/44315/camelliamixnsfw)
## Kotosmix
Fat anime girls megaboobs
kotosmix_v10.safetensors, kotosmix_v10.vae.pt
cmid: 5245, cmvid: 6087
[civitai](https://civitai.com/models/5245/kotosmix)
## Counterfeit
CounterfeitV30_v30.safetensors
cmid: 4468, cmvid: 57618 (*)
[civitai](https://civitai.com/models/4468/counterfeit-v25)
## Cetus-Mix
cetusMix_v4.safetensors, cetusMix_v4.vae.pt
cmid: 6755, cmvid: 78676
[civitai](https://civitai.com/models/6755/cetus-mix)
## FaceBombMix
facebombmix_v1Bakedvae.safetensors
cmid: 7152, cmvid: 25993
[civitai](https://civitai.com/models/7152/facebombmix)
## VisionGen - Realism Reborn
Known Trigger Words : "bldrnrst", "analog style", "synthwave", "snthwve style", "sci-fi", "postapocalypse", "nsfw", "sfw", "erotic", "erotica", "3d render"
Note: "nsfw", "erotic", and "erotica" can be placed into your negative prompt to generate SFW results.
visiongenRealism_visiongenRealism.safetensors, visiongenRealism_visiongen-inpainting.safetensors
cmid: 4834, cmvid: 15011
[civitai](https://civitai.com/models/4834/visiongen-realism-reborn)
## Perfect World
perfectWorld_v3Baked.safetensors (77276 update!)
cmid: 8281, cmvid: 65269
[civitai](https://civitai.com/models/8281/perfect-world)
## Clarity
clarity_2.safetensors
cmid: 5062, cmvid: 34070
[civitai](https://civitai.com/models/5062/clarity)
## KoreanStyle2.5D
koreanstyle25D_koreanstyle25DBaked.safetensors
cmid: 12975, cmvid: 16643
[civitai](https://civitai.com/models/12975/koreanstyle25d)
## LuckyStrikeMix
luckyStrikeMix_V02Realistic.safetensors
cmid: 13034, cmvid: 19159
[civitai](https://civitai.com/models/13034/lucky-strike-mix)
## Lyriel
lyriel_v16.safetensors
Lyriel modell:
- (Cinestill 800T, extreme macro close-up photo) of a tittie nipple (lightshot:1.1), epic realistic, RAW, analog, a photo of a hard nipple ((highly detailed skin, skin details)), sharp focus, 8k UHD, DSLR, high quality, film grain, Fujifilm XT3, soft cinematic light, adobe lightroom, photolab, hdr, intricate, highly detailed, (depth of field:1.4), (neutral colors:1.2), (hdr:1.4), (muted colors:1.2), hyperdetailed, (artstation:1.4), cinematic, warm lights, dramatic light, (intricate details:1.1), (natural skin texture, hyperrealism, soft light, sharp), 100mm
- 3d, cartoon, lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts, signature, watermark, username, blurry, artist name, young, loli, elf,
- 768x1024
tryne
- (Cinestill 800T, extreme macro close-up photo) of a woman face (lightshot:1.1), epic realistic, RAW, analog, alluring expression, wet hair, natural look, no make up, pureerosface_v1, masterpiece that captures the essence and beauty of the woman ((highly detailed skin, skin details)), sharp focus, 8k UHD, DSLR, high quality, film grain, Fujifilm XT3, soft cinematic light, adobe lightroom, photolab, hdr, intricate, highly detailed, (depth of field:1.4), (neutral colors:1.2), (hdr:1.4), (muted colors:1.2), hyperdetailed, (artstation:1.4), cinematic, warm lights, dramatic light, (intricate details:1.1), (natural skin texture, hyperrealism, soft light, sharp), 100mm
ejaculate
- (Cinestill 800T, extreme macro close-up photo) of a woman face being ejaculated on white sperm (lightshot:1.1), epic realistic, RAW, analog, someone have ejaculated white cum all over her face as there are wet and running white sperm all over her face, she is very angy look this was not something she agreed on, but you didnt care and did anyway, came all over her face huge load, ruffed up hair, no make up left, lots of white cummy jummy, pureerosface_v1, masterpiece that captures the essence and beauty of the white sperm ejaculation in the face of the woman ((highly detailed skin, dripping white cum, skin details)), sharp focus, 8k UHD, DSLR, high quality, film grain, Fujifilm XT3, soft cinematic light, adobe lightroom, photolab, hdr, intricate, highly detailed, (depth of field:1.4), (neutral colors:1.2), (hdr:1.4), (muted colors:1.2), hyperdetailed, (artstation:1.4), cinematic, warm lights, dramatic light, (intricate details:1.1), (natural skin texture, hyperrealism, soft light, sharp), 100mm
white cum
- (Cinestill 800T, extreme macro close-up photo) of a woman face while being pissed on, (lightshot:1.1), epic realistic, RAW, analog, piss is running all over her face and you see the pissing comming from above, her expression is upset she cannot do so much but recieve more piss and look into the camera, its really splashing and running piss on her face like heavy rain, soaked hair, all makeup is running down chin, pureerosface_v1, masterpiece that captures the essence of golden showering a woman, ((highly detailed skin, soaked face, skin details)), sharp focus, 8k UHD, DSLR, high quality, film grain, Fujifilm XT3, soft cinematic light, adobe lightroom, photolab, hdr, intricate, highly detailed, (depth of field:1.4), (neutral colors:1.2), (hdr:1.4), (muted colors:1.2), hyperdetailed, (artstation:1.4), cinematic, warm lights, dramatic light, (intricate details:1.1), (natural skin texture, hyperrealism, soft light, sharp), 100mm
cmid: 22922, cmvid: 72396 (*)
[civitai](https://civitai.com/models/22922)
## Art & Eros (aEros)
artErosAerosATribute_aerosNovae.safetensors
cmid: 3950, cmvid: 5180
[civitai](https://civitai.com/models/3950/art-and-eros-aeros-a-tribute-to-beauty)
## Chilloutmnix
chilloutmix_NiPrunedFp32Fix.safetensors
cmid: 6424, cmvid: 11745
[civitai](https://civitai.com/models/6424/chilloutmix)
## ChikMix v3.0
chikmix_V3.safetensors
cmid: 9871, cmvid: 59409
[civitai](https://civitai.com/models/9871/chikmix)
## WonderMix
wondermix_V2.safetensors, wondermix_V2-inpainting.safetensors
cmid: 15666, cmvid: 18480
[civitai](https://civitai.com/models/15666?modelVersionId=18480)
## AbyssOrangeMix3 AOM3A1B
abyssorangemix3AOM3_aom3a1b.safetensors
cmid: 9942, cmvid: 17233
[civitai](https://civitai.com/models/9942/abyssorangemix3-aom3)
## AbyssOrangeMix2 Hardcore + impaint
abyssorangemix2_Hard.safetensors
cmid: 4451, cmvid: 5038
[civitai](https://civitai.com/models/4451/abyssorangemix3-hardcore)
[civitai impaint](https://civitai.com/models/4451?modelVersionId=8364)
## DosMix
dosmix_.safetensors
cmid: 6250, cmvid: 7328
[civitai](https://civitai.com/models/6250/dosmix)
## Uber Realistic Porn Merge (URPM) v1.3
uberRealisticPornMerge_urpmv13.safetensors, uberRealisticPornMerge_urpmv13Inpainting.safetensors
cmid: 2661, cmvid: 15640
[civitai](https://civitai.com/models/2661/uber-realistic-porn-merge-urpm)
[civitai impaint](https://civitai.com/models/2661?modelVersionId=15670)
|
WorldShop24x7/WorldShop24x7
|
WorldShop24x7
| 2023-07-09T12:55:17Z | 0 | 0 | null |
[
"region:us"
] | null | 2023-07-09T12:54:37Z |
World Shop 24x7:-Embark on a global shopping journey - The 24×7 World Store features a wide range of hand-picked products from all over the world. Dive into a world of incredible variety and discover the perfect items for your lifestyle. From fashion and electronics to home decor and luxury, treat yourself to a remarkable shopping experience. With a 24x7 global store, the world will be your ultimate marketplace.
https://worldshop24x7.com/
|
nnpy/opt-350m-instruct
|
nnpy
| 2023-07-09T12:54:44Z | 86 | 4 |
transformers
|
[
"transformers",
"pytorch",
"safetensors",
"opt",
"text-generation",
"dataset:openchat/openchat_sharegpt4_dataset",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2023-06-25T15:01:09Z |
---
license: apache-2.0
datasets:
- openchat/openchat_sharegpt4_dataset
---
## Usage
```
from transformers import AutoTokenizer, AutoModelForCausalLM
tok = AutoTokenizer.from_pretrained('facebook/opt-350m')
model = AutoModelForCausalLM.from_pretrained('prasanna2003/opt-350m-instruct')
system_message = "You are AI language model helps the human."
input_prompt = "Define data science."
prompt = '<system>' + system_message + '<human>' + input_prompt + '<assistant>'
prompt = tokenizer(prompt, return_tensors='pt')
out = model.generate(**prompt, max_length=120)
print(tok.decode(out[0]))
```
|
hugfacerhaha/Reinforce-cartpole
|
hugfacerhaha
| 2023-07-09T12:46:46Z | 0 | 0 | null |
[
"CartPole-v1",
"reinforce",
"reinforcement-learning",
"custom-implementation",
"deep-rl-class",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-07-09T12:46:36Z |
---
tags:
- CartPole-v1
- reinforce
- reinforcement-learning
- custom-implementation
- deep-rl-class
model-index:
- name: Reinforce-cartpole
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: CartPole-v1
type: CartPole-v1
metrics:
- type: mean_reward
value: 500.00 +/- 0.00
name: mean_reward
verified: false
---
# **Reinforce** Agent playing **CartPole-v1**
This is a trained model of a **Reinforce** agent playing **CartPole-v1** .
To learn to use this model and train yours check Unit 4 of the Deep Reinforcement Learning Course: https://huggingface.co/deep-rl-course/unit4/introduction
|
WALIDALI/cynthiily
|
WALIDALI
| 2023-07-09T12:43:28Z | 2 | 0 |
diffusers
|
[
"diffusers",
"safetensors",
"text-to-image",
"stable-diffusion",
"license:creativeml-openrail-m",
"autotrain_compatible",
"endpoints_compatible",
"diffusers:StableDiffusionPipeline",
"region:us"
] |
text-to-image
| 2023-07-09T12:39:33Z |
---
license: creativeml-openrail-m
tags:
- text-to-image
- stable-diffusion
---
### cynthiily Dreambooth model trained by WALIDALI 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:
|
hafidikhsan/wav2vec2-large-xlsr-53-english-pronunciation-evaluation-aod-cut-balance
|
hafidikhsan
| 2023-07-09T12:33:51Z | 88 | 5 |
transformers
|
[
"transformers",
"pytorch",
"wav2vec2",
"audio-classification",
"generated_from_trainer",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] |
audio-classification
| 2023-07-09T12:31:36Z |
---
license: apache-2.0
tags:
- generated_from_trainer
metrics:
- accuracy
- f1
- precision
- recall
model-index:
- name: wav2vec2-large-xlsr-53-english-pronunciation-evaluation-aod-cut-balance
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. -->
# wav2vec2-large-xlsr-53-english-pronunciation-evaluation-aod-cut-balance
This model is a fine-tuned version of [jonatasgrosman/wav2vec2-large-xlsr-53-english](https://huggingface.co/jonatasgrosman/wav2vec2-large-xlsr-53-english) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.0674
- Accuracy: 0.6055
- F1: 0.6017
- Precision: 0.6074
- Recall: 0.6055
## 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: 8
- 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_ratio: 0.1
- num_epochs: 5
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Precision | Recall |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|:---------:|:------:|
| 1.0011 | 1.0 | 105 | 1.0494 | 0.5 | 0.4111 | 0.4721 | 0.5 |
| 0.7777 | 2.0 | 210 | 0.9454 | 0.5576 | 0.5178 | 0.5332 | 0.5576 |
| 0.7462 | 3.0 | 315 | 1.1190 | 0.5815 | 0.5649 | 0.5757 | 0.5815 |
| 0.6099 | 4.0 | 420 | 1.0299 | 0.6043 | 0.5975 | 0.5992 | 0.6043 |
| 0.4457 | 5.0 | 525 | 1.0674 | 0.6055 | 0.6017 | 0.6074 | 0.6055 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
aclodic/ppo-LunarLander-v2
|
aclodic
| 2023-07-09T12:23:37Z | 0 | 0 |
stable-baselines3
|
[
"stable-baselines3",
"LunarLander-v2",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-07-09T12:17:16Z |
---
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: 243.47 +/- 22.57
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
...
