modelId
stringlengths 5
139
| author
stringlengths 2
42
| last_modified
timestamp[us, tz=UTC]date 2020-02-15 11:33:14
2025-08-30 00:39:23
| downloads
int64 0
223M
| likes
int64 0
11.7k
| library_name
stringclasses 526
values | tags
listlengths 1
4.05k
| pipeline_tag
stringclasses 55
values | createdAt
timestamp[us, tz=UTC]date 2022-03-02 23:29:04
2025-08-30 00:39:08
| card
stringlengths 11
1.01M
|
|---|---|---|---|---|---|---|---|---|---|
Team-Atom/smvlb_record_blue_pick_64_40000
|
Team-Atom
| 2025-08-29T03:08:26Z | 0 | 0 |
lerobot
|
[
"lerobot",
"safetensors",
"smolvla",
"robotics",
"dataset:Team-Atom/blue_pick",
"arxiv:2506.01844",
"base_model:lerobot/smolvla_base",
"base_model:finetune:lerobot/smolvla_base",
"license:apache-2.0",
"region:us"
] |
robotics
| 2025-08-29T03:08:10Z |
---
base_model: lerobot/smolvla_base
datasets: Team-Atom/blue_pick
library_name: lerobot
license: apache-2.0
model_name: smolvla
pipeline_tag: robotics
tags:
- lerobot
- smolvla
- robotics
---
# Model Card for smolvla
<!-- Provide a quick summary of what the model is/does. -->
[SmolVLA](https://huggingface.co/papers/2506.01844) is a compact, efficient vision-language-action model that achieves competitive performance at reduced computational costs and can be deployed on consumer-grade hardware.
This policy has been trained and pushed to the Hub using [LeRobot](https://github.com/huggingface/lerobot).
See the full documentation at [LeRobot Docs](https://huggingface.co/docs/lerobot/index).
---
## How to Get Started with the Model
For a complete walkthrough, see the [training guide](https://huggingface.co/docs/lerobot/il_robots#train-a-policy).
Below is the short version on how to train and run inference/eval:
### Train from scratch
```bash
python -m lerobot.scripts.train \
--dataset.repo_id=${HF_USER}/<dataset> \
--policy.type=act \
--output_dir=outputs/train/<desired_policy_repo_id> \
--job_name=lerobot_training \
--policy.device=cuda \
--policy.repo_id=${HF_USER}/<desired_policy_repo_id>
--wandb.enable=true
```
_Writes checkpoints to `outputs/train/<desired_policy_repo_id>/checkpoints/`._
### Evaluate the policy/run inference
```bash
python -m lerobot.record \
--robot.type=so100_follower \
--dataset.repo_id=<hf_user>/eval_<dataset> \
--policy.path=<hf_user>/<desired_policy_repo_id> \
--episodes=10
```
Prefix the dataset repo with **eval\_** and supply `--policy.path` pointing to a local or hub checkpoint.
---
## Model Details
- **License:** apache-2.0
|
Sonia280825/SoniaSkewes-Replicate
|
Sonia280825
| 2025-08-29T03:07:52Z | 0 | 0 |
diffusers
|
[
"diffusers",
"flux",
"lora",
"replicate",
"text-to-image",
"en",
"base_model:black-forest-labs/FLUX.1-dev",
"base_model:adapter:black-forest-labs/FLUX.1-dev",
"license:other",
"region:us"
] |
text-to-image
| 2025-08-28T06:52:03Z |
---
license: other
license_name: flux-1-dev-non-commercial-license
license_link: https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/LICENSE.md
language:
- en
tags:
- flux
- diffusers
- lora
- replicate
base_model: "black-forest-labs/FLUX.1-dev"
pipeline_tag: text-to-image
# widget:
# - text: >-
# prompt
# output:
# url: https://...
instance_prompt: SONIA
---
# Soniaskewes Replicate
<Gallery />
## About this LoRA
This is a [LoRA](https://replicate.com/docs/guides/working-with-loras) for the FLUX.1-dev text-to-image model. It can be used with diffusers or ComfyUI.
It was trained on [Replicate](https://replicate.com/) using AI toolkit: https://replicate.com/ostris/flux-dev-lora-trainer/train
## Trigger words
You should use `SONIA` to trigger the image generation.
## Run this LoRA with an API using Replicate
```py
import replicate
input = {
"prompt": "SONIA",
"lora_weights": "https://huggingface.co/Sonia280825/SoniaSkewes-Replicate/resolve/main/lora.safetensors"
}
output = replicate.run(
"black-forest-labs/flux-dev-lora",
input=input
)
for index, item in enumerate(output):
with open(f"output_{index}.webp", "wb") as file:
file.write(item.read())
```
## Use it with the [🧨 diffusers library](https://github.com/huggingface/diffusers)
```py
from diffusers import AutoPipelineForText2Image
import torch
pipeline = AutoPipelineForText2Image.from_pretrained('black-forest-labs/FLUX.1-dev', torch_dtype=torch.float16).to('cuda')
pipeline.load_lora_weights('Sonia280825/SoniaSkewes-Replicate', weight_name='lora.safetensors')
image = pipeline('SONIA').images[0]
```
For more details, including weighting, merging and fusing LoRAs, check the [documentation on loading LoRAs in diffusers](https://huggingface.co/docs/diffusers/main/en/using-diffusers/loading_adapters)
## Training details
- Steps: 2024
- Learning rate: 0.0004
- LoRA rank: 16
## Contribute your own examples
You can use the [community tab](https://huggingface.co/Sonia280825/SoniaSkewes-Replicate/discussions) to add images that show off what you’ve made with this LoRA.
|
ambrosfitz/roberta-base-19th-century-sentiment-multi-label
|
ambrosfitz
| 2025-08-29T03:05:18Z | 0 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"roberta",
"text-classification",
"generated_from_trainer",
"base_model:ambrosfitz/roberta-base-19th-century-v2",
"base_model:finetune:ambrosfitz/roberta-base-19th-century-v2",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2025-08-29T03:04:00Z |
---
library_name: transformers
license: mit
base_model: ambrosfitz/roberta-base-19th-century-v2
tags:
- generated_from_trainer
metrics:
- f1
- accuracy
model-index:
- name: roberta-base-19th-century-sentiment-multi-label
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-19th-century-sentiment-multi-label
This model is a fine-tuned version of [ambrosfitz/roberta-base-19th-century-v2](https://huggingface.co/ambrosfitz/roberta-base-19th-century-v2) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5874
- F1: 0.2593
- Roc Auc: 0.5468
- Accuracy: 0.0
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- num_epochs: 5
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 | Roc Auc | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:------:|:-------:|:--------:|
| No log | 1.0 | 7 | 0.6639 | 0.3761 | 0.5939 | 0.0 |
| No log | 2.0 | 14 | 0.6161 | 0.2593 | 0.5468 | 0.0 |
| No log | 3.0 | 21 | 0.5961 | 0.2593 | 0.5468 | 0.0 |
| No log | 4.0 | 28 | 0.5896 | 0.2593 | 0.5468 | 0.0 |
| No log | 5.0 | 35 | 0.5874 | 0.2593 | 0.5468 | 0.0 |
### Framework versions
- Transformers 4.55.4
- Pytorch 2.8.0+cu126
- Datasets 4.0.0
- Tokenizers 0.21.4
|
losaferto22/blockassist-bc-mute_shy_clam_1756436233
|
losaferto22
| 2025-08-29T02:57:57Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"mute shy clam",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T02:57:39Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- mute shy clam
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
internlm/Intern-S1
|
internlm
| 2025-08-29T02:56:15Z | 61,482 | 234 |
transformers
|
[
"transformers",
"safetensors",
"interns1",
"text-generation",
"image-text-to-text",
"conversational",
"custom_code",
"arxiv:2508.15763",
"license:apache-2.0",
"autotrain_compatible",
"region:us"
] |
image-text-to-text
| 2025-07-24T06:05:13Z |
---
license: apache-2.0
pipeline_tag: image-text-to-text
library_name: transformers
---
## Intern-S1
<div align="center">
<img src="https://cdn-uploads.huggingface.co/production/uploads/642695e5274e7ad464c8a5ba/E43cgEXBRWjVJlU_-hdh6.png" />
<div> </div>
[💻Github Repo](https://github.com/InternLM/Intern-S1) • [🤗Model Collections](https://huggingface.co/collections/internlm/intern-s1-6882e325e8ac1c58ba108aa5) • [📜Technical Report](https://arxiv.org/abs/2508.15763) • [💬Online Chat](https://chat.intern-ai.org.cn/)
</div>
<p align="center">
👋 join us on <a href="https://discord.gg/xa29JuW87d" target="_blank">Discord</a> and <a href="https://cdn.vansin.top/intern-s1.jpg" target="_blank">WeChat</a>
</p>
## Introduction
We introduce **Intern-S1**, our **most advanced open-source multimodal reasoning model** to date. Intern-S1 combines **strong general-task capabilities with state-of-the-art performance on a wide range of scientific tasks**, rivaling leading closed-source commercial models.
Built upon a 235B MoE language model (Qwen3) and a 6B Vision encoder (InternViT), Intern-S1 has been further pretrained on **5 trillion tokens** of multimodal data, including over **2.5 trillion scientific-domain tokens**. This enables the model to retain strong general capabilities while excelling in specialized scientific domains such as **interpreting chemical structures, understanding protein sequences, and planning compound synthesis routes**, making Intern-S1 to be a capable research assistant for real-world scientific applications.
Features
- Strong performance across language and vision reasoning benchmarks, especially scientific tasks.
- Continuously pretrained on a massive 5T token dataset, with over 50% specialized scientific data, embedding deep domain expertise.
- Dynamic tokenizer enables native understanding of molecular formulas, protein sequences, and seismic signals.