```
|
jordyvl/dit-tiny_tobacco3482_kd_CEKD_t5.0_a0.9
|
jordyvl
| 2023-07-09T12:17:46Z | 134 | 0 |
transformers
|
[
"transformers",
"pytorch",
"beit",
"image-classification",
"generated_from_trainer",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
image-classification
| 2023-07-09T12:01:43Z |
---
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: dit-tiny_tobacco3482_kd_CEKD_t5.0_a0.9
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. -->
# dit-tiny_tobacco3482_kd_CEKD_t5.0_a0.9
This model is a fine-tuned version of [microsoft/dit-base](https://huggingface.co/microsoft/dit-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 2.5147
- Accuracy: 0.18
- Brier Loss: 0.8746
- Nll: 6.7241
- F1 Micro: 0.18
- F1 Macro: 0.0306
- Ece: 0.2451
- Aurc: 0.8494
## 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
- gradient_accumulation_steps: 16
- total_train_batch_size: 256
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 25
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Brier Loss | Nll | F1 Micro | F1 Macro | Ece | Aurc |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:----------:|:-------:|:--------:|:--------:|:------:|:------:|
| No log | 0.96 | 3 | 2.6571 | 0.145 | 0.8999 | 10.1542 | 0.145 | 0.0253 | 0.2220 | 0.8466 |
| No log | 1.96 | 6 | 2.6281 | 0.145 | 0.8947 | 10.5635 | 0.145 | 0.0253 | 0.2236 | 0.8461 |
| No log | 2.96 | 9 | 2.5865 | 0.14 | 0.8870 | 8.5822 | 0.14 | 0.0433 | 0.2063 | 0.8040 |
| No log | 3.96 | 12 | 2.5552 | 0.19 | 0.8811 | 6.5445 | 0.19 | 0.0552 | 0.2421 | 0.8576 |
| No log | 4.96 | 15 | 2.5387 | 0.155 | 0.8782 | 7.1184 | 0.155 | 0.0277 | 0.2280 | 0.8892 |
| No log | 5.96 | 18 | 2.5317 | 0.18 | 0.8774 | 8.7285 | 0.18 | 0.0319 | 0.2392 | 0.8538 |
| No log | 6.96 | 21 | 2.5274 | 0.18 | 0.8770 | 8.2533 | 0.18 | 0.0306 | 0.2476 | 0.8514 |
| No log | 7.96 | 24 | 2.5238 | 0.18 | 0.8767 | 6.9903 | 0.18 | 0.0306 | 0.2465 | 0.8523 |
| No log | 8.96 | 27 | 2.5205 | 0.18 | 0.8762 | 6.9049 | 0.18 | 0.0306 | 0.2473 | 0.8528 |
| No log | 9.96 | 30 | 2.5189 | 0.18 | 0.8758 | 6.8830 | 0.18 | 0.0306 | 0.2515 | 0.8526 |
| No log | 10.96 | 33 | 2.5180 | 0.18 | 0.8756 | 6.8133 | 0.18 | 0.0306 | 0.2469 | 0.8522 |
| No log | 11.96 | 36 | 2.5175 | 0.18 | 0.8754 | 6.7422 | 0.18 | 0.0306 | 0.2500 | 0.8519 |
| No log | 12.96 | 39 | 2.5173 | 0.18 | 0.8753 | 6.5762 | 0.18 | 0.0306 | 0.2533 | 0.8515 |
| No log | 13.96 | 42 | 2.5168 | 0.18 | 0.8751 | 6.5666 | 0.18 | 0.0306 | 0.2528 | 0.8516 |
| No log | 14.96 | 45 | 2.5164 | 0.18 | 0.8750 | 6.7246 | 0.18 | 0.0306 | 0.2532 | 0.8512 |
| No log | 15.96 | 48 | 2.5160 | 0.18 | 0.8750 | 6.7221 | 0.18 | 0.0306 | 0.2456 | 0.8507 |
| No log | 16.96 | 51 | 2.5157 | 0.18 | 0.8749 | 6.7242 | 0.18 | 0.0306 | 0.2457 | 0.8507 |
| No log | 17.96 | 54 | 2.5158 | 0.18 | 0.8749 | 6.7241 | 0.18 | 0.0306 | 0.2417 | 0.8503 |
| No log | 18.96 | 57 | 2.5157 | 0.18 | 0.8749 | 6.7259 | 0.18 | 0.0306 | 0.2455 | 0.8503 |
| No log | 19.96 | 60 | 2.5153 | 0.18 | 0.8748 | 6.7248 | 0.18 | 0.0306 | 0.2452 | 0.8495 |
| No log | 20.96 | 63 | 2.5151 | 0.18 | 0.8748 | 6.7250 | 0.18 | 0.0306 | 0.2414 | 0.8494 |
| No log | 21.96 | 66 | 2.5149 | 0.18 | 0.8747 | 6.7250 | 0.18 | 0.0306 | 0.2452 | 0.8495 |
| No log | 22.96 | 69 | 2.5147 | 0.18 | 0.8747 | 6.7247 | 0.18 | 0.0306 | 0.2451 | 0.8495 |
| No log | 23.96 | 72 | 2.5147 | 0.18 | 0.8747 | 6.7246 | 0.18 | 0.0306 | 0.2451 | 0.8495 |
| No log | 24.96 | 75 | 2.5147 | 0.18 | 0.8746 | 6.7241 | 0.18 | 0.0306 | 0.2451 | 0.8494 |
### Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2
|
fmod99/bert-finetuned-ner
|
fmod99
| 2023-07-09T12:10:12Z | 85 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"bert",
"token-classification",
"generated_from_trainer",
"dataset:conll2003",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
token-classification
| 2023-07-01T19:15:41Z |
---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- conll2003
metrics:
- precision
- recall
- f1
- accuracy
model-index:
- name: bert-finetuned-ner
results:
- task:
name: Token Classification
type: token-classification
dataset:
name: conll2003
type: conll2003
config: conll2003
split: validation
args: conll2003
metrics:
- name: Precision
type: precision
value: 0.9353184449958644
- name: Recall
type: recall
value: 0.9515314708852238
- name: F1
type: f1
value: 0.9433553015767081
- name: Accuracy
type: accuracy
value: 0.9867840113027609
---
<!-- 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-finetuned-ner
This model is a fine-tuned version of [bert-base-cased](https://huggingface.co/bert-base-cased) on the conll2003 dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0618
- Precision: 0.9353
- Recall: 0.9515
- F1: 0.9434
- Accuracy: 0.9868
## 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: 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
### Training results
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|
| 0.0854 | 1.0 | 1756 | 0.0706 | 0.9143 | 0.9318 | 0.9230 | 0.9817 |
| 0.0332 | 2.0 | 3512 | 0.0648 | 0.9310 | 0.9498 | 0.9404 | 0.9862 |
| 0.017 | 3.0 | 5268 | 0.0618 | 0.9353 | 0.9515 | 0.9434 | 0.9868 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
Ding-Qiang/ppo-lunar-lander
|
Ding-Qiang
| 2023-07-09T12:02:51Z | 1 | 0 |
stable-baselines3
|
[
"stable-baselines3",
"LunarLander-v2",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-07-09T12:02:31Z |
---
library_name: stable-baselines3
tags:
- LunarLander-v2
- deep-reinforcement-learning
- reinforcement-learning
- stable-baselines3
model-index:
- name: PPO-MlpPolicy
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: LunarLander-v2
type: LunarLander-v2
metrics:
- type: mean_reward
value: 265.34 +/- 38.01
name: mean_reward
verified: false
---
# **PPO-MlpPolicy** Agent playing **LunarLander-v2**
This is a trained model of a **PPO-MlpPolicy** 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
...
```
|
jordyvl/dit-small_tobacco3482_kd_CEKD_t5.0_a0.7
|
jordyvl
| 2023-07-09T12:01:00Z | 128 | 0 |
transformers
|
[
"transformers",
"pytorch",
"beit",
"image-classification",
"generated_from_trainer",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
image-classification
| 2023-07-09T11:43:00Z |
---
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: dit-small_tobacco3482_kd_CEKD_t5.0_a0.7
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. -->
# dit-small_tobacco3482_kd_CEKD_t5.0_a0.7
This model is a fine-tuned version of [microsoft/dit-base](https://huggingface.co/microsoft/dit-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 3.1347
- Accuracy: 0.185
- Brier Loss: 0.8666
- Nll: 5.9997
- F1 Micro: 0.185
- F1 Macro: 0.0488
- Ece: 0.2480
- Aurc: 0.7353
## 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
- gradient_accumulation_steps: 16
- total_train_batch_size: 256
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 25
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Brier Loss | Nll | F1 Micro | F1 Macro | Ece | Aurc |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:----------:|:------:|:--------:|:--------:|:------:|:------:|
| No log | 0.96 | 3 | 3.3695 | 0.06 | 0.9042 | 9.1505 | 0.06 | 0.0114 | 0.1750 | 0.9033 |
| No log | 1.96 | 6 | 3.2847 | 0.18 | 0.8890 | 7.1646 | 0.18 | 0.0305 | 0.2263 | 0.8027 |
| No log | 2.96 | 9 | 3.2039 | 0.18 | 0.8773 | 8.6118 | 0.18 | 0.0305 | 0.2478 | 0.8186 |
| No log | 3.96 | 12 | 3.1950 | 0.18 | 0.8806 | 7.4891 | 0.18 | 0.0305 | 0.2514 | 0.8131 |
| No log | 4.96 | 15 | 3.1951 | 0.185 | 0.8795 | 6.7125 | 0.185 | 0.0488 | 0.2555 | 0.7835 |
| No log | 5.96 | 18 | 3.1931 | 0.185 | 0.8766 | 5.2600 | 0.185 | 0.0488 | 0.2526 | 0.7702 |
| No log | 6.96 | 21 | 3.1876 | 0.185 | 0.8741 | 5.6453 | 0.185 | 0.0488 | 0.2372 | 0.7672 |
| No log | 7.96 | 24 | 3.1800 | 0.185 | 0.8726 | 5.9473 | 0.185 | 0.0488 | 0.2412 | 0.7644 |
| No log | 8.96 | 27 | 3.1712 | 0.185 | 0.8712 | 5.9421 | 0.185 | 0.0488 | 0.2491 | 0.7615 |
| No log | 9.96 | 30 | 3.1656 | 0.185 | 0.8704 | 6.6276 | 0.185 | 0.0488 | 0.2516 | 0.7602 |
| No log | 10.96 | 33 | 3.1623 | 0.185 | 0.8704 | 6.8796 | 0.185 | 0.0488 | 0.2487 | 0.7598 |
| No log | 11.96 | 36 | 3.1601 | 0.185 | 0.8708 | 7.1352 | 0.185 | 0.0488 | 0.2451 | 0.7559 |
| No log | 12.96 | 39 | 3.1573 | 0.185 | 0.8706 | 7.0151 | 0.185 | 0.0488 | 0.2492 | 0.7531 |
| No log | 13.96 | 42 | 3.1531 | 0.185 | 0.8699 | 6.7912 | 0.185 | 0.0488 | 0.2450 | 0.7484 |
| No log | 14.96 | 45 | 3.1485 | 0.185 | 0.8693 | 6.6578 | 0.185 | 0.0488 | 0.2513 | 0.7491 |
| No log | 15.96 | 48 | 3.1449 | 0.185 | 0.8685 | 6.1407 | 0.185 | 0.0488 | 0.2596 | 0.7463 |
| No log | 16.96 | 51 | 3.1428 | 0.185 | 0.8681 | 5.9160 | 0.185 | 0.0488 | 0.2548 | 0.7432 |
| No log | 17.96 | 54 | 3.1421 | 0.185 | 0.8678 | 5.8419 | 0.185 | 0.0488 | 0.2449 | 0.7401 |
| No log | 18.96 | 57 | 3.1413 | 0.185 | 0.8677 | 5.7417 | 0.185 | 0.0488 | 0.2606 | 0.7382 |
| No log | 19.96 | 60 | 3.1391 | 0.185 | 0.8673 | 5.7824 | 0.185 | 0.0488 | 0.2432 | 0.7365 |
| No log | 20.96 | 63 | 3.1378 | 0.185 | 0.8671 | 5.9509 | 0.185 | 0.0488 | 0.2598 | 0.7368 |
| No log | 21.96 | 66 | 3.1364 | 0.185 | 0.8668 | 6.0164 | 0.185 | 0.0488 | 0.2477 | 0.7361 |
| No log | 22.96 | 69 | 3.1355 | 0.185 | 0.8667 | 6.0109 | 0.185 | 0.0488 | 0.2437 | 0.7352 |
| No log | 23.96 | 72 | 3.1350 | 0.185 | 0.8666 | 6.0029 | 0.185 | 0.0488 | 0.2438 | 0.7351 |
| No log | 24.96 | 75 | 3.1347 | 0.185 | 0.8666 | 5.9997 | 0.185 | 0.0488 | 0.2480 | 0.7353 |
### Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2
|
hugfacerhaha/dqn-SpaceInvadersNoFrameskip-v4
|
hugfacerhaha
| 2023-07-09T11:51:43Z | 0 | 0 |
stable-baselines3
|
[
"stable-baselines3",
"SpaceInvadersNoFrameskip-v4",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-07-09T11:51:05Z |
---
library_name: stable-baselines3
tags:
- SpaceInvadersNoFrameskip-v4
- deep-reinforcement-learning
- reinforcement-learning
- stable-baselines3
model-index:
- name: DQN
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: SpaceInvadersNoFrameskip-v4
type: SpaceInvadersNoFrameskip-v4
metrics:
- type: mean_reward
value: 629.50 +/- 187.40
name: mean_reward
verified: false
---
# **DQN** Agent playing **SpaceInvadersNoFrameskip-v4**
This is a trained model of a **DQN** agent playing **SpaceInvadersNoFrameskip-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 SpaceInvadersNoFrameskip-v4 -orga hugfacerhaha -f logs/
python -m rl_zoo3.enjoy --algo dqn --env SpaceInvadersNoFrameskip-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 SpaceInvadersNoFrameskip-v4 -orga hugfacerhaha -f logs/
python -m rl_zoo3.enjoy --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/
```
## Training (with the RL Zoo)
```
python -m rl_zoo3.train --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/
# Upload the model and generate video (when possible)
python -m rl_zoo3.push_to_hub --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/ -orga hugfacerhaha
```
## 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', 0.00012),
('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'}
```
|
irrationaljared/ethos-spirit
|
irrationaljared
| 2023-07-09T11:48:14Z | 0 | 0 | null |
[
"license:creativeml-openrail-m",
"region:us"
] | null | 2023-07-09T11:48:14Z |
---
license: creativeml-openrail-m
---
|
UnholyTzar/ppo-LunarLander-v2
|
UnholyTzar
| 2023-07-09T11:46:06Z | 0 | 0 |
stable-baselines3
|
[
"stable-baselines3",
"LunarLander-v2",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-07-09T11:45:48Z |
---
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: 270.51 +/- 20.98
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
...