## Performance
We evaluate the Intern-S1 on various benchmarks including general datasets and scientifc datasets. We report the performance comparsion with the recent VLMs and LLMs below.
<table>
<thead>
<tr>
<th rowspan="2">Benchmarks</th>
<th colspan="2">Intern-S1</th>
<th>InternVL3-78B</th>
<th>Qwen2.5-VL-72B</th>
<th>DS-R1-0528</th>
<th>Qwen3-235B-A22B</th>
<th>Kimi-K2-Instruct</th>
<th>Gemini-2.5 Pro</th>
<th>o3</th>
<th>Grok-4</th>
</tr>
</thead>
<tbody>
<tr><td>MMLU-Pro</td><td colspan="2">83.5 ✅</td><td>73.0</td><td>72.1</td><td>83.4</td><td>82.2</td><td>82.7</td><td>86.0</td><td>85.0</td><td>85.9</td></tr>
<tr><td>MMMU</td><td colspan="2">77.7 ✅</td><td>72.2</td><td>70.2</td><td>-</td><td>-</td><td>-</td><td>81.9</td><td>80.8</td><td>77.9</td></tr>
<tr><td>GPQA</td><td colspan="2">77.3</td><td>49.9</td><td>49.0</td><td>80.6</td><td>71.1</td><td>77.8</td><td>83.8</td><td>83.3</td><td>87.5</td></tr>
<tr><td>MMStar</td><td colspan="2">74.9 ✅</td><td>72.5</td><td>70.8</td><td>-</td><td>-</td><td>-</td><td>79.3</td><td>75.1</td><td>69.6</td></tr>
<tr><td>MathVista</td><td colspan="2">81.5 👑</td><td>79.0</td><td>74.8</td><td>-</td><td>-</td><td>-</td><td>80.3</td><td>77.5</td><td>72.5</td></tr>
<tr><td>AIME2025</td><td colspan="2">86.0</td><td>10.7</td><td>10.9</td><td>87.5</td><td>81.5</td><td>51.4</td><td>83.0</td><td>88.9</td><td>91.7</td></tr>
<tr><td>MathVision</td><td colspan="2">62.5 ✅</td><td>43.1</td><td>38.1</td><td>-</td><td>-</td><td>-</td><td>73.0</td><td>67.7</td><td>67.3</td></tr>
<tr><td>IFEval</td><td colspan="2">86.7</td><td>75.6</td><td>83.9</td><td>79.7</td><td>85.0</td><td>90.2</td><td>91.5</td><td>92.2</td><td>92.8</td></tr>
<tr><td>SFE</td><td colspan="2">44.3 👑</td><td>36.2</td><td>30.5</td><td>-</td><td>-</td><td>-</td><td>43.0</td><td>37.7</td><td>31.2</td></tr>
<tr><td>Physics</td><td colspan="2">44.0 ✅</td><td>23.1</td><td>15.7</td><td>-</td><td>-</td><td>-</td><td>40.0</td><td>47.9</td><td>42.8</td></tr>
<tr><td>SmolInstruct</td><td colspan="2">51.0 👑</td><td>19.4</td><td>21.0</td><td>30.7</td><td>28.7</td><td>48.1</td><td>40.4</td><td>43.9</td><td>47.3</td></tr>
<tr><td>ChemBench</td><td colspan="2">83.4 👑</td><td>61.3</td><td>61.6</td><td>75.6</td><td>75.8</td><td>75.3</td><td>82.8</td><td>81.6</td><td>83.3</td></tr>
<tr><td>MatBench</td><td colspan="2">75.0 👑</td><td>49.3</td><td>51.5</td><td>57.7</td><td>52.1</td><td>61.7</td><td>61.7</td><td>61.6</td><td>67.9</td></tr>
<tr><td>MicroVQA</td><td colspan="2">63.9 👑</td><td>59.1</td><td>53.0</td><td>-</td><td>-</td><td>-</td><td>63.1</td><td>58.3</td><td>59.5</td></tr>
<tr><td>ProteinLMBench</td><td colspan="2">63.1</td><td>61.6</td><td>61.0</td><td>61.4</td><td>59.8</td><td>66.7</td><td>62.9</td><td>67.7</td><td>66.2</td></tr>
<tr><td>MSEarthMCQ</td><td colspan="2">65.7 👑</td><td>57.2</td><td>37.6</td><td>-</td><td>-</td><td>-</td><td>59.9</td><td>61.0</td><td>58.0</td></tr>
<tr><td>XLRS-Bench</td><td colspan="2">55.0 👑</td><td>49.3</td><td>50.9</td><td>-</td><td>-</td><td>-</td><td>45.2</td><td>43.6</td><td>45.4</td></tr>
</tbody>
</table>
> **Note**: ✅ means the best performance among open-sourced models, 👑 indicates the best performance among all models.
We use the [OpenCompass](https://github.com/open-compass/OpenCompass/) and [VLMEvalkit](https://github.com/open-compass/vlmevalkit) to evaluate all models.
Please refer to [this page](https://opencompass.readthedocs.io/en/latest/user_guides/interns1.html) to quickly start the text-only evaluation task.
## Quick Start
### Sampling Parameters
We recommend using the following hyperparameters to ensure better results
```python
top_p = 1.0
top_k = 50
min_p = 0.0
temperature = 0.7
```
### Transformers
The following provides demo code illustrating how to generate based on text and multimodal inputs.
> **Please use transformers>=4.53.0 to ensure the model works normally.**
#### Text input
```python
from transformers import AutoProcessor, AutoModelForCausalLM
import torch
model_name = "internlm/Intern-S1"
processor = AutoProcessor.from_pretrained(model_name, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(model_name, device_map="auto", torch_dtype="auto", trust_remote_code=True)
messages = [
{
"role": "user",
"content": [
{"type": "text", "text": "tell me about an interesting physical phenomenon."},
],
}
]
inputs = processor.apply_chat_template(messages, add_generation_prompt=True, tokenize=True, return_dict=True, return_tensors="pt").to(model.device, dtype=torch.bfloat16)
generate_ids = model.generate(**inputs, max_new_tokens=32768)
decoded_output = processor.decode(generate_ids[0, inputs["input_ids"].shape[1] :], skip_special_tokens=True)
print(decoded_output)
```
#### Image input
```python
from transformers import AutoProcessor, AutoModelForCausalLM
import torch
model_name = "internlm/Intern-S1"
processor = AutoProcessor.from_pretrained(model_name, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(model_name, device_map="auto", torch_dtype="auto", trust_remote_code=True)
messages = [
{
"role": "user",
"content": [
{"type": "image", "url": "http://images.cocodataset.org/val2017/000000039769.jpg"},
{"type": "text", "text": "Please describe the image explicitly."},
],
}
]
inputs = processor.apply_chat_template(messages, add_generation_prompt=True, tokenize=True, return_dict=True, return_tensors="pt").to(model.device, dtype=torch.bfloat16)
generate_ids = model.generate(**inputs, max_new_tokens=32768)
decoded_output = processor.decode(generate_ids[0, inputs["input_ids"].shape[1] :], skip_special_tokens=True)
print(decoded_output)
```
#### Video input
Please ensure that the decord video decoding library is installed via `pip install decord`.
```python
from transformers import AutoProcessor, AutoModelForCausalLM
import torch
model_name = "internlm/Intern-S1"
processor = AutoProcessor.from_pretrained(model_name, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(model_name, device_map="auto", torch_dtype="auto", trust_remote_code=True)
messages = [
{
"role": "user",
"content": [
{
"type": "video",
"url": "https://huggingface.co/datasets/hf-internal-testing/fixtures_videos/resolve/main/tennis.mp4",
},
{"type": "text", "text": "What type of shot is the man performing?"},
],
}
]
inputs = processor.apply_chat_template(
messages,
return_tensors="pt",
add_generation_prompt=True,
video_load_backend="decord",
tokenize=True,
return_dict=True,
).to(model.device, dtype=torch.float16)
generate_ids = model.generate(**inputs, max_new_tokens=32768)
decoded_output = processor.decode(generate_ids[0, inputs["input_ids"].shape[1] :], skip_special_tokens=True)
print(decoded_output)
```
### Serving
The minimum hardware requirements for deploying Intern-S1 series models are:
| Model | A100(GPUs) | H800(GPUs) | H100(GPUs) | H200(GPUs) |
| :---------------------------------------------------------------------: | :--------: | :--------: | :--------: | :--------: |
| [internlm/Intern-S1](https://huggingface.co/internlm/Intern-S1) | 8 | 8 | 8 | 4 |
| [internlm/Intern-S1-FP8](https://huggingface.co/internlm/Intern-S1-FP8) | - | 4 | 4 | 2 |
You can utilize one of the following LLM inference frameworks to create an OpenAI compatible server:
#### [lmdeploy (>=0.9.2)](https://github.com/InternLM/lmdeploy)
```bash
lmdeploy serve api_server internlm/Intern-S1 --reasoning-parser intern-s1 --tool-call-parser intern-s1 --tp 8
```
#### [vllm (>=0.10.1)](https://github.com/vllm-project/vllm)
```bash
vllm serve internlm/Intern-S1 --tensor-parallel-size 8 --trust-remote-code
```
#### [sglang](https://github.com/sgl-project/sglang)
```bash
python3 -m sglang.launch_server \
--model-path internlm/Intern-S1 \
--trust-remote-code \
--tp 8 \
--grammar-backend none
```
#### ollama for local deployment:
```bash
# install ollama
curl -fsSL https://ollama.com/install.sh | sh
# fetch model
ollama pull internlm/interns1
# run model
ollama run internlm/interns1
# then use openai client to call on http://localhost:11434/v1
```
## Advanced Usage
### Tool Calling
Many Large Language Models (LLMs) now feature **Tool Calling**, a powerful capability that allows them to extend their functionality by interacting with external tools and APIs. This enables models to perform tasks like fetching up-to-the-minute information, running code, or calling functions within other applications.
A key advantage for developers is that a growing number of open-source LLMs are designed to be compatible with the OpenAI API. This means you can leverage the same familiar syntax and structure from the OpenAI library to implement tool calling with these open-source models. As a result, the code demonstrated in this tutorial is versatile—it works not just with OpenAI models, but with any model that follows the same interface standard.
To illustrate how this works, let's dive into a practical code example that uses tool calling to get the latest weather forecast (based on lmdeploy api server).
```python
from openai import OpenAI
import json
def get_current_temperature(location: str, unit: str = "celsius"):
"""Get current temperature at a location.
Args:
location: The location to get the temperature for, in the format "City, State, Country".
unit: The unit to return the temperature in. Defaults to "celsius". (choices: ["celsius", "fahrenheit"])
Returns:
the temperature, the location, and the unit in a dict
"""
return {
"temperature": 26.1,
"location": location,
"unit": unit,
}
def get_temperature_date(location: str, date: str, unit: str = "celsius"):
"""Get temperature at a location and date.
Args:
location: The location to get the temperature for, in the format "City, State, Country".
date: The date to get the temperature for, in the format "Year-Month-Day".
unit: The unit to return the temperature in. Defaults to "celsius". (choices: ["celsius", "fahrenheit"])
Returns:
the temperature, the location, the date and the unit in a dict
"""
return {
"temperature": 25.9,
"location": location,
"date": date,
"unit": unit,
}
def get_function_by_name(name):
if name == "get_current_temperature":
return get_current_temperature
if name == "get_temperature_date":
return get_temperature_date
tools = [{
'type': 'function',
'function': {
'name': 'get_current_temperature',
'description': 'Get current temperature at a location.',
'parameters': {
'type': 'object',
'properties': {
'location': {
'type': 'string',
'description': 'The location to get the temperature for, in the format \'City, State, Country\'.'
},
'unit': {
'type': 'string',
'enum': [
'celsius',
'fahrenheit'
],
'description': 'The unit to return the temperature in. Defaults to \'celsius\'.'
}
},
'required': [
'location'
]
}
}
}, {
'type': 'function',
'function': {
'name': 'get_temperature_date',
'description': 'Get temperature at a location and date.',
'parameters': {
'type': 'object',
'properties': {
'location': {
'type': 'string',
'description': 'The location to get the temperature for, in the format \'City, State, Country\'.'
},
'date': {
'type': 'string',
'description': 'The date to get the temperature for, in the format \'Year-Month-Day\'.'
},
'unit': {
'type': 'string',
'enum': [
'celsius',
'fahrenheit'
],
'description': 'The unit to return the temperature in. Defaults to \'celsius\'.'
}
},
'required': [
'location',
'date'
]
}
}
}]
messages = [
{'role': 'user', 'content': 'Today is 2024-11-14, What\'s the temperature in San Francisco now? How about tomorrow?'}
]
openai_api_key = "EMPTY"
openai_api_base = "http://0.0.0.0:23333/v1"
client = OpenAI(
api_key=openai_api_key,
base_url=openai_api_base,
)
model_name = client.models.list().data[0].id
response = client.chat.completions.create(
model=model_name,
messages=messages,
max_tokens=32768,
temperature=0.8,
top_p=0.8,
stream=False,
extra_body=dict(spaces_between_special_tokens=False, enable_thinking=False),
tools=tools)
print(response.choices[0].message)
messages.append(response.choices[0].message)
for tool_call in response.choices[0].message.tool_calls:
tool_call_args = json.loads(tool_call.function.arguments)
tool_call_result = get_function_by_name(tool_call.function.name)(**tool_call_args)
tool_call_result = json.dumps(tool_call_result, ensure_ascii=False)
messages.append({
'role': 'tool',
'name': tool_call.function.name,
'content': tool_call_result,
'tool_call_id': tool_call.id
})
response = client.chat.completions.create(
model=model_name,
messages=messages,
temperature=0.8,
top_p=0.8,
stream=False,
extra_body=dict(spaces_between_special_tokens=False, enable_thinking=False),
tools=tools)
print(response.choices[0].message.content)
```
### Switching Between Thinking and Non-Thinking Modes
Intern-S1 enables thinking mode by default, enhancing the model's reasoning capabilities to generate higher-quality responses. This feature can be disabled by setting `enable_thinking=False` in `tokenizer.apply_chat_template`
```python
text = tokenizer.apply_chat_template(
messages,
tokenize=False,
add_generation_prompt=True,
enable_thinking=False # think mode indicator
)
```
With LMDeploy serving Intern-S1 models, you can dynamically control the thinking mode by adjusting the `enable_thinking` parameter in your requests.
```python
from openai import OpenAI
import json
messages = [
{
'role': 'user',
'content': 'who are you'
}, {
'role': 'assistant',
'content': 'I am an AI'
}, {
'role': 'user',
'content': 'AGI is?'
}]
openai_api_key = "EMPTY"
openai_api_base = "http://0.0.0.0:23333/v1"
client = OpenAI(
api_key=openai_api_key,
base_url=openai_api_base,
)
model_name = client.models.list().data[0].id
response = client.chat.completions.create(
model=model_name,
messages=messages,
temperature=0.7,
top_p=0.8,
max_tokens=2048,
extra_body={
"enable_thinking": False,
}
)
print(json.dumps(response.model_dump(), indent=2, ensure_ascii=False))
```
For vllm and sglang users, configure this through,
```python
extra_body={
"chat_template_kwargs": {"enable_thinking": False}
}
```
## Citation
If you find this work useful, feel free to give us a cite.
```
@misc{bai2025interns1scientificmultimodalfoundation,
title={Intern-S1: A Scientific Multimodal Foundation Model},
author={Lei Bai and Zhongrui Cai and Maosong Cao and Weihan Cao and Chiyu Chen and Haojiong Chen and Kai Chen and Pengcheng Chen and Ying Chen and Yongkang Chen and Yu Cheng and Yu Cheng and Pei Chu and Tao Chu and Erfei Cui and Ganqu Cui and Long Cui and Ziyun Cui and Nianchen Deng and Ning Ding and Nanqin Dong and Peijie Dong and Shihan Dou and Sinan Du and Haodong Duan and Caihua Fan and Ben Gao and Changjiang Gao and Jianfei Gao and Songyang Gao and Yang Gao and Zhangwei Gao and Jiaye Ge and Qiming Ge and Lixin Gu and Yuzhe Gu and Aijia Guo and Qipeng Guo and Xu Guo and Conghui He and Junjun He and Yili Hong and Siyuan Hou and Caiyu Hu and Hanglei Hu and Jucheng Hu and Ming Hu and Zhouqi Hua and Haian Huang and Junhao Huang and Xu Huang and Zixian Huang and Zhe Jiang and Lingkai Kong and Linyang Li and Peiji Li and Pengze Li and Shuaibin Li and Tianbin Li and Wei Li and Yuqiang Li and Dahua Lin and Junyao Lin and Tianyi Lin and Zhishan Lin and Hongwei Liu and Jiangning Liu and Jiyao Liu and Junnan Liu and Kai Liu and Kaiwen Liu and Kuikun Liu and Shichun Liu and Shudong Liu and Wei Liu and Xinyao Liu and Yuhong Liu and Zhan Liu and Yinquan Lu and Haijun Lv and Hongxia Lv and Huijie Lv and Qidang Lv and Ying Lv and Chengqi Lyu and Chenglong Ma and Jianpeng Ma and Ren Ma and Runmin Ma and Runyuan Ma and Xinzhu Ma and Yichuan Ma and Zihan Ma and Sixuan Mi and Junzhi Ning and Wenchang Ning and Xinle Pang and Jiahui Peng and Runyu Peng and Yu Qiao and Jiantao Qiu and Xiaoye Qu and Yuan Qu and Yuchen Ren and Fukai Shang and Wenqi Shao and Junhao Shen and Shuaike Shen and Chunfeng Song and Demin Song and Diping Song and Chenlin Su and Weijie Su and Weigao Sun and Yu Sun and Qian Tan and Cheng Tang and Huanze Tang and Kexian Tang and Shixiang Tang and Jian Tong and Aoran Wang and Bin Wang and Dong Wang and Lintao Wang and Rui Wang and Weiyun Wang and Wenhai Wang and Yi Wang and Ziyi Wang and Ling-I Wu and Wen Wu and Yue Wu and Zijian Wu and Linchen Xiao and Shuhao Xing and Chao Xu and Huihui Xu and Jun Xu and Ruiliang Xu and Wanghan Xu and GanLin Yang and Yuming Yang and Haochen Ye and Jin Ye and Shenglong Ye and Jia Yu and Jiashuo Yu and Jing Yu and Fei Yuan and Bo Zhang and Chao Zhang and Chen Zhang and Hongjie Zhang and Jin Zhang and Qiaosheng Zhang and Qiuyinzhe Zhang and Songyang Zhang and Taolin Zhang and Wenlong Zhang and Wenwei Zhang and Yechen Zhang and Ziyang Zhang and Haiteng Zhao and Qian Zhao and Xiangyu Zhao and Xiangyu Zhao and Bowen Zhou and Dongzhan Zhou and Peiheng Zhou and Yuhao Zhou and Yunhua Zhou and Dongsheng Zhu and Lin Zhu and Yicheng Zou},
year={2025},
eprint={2508.15763},
archivePrefix={arXiv},
primaryClass={cs.LG},
url={https://arxiv.org/abs/2508.15763},
}
```
|
John6666/mol-keun-mix-deepnavy-v2-eps-sdxl
|
John6666
| 2025-08-29T02:55:51Z | 0 | 0 |
diffusers
|
[
"diffusers",
"safetensors",
"text-to-image",
"stable-diffusion",
"stable-diffusion-xl",
"anime",
"furry",
"anthro",
"kemono",
"noobai",
"illustrious",
"en",
"base_model:Laxhar/noobai-XL-1.0",
"base_model:finetune:Laxhar/noobai-XL-1.0",
"license:other",
"autotrain_compatible",
"endpoints_compatible",
"diffusers:StableDiffusionXLPipeline",
"region:us"
] |
text-to-image
| 2025-08-29T02:48:55Z |
---
license: other
license_name: faipl-1.0-sd
license_link: https://freedevproject.org/faipl-1.0-sd/
language:
- en
library_name: diffusers
pipeline_tag: text-to-image
tags:
- text-to-image
- stable-diffusion
- stable-diffusion-xl
- anime
- furry
- anthro
- kemono
- noobai
- illustrious
base_model: Laxhar/noobai-XL-1.0
---
Original model is [here](https://civitai.com/models/135477?modelVersionId=2157057).