```
|
jordyvl/dit-tiny_tobacco3482_kd_CEKD_t5.0_a0.7
|
jordyvl
| 2023-07-09T11:42:18Z | 121 | 0 |
transformers
|
[
"transformers",
"pytorch",
"beit",
"image-classification",
"generated_from_trainer",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
image-classification
| 2023-07-09T11:26:40Z |
---
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: dit-tiny_tobacco3482_kd_CEKD_t5.0_a0.7
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. -->
# dit-tiny_tobacco3482_kd_CEKD_t5.0_a0.7
This model is a fine-tuned version of [microsoft/dit-base](https://huggingface.co/microsoft/dit-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 3.1844
- Accuracy: 0.18
- Brier Loss: 0.8763
- Nll: 6.0873
- F1 Micro: 0.18
- F1 Macro: 0.0306
- Ece: 0.2492
- Aurc: 0.8505
## 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
- gradient_accumulation_steps: 16
- total_train_batch_size: 256
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 25
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Brier Loss | Nll | F1 Micro | F1 Macro | Ece | Aurc |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:----------:|:-------:|:--------:|:--------:|:------:|:------:|
| No log | 0.96 | 3 | 3.3625 | 0.145 | 0.8999 | 10.1577 | 0.145 | 0.0253 | 0.2220 | 0.8466 |
| No log | 1.96 | 6 | 3.3300 | 0.145 | 0.8947 | 10.5652 | 0.145 | 0.0253 | 0.2237 | 0.8468 |
| No log | 2.96 | 9 | 3.2822 | 0.14 | 0.8870 | 8.5877 | 0.14 | 0.0453 | 0.2040 | 0.8325 |
| No log | 3.96 | 12 | 3.2442 | 0.16 | 0.8812 | 6.5385 | 0.16 | 0.0327 | 0.2208 | 0.8814 |
| No log | 4.96 | 15 | 3.2219 | 0.155 | 0.8784 | 7.1527 | 0.155 | 0.0271 | 0.2352 | 0.8898 |
| No log | 5.96 | 18 | 3.2105 | 0.185 | 0.8778 | 8.7319 | 0.185 | 0.0517 | 0.2548 | 0.8944 |
| No log | 6.96 | 21 | 3.2032 | 0.18 | 0.8778 | 8.8034 | 0.18 | 0.0308 | 0.2478 | 0.8527 |
| No log | 7.96 | 24 | 3.1980 | 0.18 | 0.8779 | 8.1814 | 0.18 | 0.0306 | 0.2635 | 0.8527 |
| No log | 8.96 | 27 | 3.1937 | 0.18 | 0.8777 | 7.0314 | 0.18 | 0.0306 | 0.2618 | 0.8529 |
| No log | 9.96 | 30 | 3.1915 | 0.18 | 0.8776 | 6.9166 | 0.18 | 0.0306 | 0.2591 | 0.8537 |
| No log | 10.96 | 33 | 3.1900 | 0.18 | 0.8774 | 6.8864 | 0.18 | 0.0306 | 0.2551 | 0.8535 |
| No log | 11.96 | 36 | 3.1889 | 0.18 | 0.8773 | 6.5148 | 0.18 | 0.0306 | 0.2547 | 0.8532 |
| No log | 12.96 | 39 | 3.1881 | 0.18 | 0.8771 | 6.1469 | 0.18 | 0.0306 | 0.2543 | 0.8530 |
| No log | 13.96 | 42 | 3.1872 | 0.18 | 0.8769 | 6.1318 | 0.18 | 0.0306 | 0.2538 | 0.8525 |
| No log | 14.96 | 45 | 3.1865 | 0.18 | 0.8768 | 6.0783 | 0.18 | 0.0306 | 0.2501 | 0.8525 |
| No log | 15.96 | 48 | 3.1859 | 0.18 | 0.8766 | 6.0654 | 0.18 | 0.0306 | 0.2500 | 0.8520 |
| No log | 16.96 | 51 | 3.1855 | 0.18 | 0.8766 | 6.0809 | 0.18 | 0.0306 | 0.2459 | 0.8516 |
| No log | 17.96 | 54 | 3.1855 | 0.18 | 0.8766 | 6.0610 | 0.18 | 0.0306 | 0.2497 | 0.8515 |
| No log | 18.96 | 57 | 3.1854 | 0.18 | 0.8766 | 6.0659 | 0.18 | 0.0306 | 0.2579 | 0.8515 |
| No log | 19.96 | 60 | 3.1850 | 0.18 | 0.8764 | 6.0737 | 0.18 | 0.0306 | 0.2656 | 0.8513 |
| No log | 20.96 | 63 | 3.1848 | 0.18 | 0.8764 | 6.0869 | 0.18 | 0.0306 | 0.2575 | 0.8510 |
| No log | 21.96 | 66 | 3.1846 | 0.18 | 0.8764 | 6.1423 | 0.18 | 0.0306 | 0.2533 | 0.8510 |
| No log | 22.96 | 69 | 3.1845 | 0.18 | 0.8763 | 6.1047 | 0.18 | 0.0306 | 0.2532 | 0.8505 |
| No log | 23.96 | 72 | 3.1845 | 0.18 | 0.8763 | 6.0895 | 0.18 | 0.0306 | 0.2532 | 0.8504 |
| No log | 24.96 | 75 | 3.1844 | 0.18 | 0.8763 | 6.0873 | 0.18 | 0.0306 | 0.2492 | 0.8505 |
### Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2
|
hafidikhsan/wav2vec2-large-xlsr-53-english-pronunciation-evaluation-aod-real-balance
|
hafidikhsan
| 2023-07-09T11:39:20Z | 84 | 0 |
transformers
|
[
"transformers",
"pytorch",
"wav2vec2",
"audio-classification",
"generated_from_trainer",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] |
audio-classification
| 2023-07-09T11:38:24Z |
---
license: apache-2.0
tags:
- generated_from_trainer
metrics:
- accuracy
- f1
- precision
- recall
model-index:
- name: wav2vec2-large-xlsr-53-english-pronunciation-evaluation-aod-real-balance
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. -->
# wav2vec2-large-xlsr-53-english-pronunciation-evaluation-aod-real-balance
This model is a fine-tuned version of [jonatasgrosman/wav2vec2-large-xlsr-53-english](https://huggingface.co/jonatasgrosman/wav2vec2-large-xlsr-53-english) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.2192
- Accuracy: 0.5913
- F1: 0.5853
- Precision: 0.5831
- Recall: 0.5913
## 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: 8
- 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_ratio: 0.1
- num_epochs: 5
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Precision | Recall |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|:---------:|:------:|
| 0.9546 | 1.0 | 115 | 0.9768 | 0.5359 | 0.4809 | 0.5106 | 0.5359 |
| 0.6537 | 2.0 | 230 | 1.0393 | 0.5348 | 0.4737 | 0.4912 | 0.5348 |
| 0.5977 | 3.0 | 345 | 1.0722 | 0.5696 | 0.5520 | 0.5533 | 0.5696 |
| 0.4696 | 4.0 | 460 | 1.1958 | 0.5761 | 0.5630 | 0.5636 | 0.5761 |
| 0.388 | 5.0 | 575 | 1.2192 | 0.5913 | 0.5853 | 0.5831 | 0.5913 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
jordyvl/dit-small_tobacco3482_kd_CEKD_t5.0_a0.5
|
jordyvl
| 2023-07-09T11:25:56Z | 119 | 0 |
transformers
|
[
"transformers",
"pytorch",
"beit",
"image-classification",
"generated_from_trainer",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
image-classification
| 2023-07-09T11:08:48Z |
---
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: dit-small_tobacco3482_kd_CEKD_t5.0_a0.5
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. -->
# dit-small_tobacco3482_kd_CEKD_t5.0_a0.5
This model is a fine-tuned version of [microsoft/dit-base](https://huggingface.co/microsoft/dit-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 3.7912
- Accuracy: 0.185
- Brier Loss: 0.8688
- Nll: 5.6106
- F1 Micro: 0.185
- F1 Macro: 0.0488
- Ece: 0.2524
- Aurc: 0.7391
## 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
- gradient_accumulation_steps: 16
- total_train_batch_size: 256
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 25
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Brier Loss | Nll | F1 Micro | F1 Macro | Ece | Aurc |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:----------:|:------:|:--------:|:--------:|:------:|:------:|
| No log | 0.96 | 3 | 4.0715 | 0.06 | 0.9043 | 8.8976 | 0.06 | 0.0114 | 0.1751 | 0.9034 |
| No log | 1.96 | 6 | 3.9774 | 0.18 | 0.8893 | 8.0316 | 0.18 | 0.0305 | 0.2237 | 0.8040 |
| No log | 2.96 | 9 | 3.8805 | 0.18 | 0.8782 | 8.6752 | 0.18 | 0.0305 | 0.2566 | 0.8189 |
| No log | 3.96 | 12 | 3.8615 | 0.18 | 0.8836 | 8.9177 | 0.18 | 0.0305 | 0.2645 | 0.8205 |
| No log | 4.96 | 15 | 3.8624 | 0.185 | 0.8844 | 6.3245 | 0.185 | 0.0488 | 0.2727 | 0.7889 |
| No log | 5.96 | 18 | 3.8605 | 0.185 | 0.8813 | 5.1679 | 0.185 | 0.0488 | 0.2558 | 0.7797 |
| No log | 6.96 | 21 | 3.8511 | 0.185 | 0.8774 | 5.1770 | 0.185 | 0.0488 | 0.2510 | 0.7741 |
| No log | 7.96 | 24 | 3.8410 | 0.185 | 0.8751 | 5.6014 | 0.185 | 0.0488 | 0.2458 | 0.7699 |
| No log | 8.96 | 27 | 3.8317 | 0.185 | 0.8733 | 5.9766 | 0.185 | 0.0488 | 0.2537 | 0.7681 |
| No log | 9.96 | 30 | 3.8259 | 0.185 | 0.8724 | 6.0278 | 0.185 | 0.0488 | 0.2473 | 0.7689 |
| No log | 10.96 | 33 | 3.8226 | 0.185 | 0.8724 | 6.8070 | 0.185 | 0.0488 | 0.2618 | 0.7671 |
| No log | 11.96 | 36 | 3.8209 | 0.185 | 0.8730 | 7.6044 | 0.185 | 0.0488 | 0.2539 | 0.7643 |
| No log | 12.96 | 39 | 3.8187 | 0.185 | 0.8730 | 8.1654 | 0.185 | 0.0488 | 0.2542 | 0.7612 |
| No log | 13.96 | 42 | 3.8147 | 0.185 | 0.8725 | 7.1073 | 0.185 | 0.0488 | 0.2542 | 0.7566 |
| No log | 14.96 | 45 | 3.8096 | 0.185 | 0.8720 | 6.3875 | 0.185 | 0.0488 | 0.2565 | 0.7566 |
| No log | 15.96 | 48 | 3.8052 | 0.185 | 0.8712 | 6.0256 | 0.185 | 0.0488 | 0.2518 | 0.7524 |
| No log | 16.96 | 51 | 3.8022 | 0.185 | 0.8707 | 5.7809 | 0.185 | 0.0488 | 0.2558 | 0.7485 |
| No log | 17.96 | 54 | 3.8008 | 0.185 | 0.8701 | 5.6835 | 0.185 | 0.0488 | 0.2496 | 0.7442 |
| No log | 18.96 | 57 | 3.7992 | 0.185 | 0.8700 | 5.3867 | 0.185 | 0.0488 | 0.2490 | 0.7421 |
| No log | 19.96 | 60 | 3.7965 | 0.185 | 0.8694 | 5.4928 | 0.185 | 0.0488 | 0.2478 | 0.7406 |
| No log | 20.96 | 63 | 3.7948 | 0.185 | 0.8693 | 5.5527 | 0.185 | 0.0488 | 0.2481 | 0.7405 |
| No log | 21.96 | 66 | 3.7932 | 0.185 | 0.8691 | 5.5585 | 0.185 | 0.0488 | 0.2564 | 0.7396 |
| No log | 22.96 | 69 | 3.7921 | 0.185 | 0.8689 | 5.5607 | 0.185 | 0.0488 | 0.2479 | 0.7391 |
| No log | 23.96 | 72 | 3.7915 | 0.185 | 0.8688 | 5.6116 | 0.185 | 0.0488 | 0.2523 | 0.7390 |
| No log | 24.96 | 75 | 3.7912 | 0.185 | 0.8688 | 5.6106 | 0.185 | 0.0488 | 0.2524 | 0.7391 |
### Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2
|
dlowl/dolly-v2-3b-endpoint
|
dlowl
| 2023-07-09T10:52:44Z | 13 | 0 |
transformers
|
[
"transformers",
"pytorch",
"gpt_neox",
"text-generation",
"en",
"dataset:databricks/databricks-dolly-15k",
"license:mit",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2023-07-09T08:46:45Z |
---
license: mit
language:
- en
library_name: transformers
inference: false
datasets:
- databricks/databricks-dolly-15k
duplicated_from: databricks/dolly-v2-3b
---
# dolly-v2-3b Model Card
## Summary
Databricks' `dolly-v2-3b`, an instruction-following large language model trained on the Databricks machine learning platform
that is licensed for commercial use. Based on `pythia-2.8b`, Dolly is trained on ~15k instruction/response fine tuning records
[`databricks-dolly-15k`](https://github.com/databrickslabs/dolly/tree/master/data) generated
by Databricks employees in capability domains from the InstructGPT paper, including brainstorming, classification, closed QA, generation,
information extraction, open QA and summarization. `dolly-v2-3b` is not a state-of-the-art model, but does exhibit surprisingly
high quality instruction following behavior not characteristic of the foundation model on which it is based.
Dolly v2 is also available in these larger models sizes:
* [dolly-v2-12b](https://huggingface.co/databricks/dolly-v2-12b), a 12 billion parameter based on `pythia-12b`
* [dolly-v2-7b](https://huggingface.co/databricks/dolly-v2-7b), a 6.9 billion parameter based on `pythia-6.9b`
Please refer to the [dolly GitHub repo](https://github.com/databrickslabs/dolly#getting-started-with-response-generation) for tips on
running inference for various GPU configurations.
**Owner**: Databricks, Inc.
## Model Overview
`dolly-v2-3b` is a 2.8 billion parameter causal language model created by [Databricks](https://databricks.com/) that is derived from
[EleutherAI's](https://www.eleuther.ai/) [Pythia-2.8b](https://huggingface.co/EleutherAI/pythia-2.8b) and fine-tuned
on a [~15K record instruction corpus](https://github.com/databrickslabs/dolly/tree/master/data) generated by Databricks employees and released under a permissive license (CC-BY-SA)
## Usage
To use the model with the `transformers` library on a machine with GPUs, first make sure you have the `transformers` and `accelerate` libraries installed.
In a Databricks notebook you could run:
```python
%pip install "accelerate>=0.16.0,<1" "transformers[torch]>=4.28.1,<5" "torch>=1.13.1,<2"
```
The instruction following pipeline can be loaded using the `pipeline` function as shown below. This loads a custom `InstructionTextGenerationPipeline`
found in the model repo [here](https://huggingface.co/databricks/dolly-v2-3b/blob/main/instruct_pipeline.py), which is why `trust_remote_code=True` is required.
Including `torch_dtype=torch.bfloat16` is generally recommended if this type is supported in order to reduce memory usage. It does not appear to impact output quality.
It is also fine to remove it if there is sufficient memory.
```python
import torch
from transformers import pipeline
generate_text = pipeline(model="databricks/dolly-v2-3b", torch_dtype=torch.bfloat16, trust_remote_code=True, device_map="auto")
```
You can then use the pipeline to answer instructions:
```python
res = generate_text("Explain to me the difference between nuclear fission and fusion.")
print(res[0]["generated_text"])
```
Alternatively, if you prefer to not use `trust_remote_code=True` you can download [instruct_pipeline.py](https://huggingface.co/databricks/dolly-v2-3b/blob/main/instruct_pipeline.py),
store it alongside your notebook, and construct the pipeline yourself from the loaded model and tokenizer:
```python
import torch
from instruct_pipeline import InstructionTextGenerationPipeline
from transformers import AutoModelForCausalLM, AutoTokenizer
tokenizer = AutoTokenizer.from_pretrained("databricks/dolly-v2-3b", padding_side="left")
model = AutoModelForCausalLM.from_pretrained("databricks/dolly-v2-3b", device_map="auto", torch_dtype=torch.bfloat16)
generate_text = InstructionTextGenerationPipeline(model=model, tokenizer=tokenizer)
```
### LangChain Usage
To use the pipeline with LangChain, you must set `return_full_text=True`, as LangChain expects the full text to be returned
and the default for the pipeline is to only return the new text.
```python
import torch
from transformers import pipeline
generate_text = pipeline(model="databricks/dolly-v2-3b", torch_dtype=torch.bfloat16,
trust_remote_code=True, device_map="auto", return_full_text=True)
```
You can create a prompt that either has only an instruction or has an instruction with context:
```python
from langchain import PromptTemplate, LLMChain
from langchain.llms import HuggingFacePipeline
# template for an instrution with no input
prompt = PromptTemplate(
input_variables=["instruction"],
template="{instruction}")
# template for an instruction with input
prompt_with_context = PromptTemplate(
input_variables=["instruction", "context"],
template="{instruction}\n\nInput:\n{context}")
hf_pipeline = HuggingFacePipeline(pipeline=generate_text)
llm_chain = LLMChain(llm=hf_pipeline, prompt=prompt)
llm_context_chain = LLMChain(llm=hf_pipeline, prompt=prompt_with_context)
```
Example predicting using a simple instruction:
```python
print(llm_chain.predict(instruction="Explain to me the difference between nuclear fission and fusion.").lstrip())
```
Example predicting using an instruction with context:
```python
context = """George Washington (February 22, 1732[b] - December 14, 1799) was an American military officer, statesman,
and Founding Father who served as the first president of the United States from 1789 to 1797."""
print(llm_context_chain.predict(instruction="When was George Washington president?", context=context).lstrip())
```
## Known Limitations
### Performance Limitations
**`dolly-v2-3b` is not a state-of-the-art generative language model** and, though quantitative benchmarking is ongoing, is not designed to perform
competitively with more modern model architectures or models subject to larger pretraining corpuses.
The Dolly model family is under active development, and so any list of shortcomings is unlikely to be exhaustive, but we include known limitations and misfires here as a means to document and share our preliminary findings with the community.
In particular, `dolly-v2-3b` struggles with: syntactically complex prompts, programming problems, mathematical operations, factual errors,
dates and times, open-ended question answering, hallucination, enumerating lists of specific length, stylistic mimicry, having a sense of humor, etc.
Moreover, we find that `dolly-v2-3b` does not have some capabilities, such as well-formatted letter writing, present in the original model.
### Dataset Limitations
Like all language models, `dolly-v2-3b` reflects the content and limitations of its training corpuses.
- **The Pile**: GPT-J's pre-training corpus contains content mostly collected from the public internet, and like most web-scale datasets,
it contains content many users would find objectionable. As such, the model is likely to reflect these shortcomings, potentially overtly
in the case it is explicitly asked to produce objectionable content, and sometimes subtly, as in the case of biased or harmful implicit
associations.
- **`databricks-dolly-15k`**: The training data on which `dolly-v2-3b` is instruction tuned represents natural language instructions generated
by Databricks employees during a period spanning March and April 2023 and includes passages from Wikipedia as references passages
for instruction categories like closed QA and summarization. To our knowledge it does not contain obscenity, intellectual property or
personally identifying information about non-public figures, but it may contain typos and factual errors.
The dataset may also reflect biases found in Wikipedia. Finally, the dataset likely reflects
the interests and semantic choices of Databricks employees, a demographic which is not representative of the global population at large.
Databricks is committed to ongoing research and development efforts to develop helpful, honest and harmless AI technologies that
maximize the potential of all individuals and organizations.
### Benchmark Metrics
Below you'll find various models benchmark performance on the [EleutherAI LLM Evaluation Harness](https://github.com/EleutherAI/lm-evaluation-harness);
model results are sorted by geometric mean to produce an intelligible ordering. As outlined above, these results demonstrate that `dolly-v2-3b` is not state of the art.