This model created by [Mol_Keun](https://civitai.com/user/Mol_Keun).
|
huggingtoots/TheDrummer-Behemoth-R1-123B-v2-MLX-4Bit
|
huggingtoots
| 2025-08-29T02:52:32Z | 0 | 0 |
mlx
|
[
"mlx",
"safetensors",
"mistral",
"base_model:TheDrummer/Behemoth-R1-123B-v2",
"base_model:quantized:TheDrummer/Behemoth-R1-123B-v2",
"4-bit",
"region:us"
] | null | 2025-08-29T02:40:45Z |
---
base_model: TheDrummer/Behemoth-R1-123B-v2
tags:
- mlx
---
# huggingtoots/TheDrummer-Behemoth-R1-123B-v2-MLX-4Bit
The Model [huggingtoots/TheDrummer-Behemoth-R1-123B-v2-MLX-4Bit](https://huggingface.co/huggingtoots/TheDrummer-Behemoth-R1-123B-v2-MLX-4Bit) was converted to MLX format from [TheDrummer/Behemoth-R1-123B-v2](https://huggingface.co/TheDrummer/Behemoth-R1-123B-v2) using mlx-lm version **0.26.4**.
➡️ <span style="color:#800080">If you want a free consulting session, </span>[fill out this form](https://forms.gle/xM9gw1urhypC4bWS6) <span style="color:#800080">to get in touch!</span> 🤗
## Use with mlx
```bash
pip install mlx-lm
```
```python
from mlx_lm import load, generate
model, tokenizer = load("huggingtoots/Behemoth-R1-123B-v2-mlx-4Bit")
prompt="hello"
if hasattr(tokenizer, "apply_chat_template") and tokenizer.chat_template is not None:
messages = [{"role": "user", "content": prompt}]
prompt = tokenizer.apply_chat_template(
messages, tokenize=False, add_generation_prompt=True
)
response = generate(model, tokenizer, prompt=prompt, verbose=True)
```
|
losaferto22/blockassist-bc-mute_shy_clam_1756435842
|
losaferto22
| 2025-08-29T02:51:37Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"mute shy clam",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T02:51:07Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- mute shy clam
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
hobson123/blockassist-bc-mammalian_dense_gibbon_1756435712
|
hobson123
| 2025-08-29T02:49:44Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"mammalian dense gibbon",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T02:49:29Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- mammalian dense gibbon
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
marvelskater/blockassist-bc-mighty_extinct_stork_1756435622
|
marvelskater
| 2025-08-29T02:47:53Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"mighty extinct stork",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T02:47:49Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- mighty extinct stork
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
John6666/illustrious-from-hades-v3-series-v3530-sdxl
|
John6666
| 2025-08-29T02:44:06Z | 0 | 0 |
diffusers
|
[
"diffusers",
"safetensors",
"text-to-image",
"stable-diffusion",
"stable-diffusion-xl",
"anime",
"realistic",
"semi-realistic",
"photorealism",
"2.5D",
"facial anatomy",
"emotional subtlety",
"skin detail",
"clothing",
"hair",
"prompt adherence",
"shadow rendering",
"lighting gradients",
"depth and surface control",
"anatomy",
"micro detail",
"illustrious",
"en",
"base_model:OnomaAIResearch/Illustrious-xl-early-release-v0",
"base_model:finetune:OnomaAIResearch/Illustrious-xl-early-release-v0",
"license:other",
"autotrain_compatible",
"endpoints_compatible",
"diffusers:StableDiffusionXLPipeline",
"region:us"
] |
text-to-image
| 2025-08-29T02:39:20Z |
---
license: other
license_name: faipl-1.0-sd
license_link: https://freedevproject.org/faipl-1.0-sd/
language:
- en
library_name: diffusers
pipeline_tag: text-to-image
tags:
- text-to-image
- stable-diffusion
- stable-diffusion-xl
- anime
- realistic
- semi-realistic
- photorealism
- 2.5D
- facial anatomy
- emotional subtlety
- skin detail
- clothing
- hair
- prompt adherence
- shadow rendering
- lighting gradients
- depth and surface control
- anatomy
- micro detail
- illustrious
base_model: OnomaAIResearch/Illustrious-xl-early-release-v0
---
Original model is [here](https://civitai.com/models/1653796/illustrious-from-hades?modelVersionId=2157620).
This model created by [DeViLDoNia](https://civitai.com/user/DeViLDoNia).
|
PaddlePaddle/PP-LCNet_x0_25_textline_ori
|
PaddlePaddle
| 2025-08-29T02:43:45Z | 1,185 | 0 |
PaddleOCR
|
[
"PaddleOCR",
"OCR",
"PaddlePaddle",
"textline_orientation_classification",
"image-to-text",
"en",
"zh",
"license:apache-2.0",
"region:us"
] |
image-to-text
| 2025-06-12T07:35:10Z |
---
license: apache-2.0
library_name: PaddleOCR
language:
- en
- zh
pipeline_tag: image-to-text
tags:
- OCR
- PaddlePaddle
- PaddleOCR
- textline_orientation_classification
---
# PP-LCNet_x0_25_textline_ori
## Introduction
The text line orientation classification module primarily distinguishes the orientation of text lines and corrects them using post-processing. In processes such as document scanning and license/certificate photography, to capture clearer images, the capture device may be rotated, resulting in text lines in various orientations. Standard OCR pipelines cannot handle such data well. By utilizing image classification technology, the orientation of text lines can be predetermined and adjusted, thereby enhancing the accuracy of OCR processing. The key accuracy metrics are as follow:
<table>
<tr>
<th>Model</th>
<th>Recognition Avg Accuracy(%)</th>
<th>Model Storage Size (M)</th>
<th>Introduction</th>
</tr>
<tr>
<td>PP-LCNet_x0_25_textline_ori</td>
<td>98.85</td>
<td>0.96</td>
<td>Text line classification model based on PP-LCNet_x0_25, with two classes: 0 degrees and 180 degrees</td>
</tr>
</table>
## Quick Start
### Installation
1. PaddlePaddle
Please refer to the following commands to install PaddlePaddle using pip:
```bash
# for CUDA11.8
python -m pip install paddlepaddle-gpu==3.0.0 -i https://www.paddlepaddle.org.cn/packages/stable/cu118/
# for CUDA12.6
python -m pip install paddlepaddle-gpu==3.0.0 -i https://www.paddlepaddle.org.cn/packages/stable/cu126/
# for CPU
python -m pip install paddlepaddle==3.0.0 -i https://www.paddlepaddle.org.cn/packages/stable/cpu/
```
For details about PaddlePaddle installation, please refer to the [PaddlePaddle official website](https://www.paddlepaddle.org.cn/en/install/quick).
2. PaddleOCR
Install the latest version of the PaddleOCR inference package from PyPI:
```bash
python -m pip install paddleocr
```
### Model Usage
You can quickly experience the functionality with a single command:
```bash
paddleocr text_line_orientation_classification \
--model_name PP-LCNet_x0_25_textline_ori \
-i https://cdn-uploads.huggingface.co/production/uploads/681c1ecd9539bdde5ae1733c/m3ZmUPAnst1f9xXvTVLKS.png
```
You can also integrate the model inference of the text recognition module into your project. Before running the following code, please download the sample image to your local machine.
```python
from paddleocr import TextLineOrientationClassification
model = TextLineOrientationClassification(model_name="PP-LCNet_x0_25_textline_ori")
output = model.predict(input="m3ZmUPAnst1f9xXvTVLKS.png", batch_size=1)
for res in output:
res.print()
res.save_to_img(save_path="./output/")
res.save_to_json(save_path="./output/res.json")
```
After running, the obtained result is as follows:
```json
{'res': {'input_path': '/root/.paddlex/predict_input/m3ZmUPAnst1f9xXvTVLKS.png', 'page_index': None, 'class_ids': array([1], dtype=int32), 'scores': array([0.99671], dtype=float32), 'label_names': ['180_degree']}}
```
The visualized image is as follows:
For details about usage command and descriptions of parameters, please refer to the [Document](https://paddlepaddle.github.io/PaddleOCR/latest/en/version3.x/module_usage/text_recognition.html#iii-quick-start).
## Links
[PaddleOCR Repo](https://github.com/paddlepaddle/paddleocr)
[PaddleOCR Documentation](https://paddlepaddle.github.io/PaddleOCR/latest/en/index.html)
|
liukevin666/blockassist-bc-yawning_striped_cassowary_1756435221
|
liukevin666
| 2025-08-29T02:43:43Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"yawning striped cassowary",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T02:41:16Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- yawning striped cassowary
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
bgloh/q-FrozenLake-v1-4x4-noSlippery
|
bgloh
| 2025-08-29T02:38:44Z | 0 | 0 | null |
[
"FrozenLake-v1-4x4-no_slippery",
"q-learning",
"reinforcement-learning",
"custom-implementation",
"model-index",
"region:us"
] |
reinforcement-learning
| 2025-08-29T02:38:41Z |
---
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="bgloh/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"])
```
|
ilkhom199/a5d6fcda-9c8b-4d7e-8cc0-9322c13809f4
|
ilkhom199
| 2025-08-29T02:38:24Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"llama",
"text-generation",
"llama-factory",
"conversational",
"arxiv:1910.09700",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-08-29T02:37:39Z |
---
library_name: transformers
tags:
- llama-factory
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
|
hobson123/blockassist-bc-mammalian_dense_gibbon_1756434370
|
hobson123
| 2025-08-29T02:27:21Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"mammalian dense gibbon",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T02:27:08Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- mammalian dense gibbon
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
thisiswooyeol/MIGC-diffusers
|
thisiswooyeol
| 2025-08-29T02:24:06Z | 0 | 0 | null |
[
"text-to-image",
"en",
"base_model:CompVis/stable-diffusion-v1-4",
"base_model:finetune:CompVis/stable-diffusion-v1-4",
"license:creativeml-openrail-m",
"region:us"
] |
text-to-image
| 2025-08-29T01:20:13Z |
---
license: creativeml-openrail-m
language:
- en
base_model:
- CompVis/stable-diffusion-v1-4
- limuloo1999/MIGC
pipeline_tag: text-to-image
---
# About file
<!-- Provide a quick summary of what the model is/does. -->
Diffusers version of MIGC adapter state dict. The actual values are identical to the original checkpoint file [MICG_SD14.ckpt](https://huggingface.co/limuloo1999/MIGC)
Please see the details of MIGC in the [MIGC repositiory](https://github.com/limuloo/MIGC).
# How to use
Please use modified pipeline class in `pipeline_stable_diffusion_migc.py` file.
```python
import random
import numpy as np
import safetensors.torch
import torch
from huggingface_hub import hf_hub_download
from pipeline_stable_diffusion_migc import StableDiffusionMIGCPipeline
DEVICE="cuda"
SEED=42
pipe = StableDiffusionMIGCPipeline.from_pretrained("CompVis/stable-diffusion-v1-4").to(DEVICE)
adapter_path = hf_hub_download(repo_id="thisiswooyeol/MIGC-diffusers", filename="migc_adapter_weights.safetensors")
# Load MIGC adapter to UNet attn2 layers
state_dict = safetensors.torch.load_file(adapter_path)
for name, module in pipe.unet.named_modules():
if hasattr(module, "migc"):
print(f"Found MIGC in {name}")
# Get the state dict with the incorrect keys
state_dict_to_load = {k: v for k, v in state_dict.items() if k.startswith(name)}
# Create a new state dict, removing the "attn2." prefix from each key
new_state_dict = {k.replace(f"{name}.migc.", "", 1): v for k, v in state_dict_to_load.items()}
# Load the corrected state dict
module.migc.load_state_dict(new_state_dict)
module.to(device=pipe.unet.device, dtype=pipe.unet.dtype)
# Sample inference !
prompt = "bestquality, detailed, 8k.a photo of a black potted plant and a yellow refrigerator and a brown surfboard"
phrases = [
"a black potted plant",
"a brown surfboard",
"a yellow refrigerator",
]
bboxes = [
[0.5717187499999999, 0.0, 0.8179531250000001, 0.29807511737089204],
[0.85775, 0.058755868544600943, 0.9991875, 0.646525821596244],
[0.6041562500000001, 0.284906103286385, 0.799046875, 0.9898591549295774],
]
def seed_everything(seed):
random.seed(seed)
np.random.seed(seed)
torch.manual_seed(seed)
torch.cuda.manual_seed_all(seed)
seed_everything(SEED)
image = pipe(
prompt=prompt,
phrases=phrases,
bboxes=bboxes,
negative_prompt="worst quality, low quality, bad anatomy",
generator=torch.Generator(DEVICE).manual_seed(SEED),
).images[0]
image.save("image.png")
```
This will generate following image:
If we overlay the box layouts to the generated image, we can see the model successfully follows the layout instructions.
|
Pshuaibi/Reinforce-Cart-Pole-v1
|
Pshuaibi
| 2025-08-29T02:23:57Z | 0 | 0 | null |
[
"CartPole-v1",
"reinforce",
"reinforcement-learning",
"custom-implementation",
"deep-rl-class",
"model-index",
"region:us"
] |
reinforcement-learning
| 2025-08-29T02:23:47Z |
---
tags:
- CartPole-v1
- reinforce
- reinforcement-learning
- custom-implementation
- deep-rl-class
model-index:
- name: Reinforce-Cart-Pole-v1
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: CartPole-v1
type: CartPole-v1
metrics:
- type: mean_reward
value: 404.40 +/- 191.21
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
|
anon-researcher-ua/codegpt-oss-20b
|
anon-researcher-ua
| 2025-08-29T02:23:35Z | 0 | 0 | null |
[
"safetensors",
"gpt_oss",
"code",
"programming-tasks",
"algorithmic-reasoning",
"code-generation",
"python",
"synthetic",
"text-generation",
"conversational",
"en",
"uk",
"es",
"ko",
"ja",
"zh",
"fr",
"hi",
"de",
"dataset:anon-researcher-ua/ua-codeforces-cots-open-r1",
"dataset:anon-researcher-ua/ua-codeforces-cots-open-r1-for-training",
"dataset:open-r1/codeforces-cots",
"base_model:openai/gpt-oss-20b",
"base_model:finetune:openai/gpt-oss-20b",
"license:cc-by-4.0",
"region:us"
] |
text-generation
| 2025-08-28T23:38:23Z |
---
license: cc-by-4.0
language:
- en
- uk
- es
- ko
- ja
- zh
- fr
- hi
- de
task_categories:
- text-generation
- translation
- summarization
tags:
- code
- programming-tasks
- algorithmic-reasoning
- code-generation
- python
- synthetic
datasets:
- anon-researcher-ua/ua-codeforces-cots-open-r1
- anon-researcher-ua/ua-codeforces-cots-open-r1-for-training
- open-r1/codeforces-cots
base_model:
- openai/gpt-oss-20b
pipeline_tag: text-generation
---
# Model Summary
This model is a LoRA-tuned variant of [gpt-oss-20b](https://huggingface.co/openai/gpt-oss-20b) adapted for competitive programming with reasoning traces in Ukrainian and English.
It was trained on ~36M tokens from [ua-codeforces-cots-open-r1](https://huggingface.co/datasets/anon-researcher-ua/ua-codeforces-cots-open-r1) ([ua-codeforces-cots-open-r1-for-training](https://huggingface.co/datasets/anon-researcher-ua/ua-codeforces-cots-open-r1-for-training) in fact), which extends [open-r1/codeforces-cots](https://huggingface.co/datasets/open-r1/codeforces-cots) with translated Ukrainian problem statements and low/medium/high reasoning traces.
The model supports low, medium, and high reasoning effort as well as language control for reasoning via system prompts.
This model is able to reason in other languages as well with proper sampling params ([check examples below](https://huggingface.co/anon-researcher-ua/codegpt-oss-20b#other-languages)).
# Training Details
- Base model: gpt-oss-20b
- Fine-tuning method: LoRA
- Trainable parameters: ~75M
- Target modules: all-linear
- Target parameters: ["7.mlp.experts.gate_up_proj", "7.mlp.experts.down_proj",
"15.mlp.experts.gate_up_proj",
"15.mlp.experts.down_proj",
"21.mlp.experts.gate_up_proj",
"21.mlp.experts.down_proj",
"23.mlp.experts.gate_up_proj",
"23.mlp.experts.down_proj",
]
- LoRA config: r=32, alpha=64, dropout=0.0
- Training tokens: 36M (4k train examples, avg length 7–8k tokens, ~300 >10k tokens)
- Max length: 26496
- Epochs: 3 (early stopping after epoch 2)
- Batching: per_device_batch=1, gradient_accumulation=8 (effective batch=8)
- Optimizer: paged_adamw_8bit
- Learning rate: 5e-5 (cosine schedule, min_lr_rate=0.01)
- Weight decay: 0.01
- Precision: bf16
# Capabilities
- Generates chain-of-thought reasoning traces in English or Ukrainian (works for other languages as well, check examples below)
- Supports low/medium/high reasoning effort (controllable verbosity & detail)
- Solves competitive programming tasks in Python with explanations
- Handles long contexts (up to 32k tokens)
# Limitations
- Dataset is relatively small (~36M tokens, 2 epoch) → model may overfit to training traces.
- Reasoning traces can be verbose or stylistically inconsistent (looped, truncated output for high reasoning effort).
- Only tuned for Python competitive programming tasks, not general coding.
# License
- Our contribution: CC BY 4.0.
- Original data from open-r1/codeforces-cots: ODC-By 4.0.
- Please attribute ua-codeforces-cots-open-r1 and the upstream dataset when you share or use it.