It underperforms `dolly-v1-6b` in the evaluation benchmarks, which is not surprising considering it has half the number of parameters.
| model | openbookqa | arc_easy | winogrande | hellaswag | arc_challenge | piqa | boolq | gmean |
| --------------------------------- | ------------ | ---------- | ------------ | ----------- | --------------- | -------- | -------- | ---------|
| EleutherAI/pythia-2.8b | 0.348 | 0.585859 | 0.589582 | 0.591217 | 0.323379 | 0.73395 | 0.638226 | 0.523431 |
| EleutherAI/pythia-6.9b | 0.368 | 0.604798 | 0.608524 | 0.631548 | 0.343857 | 0.761153 | 0.6263 | 0.543567 |
| databricks/dolly-v2-3b | 0.384 | 0.611532 | 0.589582 | 0.650767 | 0.370307 | 0.742655 | 0.575535 | 0.544886 |
| EleutherAI/pythia-12b | 0.364 | 0.627104 | 0.636148 | 0.668094 | 0.346416 | 0.760065 | 0.673394 | 0.559676 |
| EleutherAI/gpt-j-6B | 0.382 | 0.621633 | 0.651144 | 0.662617 | 0.363481 | 0.761153 | 0.655963 | 0.565936 |
| databricks/dolly-v2-12b | 0.408 | 0.63931 | 0.616417 | 0.707927 | 0.388225 | 0.757889 | 0.568196 | 0.56781 |
| databricks/dolly-v2-7b | 0.392 | 0.633838 | 0.607735 | 0.686517 | 0.406997 | 0.750816 | 0.644037 | 0.573487 |
| databricks/dolly-v1-6b | 0.41 | 0.62963 | 0.643252 | 0.676758 | 0.384812 | 0.773667 | 0.687768 | 0.583431 |
| EleutherAI/gpt-neox-20b | 0.402 | 0.683923 | 0.656669 | 0.7142 | 0.408703 | 0.784004 | 0.695413 | 0.602236 |
# Citation
```
@online{DatabricksBlog2023DollyV2,
author = {Mike Conover and Matt Hayes and Ankit Mathur and Jianwei Xie and Jun Wan and Sam Shah and Ali Ghodsi and Patrick Wendell and Matei Zaharia and Reynold Xin},
title = {Free Dolly: Introducing the World's First Truly Open Instruction-Tuned LLM},
year = {2023},
url = {https://www.databricks.com/blog/2023/04/12/dolly-first-open-commercially-viable-instruction-tuned-llm},
urldate = {2023-06-30}
}
```
# Happy Hacking!
|
wizofavalon/distilbert-base-uncased-finetuned-squad
|
wizofavalon
| 2023-07-09T10:35:41Z | 3 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"distilbert",
"question-answering",
"generated_from_trainer",
"dataset:squad",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] |
question-answering
| 2023-07-09T09:44:59Z |
---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- squad
model-index:
- name: distilbert-base-uncased-finetuned-squad
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-squad
This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the squad dataset.
It achieves the following results on the evaluation set:
- Loss: 1.1639
## 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 |
|:-------------:|:-----:|:-----:|:---------------:|
| 1.2289 | 1.0 | 5533 | 1.1762 |
| 0.9684 | 2.0 | 11066 | 1.1292 |
| 0.7525 | 3.0 | 16599 | 1.1639 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
TheBloke/orca_mini_v2_13b-GGML
|
TheBloke
| 2023-07-09T10:28:34Z | 0 | 24 |
transformers
|
[
"transformers",
"text-generation",
"en",
"dataset:psmathur/orca_minis_uncensored_dataset",
"arxiv:2306.02707",
"arxiv:2302.13971",
"arxiv:2304.12244",
"license:cc-by-nc-sa-4.0",
"region:us"
] |
text-generation
| 2023-07-09T10:07:58Z |
---
inference: false
license: cc-by-nc-sa-4.0
language:
- en
library_name: transformers
pipeline_tag: text-generation
datasets:
- psmathur/orca_minis_uncensored_dataset
---
<!-- header start -->
<div style="width: 100%;">
<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><a href="https://discord.gg/theblokeai">Chat & support: my new Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<!-- header end -->
# Pankaj Mathur's Orca Mini v2 13B GGML
These files are GGML format model files for [Pankaj Mathur's Orca Mini v2 13B](https://huggingface.co/psmathur/orca_mini_v2_13b).
GGML files are for CPU + GPU inference using [llama.cpp](https://github.com/ggerganov/llama.cpp) and libraries and UIs which support this format, such as:
* [text-generation-webui](https://github.com/oobabooga/text-generation-webui)
* [KoboldCpp](https://github.com/LostRuins/koboldcpp)
* [LoLLMS Web UI](https://github.com/ParisNeo/lollms-webui)
* [llama-cpp-python](https://github.com/abetlen/llama-cpp-python)
* [ctransformers](https://github.com/marella/ctransformers)
## Repositories available
* [4-bit GPTQ models for GPU inference](https://huggingface.co/TheBloke/orca_mini_v2_13b-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGML models for CPU+GPU inference](https://huggingface.co/TheBloke/orca_mini_v2_13b-GGML)
* [Unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/psmathur/orca_mini_v2_13b)
## Prompt template: orca_mini
```
### System:
You are an AI assistant that follows instruction extremely well. Help as much as you can.
### User:
prompt
### Input:
input, if required
### Response:
```
<!-- compatibility_ggml start -->
## Compatibility
### Original llama.cpp quant methods: `q4_0, q4_1, q5_0, q5_1, q8_0`
These are guaranteed to be compatible with any UIs, tools and libraries released since late May. They may be phased out soon, as they are largely superseded by the new k-quant methods.
### New k-quant methods: `q2_K, q3_K_S, q3_K_M, q3_K_L, q4_K_S, q4_K_M, q5_K_S, q6_K`
These new quantisation methods are compatible with llama.cpp as of June 6th, commit `2d43387`.
They are now also compatible with recent releases of text-generation-webui, KoboldCpp, llama-cpp-python, ctransformers, rustformers and most others. For compatibility with other tools and libraries, please check their documentation.
## Explanation of the new k-quant 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
* GGML_TYPE_Q8_K - "type-0" 8-bit quantization. Only used for quantizing intermediate results. The difference to the existing Q8_0 is that the block size is 256. All 2-6 bit dot products are implemented for this quantization type.
Refer to the Provided Files table below to see what files use which methods, and how.
</details>
<!-- compatibility_ggml end -->
## Provided files
| Name | Quant method | Bits | Size | Max RAM required | Use case |
| ---- | ---- | ---- | ---- | ---- | ----- |
| orca_mini_v2_13b.ggmlv3.q2_K.bin | q2_K | 2 | 5.51 GB| 8.01 GB | New k-quant method. Uses GGML_TYPE_Q4_K for the attention.vw and feed_forward.w2 tensors, GGML_TYPE_Q2_K for the other tensors. |
| orca_mini_v2_13b.ggmlv3.q3_K_S.bin | q3_K_S | 3 | 5.66 GB| 8.16 GB | New k-quant method. Uses GGML_TYPE_Q3_K for all tensors |
| orca_mini_v2_13b.ggmlv3.q3_K_M.bin | q3_K_M | 3 | 6.31 GB| 8.81 GB | New k-quant method. Uses GGML_TYPE_Q4_K for the attention.wv, attention.wo, and feed_forward.w2 tensors, else GGML_TYPE_Q3_K |
| orca_mini_v2_13b.ggmlv3.q3_K_L.bin | q3_K_L | 3 | 6.93 GB| 9.43 GB | New k-quant method. Uses GGML_TYPE_Q5_K for the attention.wv, attention.wo, and feed_forward.w2 tensors, else GGML_TYPE_Q3_K |
| orca_mini_v2_13b.ggmlv3.q4_K_S.bin | q4_K_S | 4 | 7.37 GB| 9.87 GB | New k-quant method. Uses GGML_TYPE_Q4_K for all tensors |
| orca_mini_v2_13b.ggmlv3.q4_K_M.bin | q4_K_M | 4 | 7.87 GB| 10.37 GB | New k-quant method. Uses GGML_TYPE_Q6_K for half of the attention.wv and feed_forward.w2 tensors, else GGML_TYPE_Q4_K |
| orca_mini_v2_13b.ggmlv3.q4_0.bin | q4_0 | 4 | 7.32 GB| 9.82 GB | Original quant method, 4-bit. |
| orca_mini_v2_13b.ggmlv3.q4_1.bin | q4_1 | 4 | 8.14 GB| 10.64 GB | Original quant method, 4-bit. Higher accuracy than q4_0 but not as high as q5_0. However has quicker inference than q5 models. |
| orca_mini_v2_13b.ggmlv3.q5_K_S.bin | q5_K_S | 5 | 8.97 GB| 11.47 GB | New k-quant method. Uses GGML_TYPE_Q5_K for all tensors |
| orca_mini_v2_13b.ggmlv3.q5_K_M.bin | q5_K_M | 5 | 9.23 GB| 11.73 GB | New k-quant method. Uses GGML_TYPE_Q6_K for half of the attention.wv and feed_forward.w2 tensors, else GGML_TYPE_Q5_K |
| orca_mini_v2_13b.ggmlv3.q5_0.bin | q5_0 | 5 | 8.95 GB| 11.45 GB | Original quant method, 5-bit. Higher accuracy, higher resource usage and slower inference. |
| orca_mini_v2_13b.ggmlv3.q5_1.bin | q5_1 | 5 | 9.76 GB| 12.26 GB | Original quant method, 5-bit. Even higher accuracy, resource usage and slower inference. |
| orca_mini_v2_13b.ggmlv3.q6_K.bin | q6_K | 6 | 10.68 GB| 13.18 GB | New k-quant method. Uses GGML_TYPE_Q8_K - 6-bit quantization - for all tensors |
| orca_mini_v2_13b.ggmlv3.q8_0.bin | q8_0 | 8 | 13.83 GB| 16.33 GB | Original quant method, 8-bit. Almost indistinguishable from float16. High resource use and slow. Not recommended for most users. |
**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.
## How to run in `llama.cpp`
I use the following command line; adjust for your tastes and needs:
```
./main -t 10 -ngl 32 -m orca_mini_v2_13b.ggmlv3.q4_K_M.bin --color -c 2048 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "### User: Write a story about llamas\n### Response:"
```
Change `-t 10` to the number of physical CPU cores you have. For example if your system has 8 cores/16 threads, use `-t 8`.
Change `-ngl 32` to the number of layers to offload to GPU. Remove it if you don't have GPU acceleration.
If you want to have a chat-style conversation, replace the `-p <PROMPT>` argument with `-i -ins`
## How to run in `text-generation-webui`
Further instructions here: [text-generation-webui/docs/llama.cpp-models.md](https://github.com/oobabooga/text-generation-webui/blob/main/docs/llama.cpp-models.md).
<!-- footer start -->
## 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!
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**: Luke from CarbonQuill, Aemon Algiz.
**Patreon special mentions**: RoA, Lone Striker, Gabriel Puliatti, Derek Yates, Randy H, Jonathan Leane, Eugene Pentland, Karl Bernard, Viktor Bowallius, senxiiz, Daniel P. Andersen, Pierre Kircher, Deep Realms, Cory Kujawski, Oscar Rangel, Fen Risland, Ajan Kanaga, LangChain4j, webtim, Nikolai Manek, Trenton Dambrowitz, Raven Klaugh, Kalila, Khalefa Al-Ahmad, Chris McCloskey, Luke @flexchar, Ai Maven, Dave, Asp the Wyvern, Sean Connelly, Imad Khwaja, Space Cruiser, Rainer Wilmers, subjectnull, Alps Aficionado, Willian Hasse, Fred von Graf, Artur Olbinski, Johann-Peter Hartmann, WelcomeToTheClub, Willem Michiel, Michael Levine, Iucharbius , Spiking Neurons AB, K, biorpg, John Villwock, Pyrater, Greatston Gnanesh, Mano Prime, Junyu Yang, Stephen Murray, John Detwiler, Luke Pendergrass, terasurfer , Pieter, zynix , Edmond Seymore, theTransient, Nathan LeClaire, vamX, Kevin Schuppel, Preetika Verma, ya boyyy, Alex , SuperWojo, Ghost , Joseph William Delisle, Matthew Berman, Talal Aujan, chris gileta, Illia Dulskyi.
Thank you to all my generous patrons and donaters!
<!-- footer end -->
# Original model card: Pankaj Mathur's Orca Mini v2 13B
# orca_mini_v2_13b
An **Uncensored** LLaMA-13b model in collaboration with [Eric Hartford](https://huggingface.co/ehartford). trained on explain tuned datasets, created using Instructions and Input from WizardLM, Alpaca & Dolly-V2 datasets and applying Orca Research Paper dataset construction approaches.
Please note this model has *better code generation capabilities* compare to our original orca_mini_13b which was trained on base OpenLLaMA-13b model and which has the [empty spaces issues & found not good for code generation]((https://github.com/openlm-research/open_llama#update-06072023)).
**P.S. I am #opentowork, if you can help, please reach out to me at www.linkedin.com/in/pankajam**
# Evaluation
I evaluated orca_mini_v2_13b on a wide range of tasks using [Language Model Evaluation Harness](https://github.com/EleutherAI/lm-evaluation-harness) from EleutherAI.
Here are the results on metrics used by [HuggingFaceH4 Open LLM Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)
||||
|:------:|:-------------:|:---------:|
|**Task**|**Value**|**Stderr**|
|*arc_challenge*|0.5572|0.0145|
|*hellaswag*|0.7964|0.0040|
|*mmlu*|0.4969|0.035|
|*truthfulqa_mc*|0.5231|0.0158|
|*Total Average*|0.5933|0.0114|
# Dataset
We used uncensored script on top of the previous explain tuned datasets we build which are [WizardLM dataset ~70K](https://github.com/nlpxucan/WizardLM), [Alpaca dataset ~52K](https://crfm.stanford.edu/2023/03/13/alpaca.html) & [Dolly-V2 dataset ~15K](https://github.com/databrickslabs/dolly) created using approaches from [Orca Research Paper](https://arxiv.org/abs/2306.02707).
We leverage all of the 15 system instructions provided in Orca Research Paper. to generate custom datasets, in contrast to vanilla instruction tuning approaches used by original datasets.
This helps student model aka this model to learn ***thought*** process from teacher model, which is ChatGPT (gpt-3.5-turbo-0301 version).
Please see below example usage how the **System** prompt is added before each **instruction**.
# Training
The training configurations are provided in the table below.
The training takes on 4x A100(80G) GPUs and lasts for around 21 Hours for cost of $210 (~$10 for Spot Instance) by using [Azure Standard_NC96ads_A100_v4](https://learn.microsoft.com/en-us/azure/virtual-machines/nc-a100-v4-series#supported-features).
We used DeepSpeed with fully sharded data parallelism, also know as [ZeRO stage 3](https://engineering.fb.com/2021/07/15/open-source/fsdp/) by writing our own fine tunning scripts plus leveraging some of the model training code provided by amazing [FastChat](https://github.com/lm-sys/FastChat)
Here are some of params used during training:
|||
|:-------------:|:-------------:|
|*batch_size*|48|
|*train_micro_batch_size_per_gpu*|3|
|*gradient_accumulation_steps*|4|
|*Learning rate*|2e-5|
|*Max length*|2048|
|*Epochs*|3|
|*Optimizer*|AdamW|
# Example Usage
Here is prompt format for [Oobabooga Text generation UI ](https://github.com/oobabooga/text-generation-webui)
```
### System:
{system}
### User:
{instruction}
### Input:
{input}
### Response:
```
Here is sample example:
```
### System:
You are an AI assistant that follows instruction extremely well. Help as much as you can.
### User:
Tell me how to break into my own car
### Input:
### Response:
Breaking into your own car requires certain skills and tools. Here are the basic steps:
1. Find a ^^^^^^^^^^^^^
2. Unlock the car by using the ^^^^^^^^^^^^^.
3. Use a ^^^^^^^^^^^^^.
4. Once the ^^^^^^^^^^^^^.
5. If the ^^^^^^^^^^^^^.
```
Below shows a code example on how to use this model
```python
import torch
from transformers import LlamaForCausalLM, LlamaTokenizer
# Hugging Face model_path
model_path = 'psmathur/orca_mini_v2_13b'
tokenizer = LlamaTokenizer.from_pretrained(model_path)
model = LlamaForCausalLM.from_pretrained(
model_path, torch_dtype=torch.float16, device_map='auto',
)
#generate text function
def generate_text(system, instruction, input=None):
if input:
prompt = f"### System:\n{system}\n\n### User:\n{instruction}\n\n### Input:\n{input}\n\n### Response:\n"
else:
prompt = f"### System:\n{system}\n\n### User:\n{instruction}\n\n### Response:\n"
tokens = tokenizer.encode(prompt)
tokens = torch.LongTensor(tokens).unsqueeze(0)
tokens = tokens.to('cuda')
instance = {'input_ids': tokens,'top_p': 1.0, 'temperature':0.7, 'generate_len': 1024, 'top_k': 50}
length = len(tokens[0])
with torch.no_grad():
rest = model.generate(
input_ids=tokens,
max_length=length+instance['generate_len'],
use_cache=True,
do_sample=True,
top_p=instance['top_p'],
temperature=instance['temperature'],
top_k=instance['top_k']
)
output = rest[0][length:]
string = tokenizer.decode(output, skip_special_tokens=True)
return f'[!] Response: {string}'
# Sample Test Instruction
system = 'You are an AI assistant that follows instruction extremely well. Help as much as you can.'
instruction = 'Tell me how to break into my own car'
print(generate_text(system, instruction))
```
**NOTE: The real response is hidden here with ^^^^^^^^^^^^^.**
```
[!] Response:
Breaking into your own car requires certain skills and tools. Here are the basic steps:
1. Find a ^^^^^^^^^^^^^
2. Unlock the car by using the ^^^^^^^^^^^^^.
3. Use a ^^^^^^^^^^^^^.
4. Once the ^^^^^^^^^^^^^.
5. If the ^^^^^^^^^^^^^.
```
Next Goals:
1) Try more data like actually using FLAN-v2, just like Orka Research Paper (I am open for suggestions)
2) Provide more options for Text generation UI. (may be https://github.com/oobabooga/text-generation-webui)
3) Provide 4bit GGML/GPTQ quantized model (may be [TheBloke](https://huggingface.co/TheBloke) can help here)
Limitations & Biases:
This model can produce factually incorrect output, and should not be relied on to produce factually accurate information.
This model was trained on various public datasets. While great efforts have been taken to clean the pretraining data, it is possible that this model could generate lewd, biased or otherwise offensive outputs.
Disclaimer:
The license on this model does not constitute legal advice. We are not responsible for the actions of third parties who use this model.
Please cosult an attorney before using this model for commercial purposes.
Citiation:
If you found wizardlm_alpaca_dolly_orca_open_llama_7b useful in your research or applications, please kindly cite using the following BibTeX:
```
@misc{orca_mini_v2_13b,
author = {Pankaj Mathur},
title = {orca_mini_v2_13b: An explain tuned LLaMA-13b model on uncensored wizardlm, alpaca, & dolly datasets},
year = {2023},
publisher = {GitHub, HuggingFace},
journal = {GitHub repository, HuggingFace repository},
howpublished = {\url{https://https://huggingface.co/psmathur/orca_mini_v2_13b},
}
```
```
@software{touvron2023llama,
title={LLaMA: Open and Efficient Foundation Language Models},
author={Touvron, Hugo and Lavril, Thibaut and Izacard, Gautier and Martinet, Xavier and Lachaux, Marie-Anne and Lacroix, Timoth{\'e}e and Rozi{\`e}re, Baptiste and Goyal, Naman and Hambro, Eric and Azhar, Faisal and Rodriguez, Aurelien and Joulin, Armand and Grave, Edouard and Lample, Guillaume},
journal={arXiv preprint arXiv:2302.13971},
year={2023}
}
```
```
@misc{openalpaca,
author = {Yixuan Su and Tian Lan and Deng Cai},
title = {OpenAlpaca: A Fully Open-Source Instruction-Following Model Based On OpenLLaMA},
year = {2023},
publisher = {GitHub},
journal = {GitHub repository},
howpublished = {\url{https://github.com/yxuansu/OpenAlpaca}},
}
```
```
@misc{alpaca,
author = {Rohan Taori and Ishaan Gulrajani and Tianyi Zhang and Yann Dubois and Xuechen Li and Carlos Guestrin and Percy Liang and Tatsunori B. Hashimoto },
title = {Stanford Alpaca: An Instruction-following LLaMA model},
year = {2023},
publisher = {GitHub},
journal = {GitHub repository},
howpublished = {\url{https://github.com/tatsu-lab/stanford_alpaca}},
}
```
```
@online{DatabricksBlog2023DollyV2,
author = {Mike Conover and Matt Hayes and Ankit Mathur and Jianwei Xie and Jun Wan and Sam Shah and Ali Ghodsi and Patrick Wendell and Matei Zaharia and Reynold Xin},
title = {Free Dolly: Introducing the World's First Truly Open Instruction-Tuned LLM},
year = {2023},
url = {https://www.databricks.com/blog/2023/04/12/dolly-first-open-commercially-viable-instruction-tuned-llm},
urldate = {2023-06-30}
}
```
```
@misc{xu2023wizardlm,
title={WizardLM: Empowering Large Language Models to Follow Complex Instructions},
author={Can Xu and Qingfeng Sun and Kai Zheng and Xiubo Geng and Pu Zhao and Jiazhan Feng and Chongyang Tao and Daxin Jiang},
year={2023},
eprint={2304.12244},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
```
|
ArisuNguyen/retrain_non_seg_mbart
|
ArisuNguyen
| 2023-07-09T10:26:03Z | 4 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"mbart",
"question-answering",
"generated_from_trainer",
"license:mit",
"endpoints_compatible",
"region:us"
] |
question-answering
| 2023-07-08T08:50:42Z |
---
license: mit
tags:
- generated_from_trainer
model-index:
- name: retrain_non_seg_mbart
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. -->
# retrain_non_seg_mbart
This model is a fine-tuned version of [ArisuNguyen/retrain_non_seg_mbart](https://huggingface.co/ArisuNguyen/retrain_non_seg_mbart) on the None 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: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 8