# Usage
## Transformers inference (Slow)
```python
from transformers import AutoModelForCausalLM, AutoTokenizer
from datasets import load_dataset
tokenizer = AutoTokenizer.from_pretrained("anon-researcher-ua/codegpt-oss-20b")
model_kwargs = dict(attn_implementation="flash_attention_2", torch_dtype="auto", use_cache=True, device_map="auto")
model = AutoModelForCausalLM.from_pretrained("anon-researcher-ua/codegpt-oss-20b", **model_kwargs).cuda()
train_dataset = load_dataset("anon-researcher-ua/ua-codeforces-cots-open-r1-for-training")
def format_row(row):
return tokenizer.apply_chat_template(
row["messages_gpt"][:-1],
tokenize=True,
return_tensors="pt",
reasoning_effort=row['reasoning_effort']
)
gen_kwargs = {"max_new_tokens": 32768, "do_sample": True, "temperature": 0.2, "top_p": 1.0, "repetition_penalty": 1}
for ridx, row in enumerate(train_dataset['validation']):
input_ids = format_row(row).to(model.device)
output_ids = model.generate(input_ids, **gen_kwargs)
response = tokenizer.batch_decode(output_ids)[0]
print(gen_kwargs, "\n")
print(response, "\n\n=================================================\n\n")
```
## VLLM inference (Fast)
### Prerequisites
```bash
%pip install --upgrade -qqq uv
%uv pip install --pre vllm==0.10.1.1 --extra-index-url https://wheels.vllm.ai/gpt-oss --extra-index-url https://download.pytorch.org/whl/nightly/cu128 --index-strategy unsafe-best-match --no-cache
```
```python
from vllm import LLM, SamplingParams
from vllm.inputs import TokensPrompt
from transformers import AutoTokenizer
model = "anon-researcher-ua/codegpt-oss-20b"
tokenizer = AutoTokenizer.from_pretrained(model)
llm = LLM(model=model, dtype="bfloat16")
def run_predict(messages, sampling_params, reasoning_effort="low"):
formatted_prompt = tokenizer.apply_chat_template(
messages,
tokenize=False,
add_generation_prompt=True,
reasoning_effort=reasoning_effort
)
input_ids = tokenizer(
formatted_prompt,
add_special_tokens=False
).input_ids
prompt = TokensPrompt(prompt_token_ids=input_ids)
output = llm.generate(prompt, sampling_params)
return output[0].outputs[0].text
sampling_params = SamplingParams(
max_tokens=32768,
temperature=0.2, # recommended temp=1.0 leads to poor performance
top_p=1,
repetition_penalty=1,
stop_token_ids=[200002, 199999, 200012],
skip_special_tokens=False
)
messages = [
{"role": "system", "content": "reasoning language: Ukrainian"},
{"role": "user", "content": "Обчисліть площу бічної поверхні правильної трикутної піраміди, сторона основи якої дорівнює 8 см, а апофема на 2 см більша за сторону основи піраміди."},
]
print(run_predict(messages, sampling_params, reasoning_effort="medium"))
```
## Reasoning language: Ukrainian, reasoning effort: medium
```
<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.
Knowledge cutoff: 2024-06
Current date: 2025-08-29
Reasoning: medium
# Valid channels: analysis, commentary, final. Channel must be included for every message.<|end|><|start|>developer<|message|># Instructions
reasoning language: Ukrainian
<|end|><|start|>user<|message|>Обчисліть площу бічної поверхні правильної трикутної піраміди, сторона основи якої дорівнює 8 см, а апофема на 2 см більша за сторону основи піраміди.<|end|><|start|>assistant
<|channel|>analysis<|message|>Задача: обчислити площу бічної поверхні правильної трикутної піраміди. Сторона основи = 8 см. Апофема (висота бічної поверхні) = 8 + 2 = 10 см. Правильна трикутна піраміда: основа рівносторонній трикутнику. Бічна поверхня складається з трикутників, кожен з яких має однакові сторони. Бічна поверхня піраміди: 3 * (площа одного трикутника).
Площа трикутника: (1/2) * a * h, де a — сторона трикутника, h — висота трикутника. Отже, площа бічної поверхні = 3 * (1/2 * a * h) = (3a h) / 2.
Підставляємо a = 8, h = 10: (3 * 8 * 10) / 2 = 240 / 2 = 120 см².
Потрібно вивести результат. Формат: "120 см²" (або "120 см²" з правильною одиницею). У деяких стандартах використовують "см²" або "см²". В українській мові одиниця площі позначається "см²" (або "см²"). Але в тексті задачі використано "см". Тому вивід має бути "120 см²".
Підтвердити, що інші параметри (наприклад, висота основи піраміди) не впливають на площу бічної поверхні. Ні, оскільки апофема — це висота бічної поверхні.
Отже, правильна відповідь: 120 см².
Вивести: "120 см²".
Підтвердити, що апофема — це висота бічної поверхні, а не висота основи. У задачі сказано "апофема на 2 см більша за сторону основи піраміди". Отже, 10 см.
Тож відповідь: 120 см².<|end|><|start|>assistant<|channel|>final<|message|>Нам потрібно обчислити площу бічної поверхні правильної трикутної піраміди.
- Сторона основи \(a = 8\) см.
- Апофема (висота бічної поверхні) \(h = a + 2 = 8 + 2 = 10\) см.
Бічна поверхня піраміди складається з трьох трикутників, кожен з яких має однакові сторони. Загальна площа бічної поверхні обчислюється як:
\[
\text{Площа} = 3 \times \frac{1}{2} \times a \times h = \frac{3a h}{2}
\]
Підставляємо відомі значення:
\[
\frac{3 \times 8 \times 10}{2} = \frac{240}{2} = 120 \text{ см}^2
\]
Отже, площа бічної поверхні правильної трикутної піраміди становить **120 см²**.
```
## Reasoning language: English, reasoning effort: high
```
<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.
Knowledge cutoff: 2024-06
Current date: 2025-08-29
Reasoning: high
# Valid channels: analysis, commentary, final. Channel must be included for every message.<|end|><|start|>developer<|message|># Instructions
reasoning language: English
<|end|><|start|>user<|message|>Обчисліть площу бічної поверхні правильної трикутної піраміди, сторона основи якої дорівнює 8 см, а апофема на 2 см більша за сторону основи піраміди.<|end|><|start|>assistant
<|channel|>analysis<|message|>Обчисліть площу бічної поверхні правильної трикутної піраміди, де сторона основи дорівнює 8 см, а апофема на 2 см більша за сторону основи піраміди. Площа бічної поверхні складається з трикутників, кожен з яких має однакову висоту (висоту трикутника). Сума всіх трикутників — це бічна поверхня. Площа бічної поверхні — це сума площ усіх трикутників.
Завдання: знайти площу бічної lateral
We need to compute the lateral surface area of a regular triangular pyramid (tetrahedron) with base side length 8 cm. The slant height (apothem?) is 2 cm longer than the base side. Wait: "апофема на 2 см більша за сторону основи піраміди піраміди" maybe apothem of the pyramid (slant height) is 2 cm greater than the base side. So slant height = 8 + 2 = 10 cm. But need to confirm what exactly the apothem refers to here.
The lateral surface area of a regular triangular pyramid (tetrahedron) is the sum of the areas of the three lateral faces. Each lateral face is a triangle with base side length 8 and height equal to the slant height. So each triangle area is (1/2 * base * height). So each triangle is 1/2 * 8 * 10 = 40. Multiply by 3 gives 120 cm².
But maybe the apothem refers to the apothem of the base (the base is a regular triangle). Wait, the problem statement: "площа бічної поверхні правильної трикутної піраміди, сторона основи якої дорівнює 8 см, а апофема на 2 см більша за сторону основи піраміди піраміди." So the apothem (apothema) of the pyramid is 2 cm greater than the side of the base. So slant height (apothem of the pyramid) = 8 + 2 = 10. So each lateral face is a triangle with base 8 and height 10. So each area is (1/2*8*10) = 40. 3 faces => 120.
But maybe there is a trick: the apothem of the pyramid is the slant height of the lateral faces. So yes.
But let's double-check the geometry. For a regular triangular pyramid, each lateral face is a triangle with base side length equal to the base side of the pyramid. The height of each lateral face is the slant height (apothem) of the pyramid. So yes.
Thus the lateral surface area = 3 * (1/2 * base * slant height) = 3 * (1/2 * 8 * 10) = 3 * 40 = 120.
But maybe the problem is not that trivial. Let's read the original Ukrainian text: "Обчисліть площу бічної поверхні правильної трикутної піраміди, сторона основи якої дорівнює 8 см, а апофема на 2 см більша за сторону основи піраміди піраміди. Площа бічної поверхні складається з трикутників, кожен з яких має однакову висоту (висоту трикутника). Сума всіх трикутників — це бічна поверхня. Площа бічної поверхні — це сума площ усіх трикутників."
So the lateral surface area is the sum of the areas of the three triangles. Each triangle has same height (the height of the triangle). The height of the triangle is the slant height of the pyramid. So each triangle area is (1/2 * base side * height). So indeed 1/2 * 8 * (8+2) = 1/2 * 8 * 10 = 40. Then sum for 3 triangles is 3 * 40 = 120.
But let's check if there's any other interpretation. For a regular triangular pyramid, the base is a regular triangle. The apothem of the pyramid is the slant height of the pyramid. So yes.
Alternatively, maybe "апофема" could be the apothem of the base (the distance from the center of the base to a side). But the problem says "апофема на 2 см більша за сторону основи піраміди піраміди" which suggests it's the apothem of the pyramid (the slant height). So yes.
Thus the answer should be 120 cm².
But maybe the problem expects a different approach. Let's re-read the problem:
"Обчисліть площу бічної поверхні правильної трикутної піраміди, сторона основи якої дорівнює 8 см, а апофема на 2 см більша за сторону основи піраміди піраміди. Площа бічної поверхні складається з трикутників, кожен з яких має однакову висоту (висоту трикутника). Сума всіх трикутників — це бічна поверхня. Площа бічної поверхні — це сума площ усіх трикутників."
So yes, it's exactly that. So answer is 120 cm².
But maybe the apothem is something else. Let's think: In a pyramid, the apothem (slant height) is the height of each lateral face. So each lateral face is a triangle with height equal to the apothem. So yes.
Thus each triangle area is 1/2 * base * apothem. So 1/2*8*10 = 40. Sum 3*40 = 120.