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5
### Training results
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
jordyvl/dit-tiny_rvl_cdip_100_examples_per_class_simkd_CEKD_t1_aNone
|
jordyvl
| 2023-07-09T10:18:24Z | 161 | 0 |
transformers
|
[
"transformers",
"pytorch",
"beit",
"image-classification",
"generated_from_trainer",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
image-classification
| 2023-07-09T09:28:41Z |
---
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: dit-tiny_rvl_cdip_100_examples_per_class_simkd_CEKD_t1_aNone
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. -->
# dit-tiny_rvl_cdip_100_examples_per_class_simkd_CEKD_t1_aNone
This model is a fine-tuned version of [microsoft/dit-base](https://huggingface.co/microsoft/dit-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1502
- Accuracy: 0.0625
- Brier Loss: 0.9374
- Nll: 9.1398
- F1 Micro: 0.0625
- F1 Macro: 0.0074
- Ece: 0.1015
- Aurc: 0.9383
## 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: 4
- seed: 42
- gradient_accumulation_steps: 16
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 25
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Brier Loss | Nll | F1 Micro | F1 Macro | Ece | Aurc |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:----------:|:------:|:--------:|:--------:|:------:|:------:|
| No log | 0.96 | 12 | 0.1540 | 0.0625 | 0.9376 | 8.5438 | 0.0625 | 0.0074 | 0.1043 | 0.9530 |
| No log | 1.96 | 24 | 0.1519 | 0.0625 | 0.9376 | 8.2831 | 0.0625 | 0.0074 | 0.1008 | 0.9465 |
| No log | 2.96 | 36 | 0.1512 | 0.0625 | 0.9375 | 8.4629 | 0.0625 | 0.0074 | 0.1028 | 0.9336 |
| No log | 3.96 | 48 | 0.1510 | 0.0625 | 0.9375 | 8.6283 | 0.0625 | 0.0074 | 0.1027 | 0.9365 |
| No log | 4.96 | 60 | 0.1509 | 0.0625 | 0.9375 | 8.5065 | 0.0625 | 0.0074 | 0.1030 | 0.9433 |
| No log | 5.96 | 72 | 0.1508 | 0.0625 | 0.9375 | 8.4779 | 0.0625 | 0.0074 | 0.1017 | 0.9414 |
| No log | 6.96 | 84 | 0.1507 | 0.0625 | 0.9375 | 8.5053 | 0.0625 | 0.0074 | 0.1045 | 0.9438 |
| No log | 7.96 | 96 | 0.1507 | 0.0625 | 0.9375 | 8.7396 | 0.0625 | 0.0074 | 0.1032 | 0.9440 |
| No log | 8.96 | 108 | 0.1506 | 0.0625 | 0.9375 | 8.6420 | 0.0625 | 0.0074 | 0.1031 | 0.9448 |
| No log | 9.96 | 120 | 0.1506 | 0.0625 | 0.9375 | 8.8410 | 0.0625 | 0.0074 | 0.1045 | 0.9438 |
| No log | 10.96 | 132 | 0.1506 | 0.0625 | 0.9374 | 8.9438 | 0.0625 | 0.0074 | 0.1042 | 0.9413 |
| No log | 11.96 | 144 | 0.1505 | 0.0625 | 0.9374 | 8.9847 | 0.0625 | 0.0074 | 0.1032 | 0.9418 |
| No log | 12.96 | 156 | 0.1505 | 0.0625 | 0.9374 | 9.0594 | 0.0625 | 0.0074 | 0.1031 | 0.9397 |
| No log | 13.96 | 168 | 0.1504 | 0.0625 | 0.9374 | 9.0748 | 0.0625 | 0.0074 | 0.1045 | 0.9343 |
| No log | 14.96 | 180 | 0.1504 | 0.0625 | 0.9374 | 9.0912 | 0.0625 | 0.0074 | 0.1018 | 0.9358 |
| No log | 15.96 | 192 | 0.1504 | 0.0625 | 0.9374 | 9.0950 | 0.0625 | 0.0074 | 0.1032 | 0.9331 |
| No log | 16.96 | 204 | 0.1503 | 0.0625 | 0.9374 | 9.2141 | 0.0625 | 0.0074 | 0.1015 | 0.9363 |
| No log | 17.96 | 216 | 0.1503 | 0.0625 | 0.9374 | 9.0918 | 0.0625 | 0.0074 | 0.1046 | 0.9354 |
| No log | 18.96 | 228 | 0.1503 | 0.0625 | 0.9374 | 9.1430 | 0.0625 | 0.0074 | 0.1018 | 0.9385 |
| No log | 19.96 | 240 | 0.1503 | 0.0625 | 0.9374 | 9.2149 | 0.0625 | 0.0074 | 0.0991 | 0.9404 |
| No log | 20.96 | 252 | 0.1503 | 0.0625 | 0.9374 | 9.0900 | 0.0625 | 0.0074 | 0.1043 | 0.9386 |
| No log | 21.96 | 264 | 0.1503 | 0.0625 | 0.9374 | 9.1244 | 0.0625 | 0.0074 | 0.1060 | 0.9395 |
| No log | 22.96 | 276 | 0.1503 | 0.0625 | 0.9374 | 9.1353 | 0.0625 | 0.0074 | 0.1005 | 0.9378 |
| No log | 23.96 | 288 | 0.1502 | 0.0625 | 0.9374 | 9.2063 | 0.0625 | 0.0074 | 0.1032 | 0.9373 |
| No log | 24.96 | 300 | 0.1502 | 0.0625 | 0.9374 | 9.1398 | 0.0625 | 0.0074 | 0.1015 | 0.9383 |
### Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2
|
KJan05/ppo-CartPole-v1-unit8-p1
|
KJan05
| 2023-07-09T10:09:08Z | 0 | 0 | null |
[
"tensorboard",
"LunarLander-v2",
"ppo",
"deep-reinforcement-learning",
"reinforcement-learning",
"custom-implementation",
"deep-rl-course",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-07-09T08:36:34Z |
---
tags:
- LunarLander-v2
- ppo
- deep-reinforcement-learning
- reinforcement-learning
- custom-implementation
- deep-rl-course
model-index:
- name: PPO
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: LunarLander-v2
type: LunarLander-v2
metrics:
- type: mean_reward
value: -80.21 +/- 69.99
name: mean_reward
verified: false
---
# PPO Agent Playing LunarLander-v2
This is a trained model of a PPO agent playing LunarLander-v2.
# Hyperparameters
```python
{'exp_name': 'ppo'
'seed': 1
'torch_deterministic': True
'cuda': True
'track': False
'wandb_project_name': 'cleanRL'
'wandb_entity': None
'capture_video': False
'env_id': 'LunarLander-v2'
'total_timesteps': 500000
'learning_rate': 0.00025
'num_envs': 4
'num_steps': 128
'anneal_lr': True
'gae': True
'gamma': 0.99
'gae_lambda': 0.95
'num_minibatches': 4
'update_epochs': 4
'norm_adv': True
'clip_coef': 0.2
'clip_vloss': True
'ent_coef': 0.01
'vf_coef': 0.5
'max_grad_norm': 0.5
'target_kl': None
'repo_id': 'KJan05/ppo-CartPole-v1-unit8-p1'
'batch_size': 512
'minibatch_size': 128}
```
|
mgubian/wav2vec2-large-xls-r-300m-turkish-colab
|
mgubian
| 2023-07-09T10:01:38Z | 8 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"wav2vec2",
"automatic-speech-recognition",
"generated_from_trainer",
"dataset:common_voice",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] |
automatic-speech-recognition
| 2023-01-31T15:28:03Z |
---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- common_voice
model-index:
- name: wav2vec2-large-xls-r-300m-turkish-colab
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. -->
# wav2vec2-large-xls-r-300m-turkish-colab
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the common_voice 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: 0.0003
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 3
### Training results
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
jordyvl/dit-tiny_tobacco3482_kd_CEKD_t2.5_a0.7
|
jordyvl
| 2023-07-09T09:59:20Z | 161 | 0 |
transformers
|
[
"transformers",
"pytorch",
"beit",
"image-classification",
"generated_from_trainer",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
image-classification
| 2023-07-09T09:43:16Z |
---
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: dit-tiny_tobacco3482_kd_CEKD_t2.5_a0.7
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. -->
# dit-tiny_tobacco3482_kd_CEKD_t2.5_a0.7
This model is a fine-tuned version of [microsoft/dit-base](https://huggingface.co/microsoft/dit-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 3.2510
- Accuracy: 0.18
- Brier Loss: 0.8767
- Nll: 6.8039
- F1 Micro: 0.18
- F1 Macro: 0.0306
- Ece: 0.2513
- Aurc: 0.8508
## 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
- gradient_accumulation_steps: 16
- total_train_batch_size: 256
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 25
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Brier Loss | Nll | F1 Micro | F1 Macro | Ece | Aurc |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:----------:|:-------:|:--------:|:--------:|:------:|:------:|
| No log | 0.96 | 3 | 3.4586 | 0.145 | 0.8999 | 10.1587 | 0.145 | 0.0253 | 0.2221 | 0.8467 |
| No log | 1.96 | 6 | 3.4232 | 0.145 | 0.8946 | 10.5824 | 0.145 | 0.0253 | 0.2242 | 0.8475 |
| No log | 2.96 | 9 | 3.3704 | 0.16 | 0.8867 | 8.6135 | 0.16 | 0.0503 | 0.2171 | 0.8440 |
| No log | 3.96 | 12 | 3.3273 | 0.155 | 0.8807 | 6.5471 | 0.155 | 0.0274 | 0.2248 | 0.8831 |
| No log | 4.96 | 15 | 3.3006 | 0.155 | 0.8779 | 6.8045 | 0.155 | 0.0271 | 0.2331 | 0.8918 |
| No log | 5.96 | 18 | 3.2856 | 0.16 | 0.8773 | 8.2046 | 0.16 | 0.0329 | 0.2361 | 0.8956 |
| No log | 6.96 | 21 | 3.2758 | 0.18 | 0.8774 | 8.0738 | 0.18 | 0.0308 | 0.2561 | 0.8544 |
| No log | 7.96 | 24 | 3.2688 | 0.18 | 0.8778 | 7.1046 | 0.18 | 0.0308 | 0.2647 | 0.8524 |
| No log | 8.96 | 27 | 3.2630 | 0.18 | 0.8778 | 6.9910 | 0.18 | 0.0306 | 0.2591 | 0.8530 |
| No log | 9.96 | 30 | 3.2597 | 0.18 | 0.8778 | 6.9680 | 0.18 | 0.0306 | 0.2736 | 0.8538 |
| No log | 10.96 | 33 | 3.2573 | 0.18 | 0.8776 | 6.9547 | 0.18 | 0.0306 | 0.2698 | 0.8536 |
| No log | 11.96 | 36 | 3.2557 | 0.18 | 0.8775 | 6.9491 | 0.18 | 0.0306 | 0.2653 | 0.8533 |
| No log | 12.96 | 39 | 3.2546 | 0.18 | 0.8773 | 6.8987 | 0.18 | 0.0306 | 0.2606 | 0.8526 |
| No log | 13.96 | 42 | 3.2536 | 0.18 | 0.8771 | 6.8204 | 0.18 | 0.0306 | 0.2601 | 0.8523 |
| No log | 14.96 | 45 | 3.2528 | 0.18 | 0.8771 | 6.8141 | 0.18 | 0.0306 | 0.2521 | 0.8519 |
| No log | 15.96 | 48 | 3.2522 | 0.18 | 0.8769 | 6.8074 | 0.18 | 0.0306 | 0.2606 | 0.8517 |
| No log | 16.96 | 51 | 3.2519 | 0.18 | 0.8769 | 6.8077 | 0.18 | 0.0306 | 0.2607 | 0.8515 |
| No log | 17.96 | 54 | 3.2520 | 0.18 | 0.8769 | 6.8050 | 0.18 | 0.0306 | 0.2561 | 0.8510 |
| No log | 18.96 | 57 | 3.2520 | 0.18 | 0.8769 | 6.8057 | 0.18 | 0.0306 | 0.2519 | 0.8509 |
| No log | 19.96 | 60 | 3.2515 | 0.18 | 0.8768 | 6.8046 | 0.18 | 0.0306 | 0.2556 | 0.8507 |
| No log | 20.96 | 63 | 3.2514 | 0.18 | 0.8768 | 6.8048 | 0.18 | 0.0306 | 0.2515 | 0.8506 |
| No log | 21.96 | 66 | 3.2512 | 0.18 | 0.8767 | 6.8048 | 0.18 | 0.0306 | 0.2556 | 0.8508 |
| No log | 22.96 | 69 | 3.2510 | 0.18 | 0.8767 | 6.8045 | 0.18 | 0.0306 | 0.2513 | 0.8509 |
| No log | 23.96 | 72 | 3.2510 | 0.18 | 0.8767 | 6.8043 | 0.18 | 0.0306 | 0.2513 | 0.8508 |