Thus answer: 120 cm².
But wait, the problem might be a trick: maybe the apothem is the apothem of the base (the distance from the center to a side). But the problem says "апофема на 2 см більша за сторону основи піраміди піраміди." So it's not the base's apothem. It's the pyramid's apothem. So yes.
So answer: 120.
But let's check the geometry: The lateral faces of a regular pyramid are congruent isosceles triangles. Each has base side length equal to the base side of the pyramid. The height of each lateral face is the slant height (apothem) of the pyramid. So yes.
Thus the lateral surface area is 3 * (1/2 * base side * slant height) = 3 * (1/2 * 8 * 10) = 3 * 40 = 120.
Thus answer: 120 cm².
But let's double-check that there is no confusion between apothem and slant height. In a pyramid, the apothem is the slant height. So yes.
Thus answer: 120.
But maybe the problem expects the answer in terms of cm². So 120 cm².
Thus the correct answer is 120 cm².
But let's think if there's any other possible misinterpretation. For example, maybe the apothem is the apothem of the base (the distance from the center of the base to a side). But the problem says "апофема на 2 см більша за сторону основи піраміди піраміди." So it's the apothem of the pyramid. So no.
Thus answer: 120 cm².
So final answer: 120 cm².
But let's check the original language: "правильної трикутної піраміди" which is a regular triangular pyramid. So yes.
Thus answer: 120 cm².
Thus the correct answer is 120. But maybe the units: cm². So 120 cm².
Thus final answer: 120 cm².
But let's check if there's any other geometry: For the lateral faces, each is a triangle with base side 8 and height 10. So each area is 40. So sum 3*40 = 120.
Yes.
Thus answer: 120 cm².
Thus the solution is 120 cm².
Thus the correct answer is 120.
Thus final answer: 120.
Thus the answer: 120 cm².
Thus the solution is 120.
So the answer: 120.
Thus the correct answer is 120.
So final answer: 120 cm².<|end|><|start|>assistant<|channel|>final<|message|>The lateral surface area of a regular triangular pyramid is the sum of the areas of its three lateral faces. Each face is a triangle with a base equal to the side of the pyramid (8 cm) and a height equal to the pyramid's apothem (10 cm, since it is 2 cm greater than the base side).
Area of one triangle:
\[
\frac{1}{2} \times 8 \, \text{cm} \times 10 \, \text{cm} = 40 \, \text{cm}^2
\]
Since there are three such triangles:
\[
3 \times 40 \, \text{cm}^2 = 120 \, \text{cm}^2
\]
\[
\boxed{120 \text{ cm}^2}
\]
```
# Other languages
## Reasoning language: Korean, reasoning effort: low
```
<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.
Knowledge cutoff: 2024-06
Current date: 2025-08-29
Reasoning: low
# Valid channels: analysis, commentary, final. Channel must be included for every message.<|end|><|start|>developer<|message|># Instructions
reasoning language: Korean
<|end|><|start|>user<|message|>Обчисліть площу бічної поверхні правильної трикутної піраміди, сторона основи якої дорівнює 8 см, а апофема на 2 см більша за сторону основи піраміди.<|end|><|start|>assistant
<|channel|>analysis<|message|>문제는 정삼각 피라미드의 옆면 표면적을 구하는 것. 밑면의 한 변이 8cm, 각 옆면의 높이(apofema)가 밑면 변보다 2cm 더 크다. 정삼각 피라미드의 옆면은 정삼각형이므로, 옆면의 높이와 밑면 변이 주어지면 각 옆면의 넓이를 구할 수 있다. 옆면 넓이 = (밑변 * 높이)/2. 주어진 값으로 대입하면 (8 * (8+2))/2 = 40cm². 피라미드의 옆면은 3개이므로 총 옆면 표면적은 3 * 40 = 120cm². 따라서 정답은 120cm². 출력 형식은 "120 cm²" 또는 "120 cm²"와 같은 형태로 제공.
정리: 120 cm².<|end|><|start|>assistant<|channel|>final<|message|>Обчислимо площу бічної поверхні трикутної піраміди. Для правильної трикутної піраміди кожна бічна поверхня є рівнобедрним трикутником.
1. **Дано:**
- Сторона основи піраміди: \(a = 8\) см
- Апофема (висота бічної поверхні): \(h = a + 2 = 8 + 2 = 10\) см
2. **Площа одного бічного трикутника:**
\[
S_{\text{бічна}} = \frac{a \cdot h}{2} = \frac{8 \cdot 10}{2} = 40 \text{ см}^2
\]
3. **Кількість бічних трикутників:**
Для правильної трикутної піраміди трикутна основа має три бічні трикутники.
4. **Загальна площа бічної поверхні:**
\[
S_{\text{загальна}} = 3 \cdot S_{\text{бічна}} = 3 \cdot 40 = 120 \text{ см}^2
\]
Отже, площа бічної поверхні правильної трикутної піраміди становить **120 см²**.
```
## Reasoning language: Spanish, reasoning effort: medium (repetition_penalty=1.1)
```
<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.
Knowledge cutoff: 2024-06
Current date: 2025-08-29
Reasoning: medium
# Valid channels: analysis, commentary, final. Channel must be included for every message.<|end|><|start|>developer<|message|># Instructions
Reasoning language: Spanish
<|end|><|start|>user<|message|>You’re given a competitive programming problem.
Analyze the maximum input constraints and identify the optimal algorithmic approach and data structures needed to process the largest possible test cases within the time and memory limits, then explain why your chosen implementation strategy is the most efficient solution. Please reason step by step about your solution approach, then provide a complete implementation in Python 3 that is thoroughly optimized for both speed and memory usage.
Your solution must read input from standard input (input()), write output to standard output (print()).
Do not include any debug prints or additional output.
Put your final solution within a single code block:
'''python
<your code here>
'''
# Problem
You're given an integer $$$n$$$. For every integer $$$i$$$ from $$$2$$$ to $$$n$$$, assign a positive integer $$$a_i$$$ such that the following conditions hold:
- For any pair of integers $$$(i,j)$$$, if $$$i$$$ and $$$j$$$ are coprime, $$$a_i
eq a_j$$$.
- The maximal value of all $$$a_i$$$ should be minimized (that is, as small as possible).
A pair of integers is called coprime if their greatest common divisor is $$$1$$$.
## Constraints
Time limit per test: 1.0 seconds
Memory limit per test: 256.0 megabytes
## Input Format
The only line contains the integer $$$n$$$ ($$$2 \le n \le 10^5$$$).
## Output Format
Print $$$n-1$$$ integers, $$$a_2$$$, $$$a_3$$$, $$$\ldots$$$, $$$a_n$$$ ($$$1 \leq a_i \leq n$$$).
If there are multiple solutions, print any of them.
## Examples
'''input
4
'''
'''output
1 2 1
'''
'''input
3
'''
'''output
2 1
'''<|end|><|start|>assistant
<|channel|>analysis<|message|>Para resolver este problema necesitamos asignar a cada entero i en el rango [2..n] un valor positivo a_i tal que si dos números son coprimos entonces sus valores sean distintos y el máximo de todos los valores sea lo más pequeño posible.
Observaciones clave:
1. Si dos números comparten un factor primo p, pueden compartir el mismo valor porque no son coprimos. Por lo tanto, podemos agrupar los enteros según los factores primos comunes.
2. Un algoritmo eficiente para generar estos grupos es usar la técnica de **sieve** (cribado). Para cada número primo p, marcamos todos sus múltiplos y los agrupamos bajo el mismo valor.
3. El proceso:
- Inicializamos una lista `a` con ceros del tamaño n+1.
- Usamos una lista booleana `is_prime` para identificar los primos.
- Iteramos desde 2 hasta n. Si el número es primo, asignamos un nuevo valor único a todos sus múltiplos que aún no hayan sido asignados.
- Este método garantiza que cualquier par coprimo tendrá valores diferentes y minimiza el máximo valor asignado.
4. Complejidad: O(n log log n), adecuada para n ≤ 100000.
Ejemplo:
- Para n=4, el resultado es 1 2 1 (maximo 2).
- Para n=3, el resultado es 2 1 (maximo 2).
Implementación en Python:
- Se utiliza un sieve clásico para marcar primos.
- Se mantiene un contador `current_val` para asignar valores únicos a cada grupo de múltiplos de un primo.
- Al finalizar, se imprime la lista `a[2:]`.
Código optimizado:
'''python
n = int(input())
a = [0] * (n + 1)
is_prime = [True] * (n + 1)
if n >= 0:
is_prime[0] = False
if n >= 1:
is_prime[1] = False
for i in range(2, int(n**0.5) + 1):
if is_prime[i]:
for j in range(i*i, n + 1, i):
is_prime[j] = False
current_val = 1
for p in range(2, n + 1):
if is_prime[p]:
for m in range(p, n + 1, p):
if a[m] == 0:
a[m] = current_val
current_val += 1
print(' '.join(map(str, a[2:n+1]))
'''
Esta solución cumple con las condiciones del problema y funciona eficientemente dentro de los límites dados.<|end|><|start|>assistant<|channel|>final<|message|>To solve this problem, we need to assign values to each integer \(i\) from 2 to \(n\) such that if two numbers are coprimes, they receive different values, and the maximum assigned value is minimized.
### Approach
The key insight is to use the properties of prime factors. If two numbers share a common prime factor, they can have the same value because they are not coprimes. We can group numbers based on their smallest prime factor using a sieve-like approach. This ensures that any two coprimes will belong to different groups and thus have different values.
### Algorithm
1. Use a sieve to find all primes up to \(n\).
2. Assign a unique value to each prime and its multiples that haven't been assigned yet.
3. This ensures that numbers sharing a prime factor get the same value, while coprimes get different values.
### Complexity
The algorithm runs in \(O(n \log \log n)\) due to the sieve operation, which is efficient for \(n\) up to \(10^5\).