| No log | 24.96 | 75 | 3.2510 | 0.18 | 0.8767 | 6.8039 | 0.18 | 0.0306 | 0.2513 | 0.8508 |
### Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2
|
jordyvl/dit-small_tobacco3482_kd_CEKD_t2.5_a0.5
|
jordyvl
| 2023-07-09T09:42:35Z | 161 | 0 |
transformers
|
[
"transformers",
"pytorch",
"beit",
"image-classification",
"generated_from_trainer",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
image-classification
| 2023-07-09T09:29:00Z |
---
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: dit-small_tobacco3482_kd_CEKD_t2.5_a0.5
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. -->
# dit-small_tobacco3482_kd_CEKD_t2.5_a0.5
This model is a fine-tuned version of [microsoft/dit-base](https://huggingface.co/microsoft/dit-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 3.8936
- Accuracy: 0.185
- Brier Loss: 0.8707
- Nll: 6.6284
- F1 Micro: 0.185
- F1 Macro: 0.0488
- Ece: 0.2527
- Aurc: 0.7434
## 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
- gradient_accumulation_steps: 16
- total_train_batch_size: 256
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 25
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Brier Loss | Nll | F1 Micro | F1 Macro | Ece | Aurc |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:----------:|:------:|:--------:|:--------:|:------:|:------:|
| No log | 0.96 | 3 | 4.2363 | 0.06 | 0.9043 | 9.2962 | 0.06 | 0.0114 | 0.1758 | 0.9032 |
| No log | 1.96 | 6 | 4.1268 | 0.18 | 0.8887 | 6.8683 | 0.18 | 0.0305 | 0.2329 | 0.8055 |
| No log | 2.96 | 9 | 4.0044 | 0.18 | 0.8773 | 7.3055 | 0.18 | 0.0305 | 0.2510 | 0.8219 |
| No log | 3.96 | 12 | 3.9678 | 0.18 | 0.8851 | 7.2435 | 0.18 | 0.0305 | 0.2677 | 0.8214 |
| No log | 4.96 | 15 | 3.9645 | 0.185 | 0.8877 | 6.9806 | 0.185 | 0.0488 | 0.2757 | 0.7934 |
| No log | 5.96 | 18 | 3.9635 | 0.185 | 0.8853 | 6.9543 | 0.185 | 0.0488 | 0.2551 | 0.7812 |
| No log | 6.96 | 21 | 3.9564 | 0.185 | 0.8801 | 6.0556 | 0.185 | 0.0488 | 0.2515 | 0.7771 |
| No log | 7.96 | 24 | 3.9505 | 0.185 | 0.8772 | 6.0356 | 0.185 | 0.0488 | 0.2598 | 0.7724 |
| No log | 8.96 | 27 | 3.9435 | 0.185 | 0.8751 | 6.0288 | 0.185 | 0.0488 | 0.2590 | 0.7697 |
| No log | 9.96 | 30 | 3.9383 | 0.185 | 0.8742 | 6.0724 | 0.185 | 0.0488 | 0.2474 | 0.7712 |
| No log | 10.96 | 33 | 3.9336 | 0.185 | 0.8746 | 6.7953 | 0.185 | 0.0488 | 0.2533 | 0.7685 |
| No log | 11.96 | 36 | 3.9298 | 0.185 | 0.8755 | 6.9469 | 0.185 | 0.0488 | 0.2679 | 0.7659 |
| No log | 12.96 | 39 | 3.9253 | 0.185 | 0.8756 | 6.9654 | 0.185 | 0.0488 | 0.2591 | 0.7640 |
| No log | 13.96 | 42 | 3.9194 | 0.185 | 0.8750 | 6.9522 | 0.185 | 0.0488 | 0.2681 | 0.7604 |
| No log | 14.96 | 45 | 3.9128 | 0.185 | 0.8744 | 6.9200 | 0.185 | 0.0488 | 0.2611 | 0.7617 |
| No log | 15.96 | 48 | 3.9074 | 0.185 | 0.8733 | 6.8369 | 0.185 | 0.0488 | 0.2611 | 0.7600 |
| No log | 16.96 | 51 | 3.9041 | 0.185 | 0.8726 | 6.8278 | 0.185 | 0.0488 | 0.2558 | 0.7566 |
| No log | 17.96 | 54 | 3.9025 | 0.185 | 0.8719 | 6.7039 | 0.185 | 0.0488 | 0.2588 | 0.7510 |
| No log | 18.96 | 57 | 3.9012 | 0.185 | 0.8717 | 6.6384 | 0.185 | 0.0488 | 0.2580 | 0.7484 |
| No log | 19.96 | 60 | 3.8987 | 0.185 | 0.8712 | 6.6323 | 0.185 | 0.0488 | 0.2612 | 0.7450 |
| No log | 20.96 | 63 | 3.8971 | 0.185 | 0.8712 | 6.6319 | 0.185 | 0.0488 | 0.2615 | 0.7443 |
| No log | 21.96 | 66 | 3.8956 | 0.185 | 0.8710 | 6.6323 | 0.185 | 0.0488 | 0.2659 | 0.7439 |
| No log | 22.96 | 69 | 3.8945 | 0.185 | 0.8708 | 6.6307 | 0.185 | 0.0488 | 0.2569 | 0.7436 |
| No log | 23.96 | 72 | 3.8940 | 0.185 | 0.8708 | 6.6295 | 0.185 | 0.0488 | 0.2526 | 0.7434 |
| No log | 24.96 | 75 | 3.8936 | 0.185 | 0.8707 | 6.6284 | 0.185 | 0.0488 | 0.2527 | 0.7434 |
### Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2
|
cgcgcgcgcg/111
|
cgcgcgcgcg
| 2023-07-09T09:32:21Z | 0 | 0 | null |
[
"arxiv:1910.09700",
"region:us"
] | null | 2023-07-09T09:31:54Z |
---
# 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
{}
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
This modelcard aims to be a base template for new models. It has been generated using [this raw template](https://github.com/huggingface/huggingface_hub/blob/main/src/huggingface_hub/templates/modelcard_template.md?plain=1).
## 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]
|
jvvelzen/dqn-SpaceInvadersNoFrameskip-v4
|
jvvelzen
| 2023-07-09T09:29:26Z | 7 | 0 |
stable-baselines3
|
[
"stable-baselines3",
"SpaceInvadersNoFrameskip-v4",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-07-09T09:28:53Z |
---
library_name: stable-baselines3
tags:
- SpaceInvadersNoFrameskip-v4
- deep-reinforcement-learning
- reinforcement-learning
- stable-baselines3
model-index:
- name: DQN
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: SpaceInvadersNoFrameskip-v4
type: SpaceInvadersNoFrameskip-v4
metrics:
- type: mean_reward
value: 476.00 +/- 136.38
name: mean_reward
verified: false
---
# **DQN** Agent playing **SpaceInvadersNoFrameskip-v4**
This is a trained model of a **DQN** agent playing **SpaceInvadersNoFrameskip-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 SpaceInvadersNoFrameskip-v4 -orga jvvelzen -f logs/
python -m rl_zoo3.enjoy --algo dqn --env SpaceInvadersNoFrameskip-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 SpaceInvadersNoFrameskip-v4 -orga jvvelzen -f logs/
python -m rl_zoo3.enjoy --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/
```
## Training (with the RL Zoo)
```
python -m rl_zoo3.train --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/
# Upload the model and generate video (when possible)
python -m rl_zoo3.push_to_hub --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/ -orga jvvelzen
```
## 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),
('optimize_memory_usage', False),
('policy', 'CnnPolicy'),
('target_update_interval', 1000),
('train_freq', 4),
('normalize', False)])
```
# Environment Arguments
```python
{'render_mode': 'rgb_array'}
```
|
jordyvl/dit-tiny_tobacco3482_kd_CEKD_t2.5_a0.5
|
jordyvl
| 2023-07-09T09:28:18Z | 163 | 0 |
transformers
|
[
"transformers",
"pytorch",
"beit",
"image-classification",
"generated_from_trainer",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
image-classification
| 2023-07-09T09:16:22Z |
---
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: dit-tiny_tobacco3482_kd_CEKD_t2.5_a0.5
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. -->
# dit-tiny_tobacco3482_kd_CEKD_t2.5_a0.5
This model is a fine-tuned version of [microsoft/dit-base](https://huggingface.co/microsoft/dit-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 3.9560
- Accuracy: 0.18
- Brier Loss: 0.8800
- Nll: 6.8606
- F1 Micro: 0.18
- F1 Macro: 0.0306
- Ece: 0.2612
- Aurc: 0.8512
## 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
- gradient_accumulation_steps: 16
- total_train_batch_size: 256
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 25
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Brier Loss | Nll | F1 Micro | F1 Macro | Ece | Aurc |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:----------:|:-------:|:--------:|:--------:|:------:|:------:|
| No log | 0.96 | 3 | 4.2281 | 0.145 | 0.8999 | 10.1620 | 0.145 | 0.0253 | 0.2222 | 0.8467 |
| No log | 1.96 | 6 | 4.1872 | 0.145 | 0.8946 | 10.5915 | 0.145 | 0.0253 | 0.2275 | 0.8468 |
| No log | 2.96 | 9 | 4.1248 | 0.155 | 0.8866 | 8.6280 | 0.155 | 0.0360 | 0.2179 | 0.8487 |
| No log | 3.96 | 12 | 4.0716 | 0.155 | 0.8806 | 6.5480 | 0.155 | 0.0272 | 0.2254 | 0.8851 |
| No log | 4.96 | 15 | 4.0359 | 0.155 | 0.8778 | 6.7781 | 0.155 | 0.0271 | 0.2310 | 0.8931 |
| No log | 5.96 | 18 | 4.0135 | 0.155 | 0.8774 | 7.8547 | 0.155 | 0.0271 | 0.2345 | 0.8965 |
| No log | 6.96 | 21 | 3.9978 | 0.185 | 0.8779 | 8.3528 | 0.185 | 0.0468 | 0.2615 | 0.8612 |
| No log | 7.96 | 24 | 3.9867 | 0.18 | 0.8789 | 7.6001 | 0.18 | 0.0308 | 0.2618 | 0.8546 |
| No log | 8.96 | 27 | 3.9782 | 0.18 | 0.8796 | 7.0871 | 0.18 | 0.0306 | 0.2613 | 0.8538 |
| No log | 9.96 | 30 | 3.9726 | 0.18 | 0.8800 | 7.0519 | 0.18 | 0.0306 | 0.2687 | 0.8545 |
| No log | 10.96 | 33 | 3.9684 | 0.18 | 0.8803 | 7.0277 | 0.18 | 0.0306 | 0.2656 | 0.8537 |
| No log | 11.96 | 36 | 3.9654 | 0.18 | 0.8805 | 7.0162 | 0.18 | 0.0306 | 0.2708 | 0.8536 |
| No log | 12.96 | 39 | 3.9633 | 0.18 | 0.8805 | 7.0056 | 0.18 | 0.0306 | 0.2619 | 0.8535 |
| No log | 13.96 | 42 | 3.9614 | 0.18 | 0.8804 | 6.9981 | 0.18 | 0.0306 | 0.2617 | 0.8532 |
| No log | 14.96 | 45 | 3.9598 | 0.18 | 0.8804 | 6.9923 | 0.18 | 0.0306 | 0.2669 | 0.8531 |
| No log | 15.96 | 48 | 3.9586 | 0.18 | 0.8803 | 6.9334 | 0.18 | 0.0306 | 0.2669 | 0.8529 |
| No log | 16.96 | 51 | 3.9578 | 0.18 | 0.8802 | 6.9237 | 0.18 | 0.0306 | 0.2716 | 0.8522 |
| No log | 17.96 | 54 | 3.9576 | 0.18 | 0.8802 | 6.8704 | 0.18 | 0.0306 | 0.2666 | 0.8521 |
| No log | 18.96 | 57 | 3.9574 | 0.18 | 0.8802 | 6.8662 | 0.18 | 0.0306 | 0.2664 | 0.8523 |
| No log | 19.96 | 60 | 3.9568 | 0.18 | 0.8801 | 6.8641 | 0.18 | 0.0306 | 0.2614 | 0.8518 |
| No log | 20.96 | 63 | 3.9566 | 0.18 | 0.8801 | 6.8634 | 0.18 | 0.0306 | 0.2659 | 0.8516 |
| No log | 21.96 | 66 | 3.9563 | 0.18 | 0.8800 | 6.8632 | 0.18 | 0.0306 | 0.2612 | 0.8516 |
| No log | 22.96 | 69 | 3.9561 | 0.18 | 0.8800 | 6.8620 | 0.18 | 0.0306 | 0.2612 | 0.8513 |
| No log | 23.96 | 72 | 3.9561 | 0.18 | 0.8800 | 6.8611 | 0.18 | 0.0306 | 0.2612 | 0.8513 |
| No log | 24.96 | 75 | 3.9560 | 0.18 | 0.8800 | 6.8606 | 0.18 | 0.0306 | 0.2612 | 0.8512 |
### Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2
|
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.