### Solution Code
Here is the Python implementation of the described approach:
'''python
def main():
import sys
n = int(sys.stdin.readline())
# Initialize array to store the values for each number
a = [0] * (n + 1)
# Sieve of Eratosthenes to find primes
is_prime = [True] * (n + 1)
if n >= 0:
is_prime[0] = False
if n >= 1:
is_prime[1] = False
for i in range(2, int(n**0.5) + 1):
if is_prime[i]:
for j in range(i * i, n + 1, i):
is_prime[j] = False
current_value = 1
for p in range(2, n + 1):
if is_prime[p]:
# Assign current_value to all multiples of p that are not yet assigned
for m in range(p, n + 1, p):
if a[m] == 0:
a[m] = current_value
current_value += 1
# Output the result for a[2] to a[n]
print(' '.join(map(str, a[2:n+1]))
if __name__ == "__main__":
main()
'''
This code reads the input, computes the necessary values using a sieve-based approach, assigns values efficiently, and outputs the results ensuring that the maximum value is minimized while satisfying the coprimality condition.
```
|
katanyasekolah/blockassist-bc-silky_sprightly_cassowary_1756432426
|
katanyasekolah
| 2025-08-29T02:21:55Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"silky sprightly cassowary",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T02:21:52Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- silky sprightly cassowary
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
felixZzz/student_sft_len32k_sub1k_nonoverlap_multiZ_acc-0829
|
felixZzz
| 2025-08-29T02:21:49Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"qwen2",
"text-generation",
"conversational",
"arxiv:1910.09700",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-08-29T02:16:55Z |
---
library_name: transformers
tags: []
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
|
MauoSama/DP_depth_cut_rope_randomized_onlywrist
|
MauoSama
| 2025-08-29T02:20:56Z | 0 | 0 |
lerobot
|
[
"lerobot",
"safetensors",
"robotics",
"diffusion",
"dataset:MauoSama/depth_cut_rope_randomized_onlywrist",
"arxiv:2303.04137",
"license:apache-2.0",
"region:us"
] |
robotics
| 2025-08-29T02:20:49Z |
---
datasets: MauoSama/depth_cut_rope_randomized_onlywrist
library_name: lerobot
license: apache-2.0
model_name: diffusion
pipeline_tag: robotics
tags:
- robotics
- diffusion
- lerobot
---
# Model Card for diffusion
<!-- Provide a quick summary of what the model is/does. -->
[Diffusion Policy](https://huggingface.co/papers/2303.04137) treats visuomotor control as a generative diffusion process, producing smooth, multi-step action trajectories that excel at contact-rich manipulation.
This policy has been trained and pushed to the Hub using [LeRobot](https://github.com/huggingface/lerobot).
See the full documentation at [LeRobot Docs](https://huggingface.co/docs/lerobot/index).
---
## How to Get Started with the Model
For a complete walkthrough, see the [training guide](https://huggingface.co/docs/lerobot/il_robots#train-a-policy).
Below is the short version on how to train and run inference/eval:
### Train from scratch
```bash
python -m lerobot.scripts.train \
--dataset.repo_id=${HF_USER}/<dataset> \
--policy.type=act \
--output_dir=outputs/train/<desired_policy_repo_id> \
--job_name=lerobot_training \
--policy.device=cuda \
--policy.repo_id=${HF_USER}/<desired_policy_repo_id>
--wandb.enable=true
```
_Writes checkpoints to `outputs/train/<desired_policy_repo_id>/checkpoints/`._
### Evaluate the policy/run inference
```bash
python -m lerobot.record \
--robot.type=so100_follower \
--dataset.repo_id=<hf_user>/eval_<dataset> \
--policy.path=<hf_user>/<desired_policy_repo_id> \
--episodes=10
```
Prefix the dataset repo with **eval\_** and supply `--policy.path` pointing to a local or hub checkpoint.
---
## Model Details
- **License:** apache-2.0
|
bluparsons/BART-base-HumAID
|
bluparsons
| 2025-08-29T02:20:24Z | 10 | 0 | null |
[
"safetensors",
"bart",
"en",
"dataset:bluparsons/HumAIDSum",
"base_model:facebook/bart-base",
"base_model:finetune:facebook/bart-base",
"license:cc-by-nc-sa-4.0",
"region:us"
] | null | 2025-08-06T22:27:27Z |
---
license: cc-by-nc-sa-4.0
datasets:
- bluparsons/HumAIDSum
language:
- en
base_model:
- facebook/bart-base
---
#### Release as part of the paper Leveraging Language Models for Temporal Event Detection and Text Summarisation from Natural Disaster Incidents - An Insight using Social Media Data from Twitter/X
|
liukevin666/blockassist-bc-yawning_striped_cassowary_1756433893
|
liukevin666
| 2025-08-29T02:19:22Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"yawning striped cassowary",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T02:19:14Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- yawning striped cassowary
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
feVeRin/enko-transliterator
|
feVeRin
| 2025-08-29T02:19:09Z | 18 | 0 | null |
[
"safetensors",
"marian",
"license:apache-2.0",
"region:us"
] | null | 2025-08-27T03:51:38Z |
---
license: apache-2.0
---
# Please Check: https://github.com/feVeRin/enko_transliterator
```python
# git clone https://github.com/feVeRin/enko_transliterator.git
from transliteration import Transliterator
model = Transliterator.from_pretrained('feVeRin/enko-transliterator')
result = model.transliterate('LORA IS ALL YOU NEED')
```
|
hobson123/blockassist-bc-mammalian_dense_gibbon_1756433745
|
hobson123
| 2025-08-29T02:16:53Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"mammalian dense gibbon",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T02:16:38Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- mammalian dense gibbon
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
pempekmangedd/blockassist-bc-patterned_sturdy_dolphin_1756432322
|
pempekmangedd
| 2025-08-29T02:15:56Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"patterned sturdy dolphin",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T02:15:52Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- patterned sturdy dolphin
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
aleebaster/blockassist-bc-sly_eager_boar_1756431832
|
aleebaster
| 2025-08-29T02:11:25Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"sly eager boar",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T02:11:21Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- sly eager boar
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
Dnq2025/swin-large-mask2former-pretrain-PAM
|
Dnq2025
| 2025-08-29T02:10:02Z | 0 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"mask2former",
"image-segmentation",
"vision",
"generated_from_trainer",
"base_model:Dnq2025/swin-large-mask2former-finetuned-ER-Mito-LD8",
"base_model:finetune:Dnq2025/swin-large-mask2former-finetuned-ER-Mito-LD8",
"license:other",
"endpoints_compatible",
"region:us"
] |
image-segmentation
| 2025-08-29T02:01:54Z |
---
library_name: transformers
license: other
base_model: Dnq2025/swin-large-mask2former-finetuned-ER-Mito-LD8
tags:
- image-segmentation
- vision
- generated_from_trainer
model-index:
- name: swin-large-mask2former-pretrain-PAM
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. -->
# swin-large-mask2former-pretrain-PAM
This model is a fine-tuned version of [Dnq2025/swin-large-mask2former-finetuned-ER-Mito-LD8](https://huggingface.co/Dnq2025/swin-large-mask2former-finetuned-ER-Mito-LD8) on the Dnq2025/Mask2former_Pretrain_PAM dataset.
It achieves the following results on the evaluation set:
- Mean Iou: 0.4801
- Loss: 75.6712
## 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: 0.0004
- train_batch_size: 4
- eval_batch_size: 4
- seed: 1337
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: polynomial
- training_steps: 672
### Training results
| Training Loss | Epoch | Step | Mean Iou | Validation Loss |
|:-------------:|:-----:|:----:|:--------:|:---------------:|
| 86.7074 | 1.0 | 14 | 0.0649 | 73.8587 |
| 60.1158 | 2.0 | 28 | 0.4315 | 61.3078 |
| 47.1427 | 3.0 | 42 | 0.3416 | 65.8193 |
| 39.335 | 4.0 | 56 | 0.4496 | 62.5412 |
| 33.9598 | 5.0 | 70 | 0.5278 | 51.6471 |
| 31.4424 | 6.0 | 84 | 0.4708 | 62.5557 |
| 29.2716 | 7.0 | 98 | 0.4481 | 69.7683 |
| 27.8051 | 8.0 | 112 | 0.4917 | 51.5576 |
| 26.4282 | 9.0 | 126 | 0.5110 | 50.2866 |
| 25.0882 | 10.0 | 140 | 0.4807 | 66.0714 |
| 23.8018 | 11.0 | 154 | 0.4920 | 49.7601 |
| 23.3043 | 12.0 | 168 | 0.4963 | 51.5394 |
| 22.6939 | 13.0 | 182 | 0.4844 | 69.5330 |
| 21.9263 | 14.0 | 196 | 0.5126 | 54.2508 |
| 21.2466 | 15.0 | 210 | 0.5297 | 47.1237 |
| 20.9495 | 16.0 | 224 | 0.5158 | 68.3420 |
| 20.2615 | 17.0 | 238 | 0.5164 | 51.2259 |
| 19.7151 | 18.0 | 252 | 0.5028 | 60.2969 |
| 19.3651 | 19.0 | 266 | 0.4946 | 63.1645 |
| 18.9289 | 20.0 | 280 | 0.5142 | 56.4129 |
| 18.4501 | 21.0 | 294 | 0.5048 | 69.7263 |
| 17.9105 | 22.0 | 308 | 0.4764 | 73.4659 |
| 17.7714 | 23.0 | 322 | 0.4847 | 65.6970 |
| 17.5999 | 24.0 | 336 | 0.5102 | 55.8619 |
| 17.2178 | 25.0 | 350 | 0.4744 | 64.8820 |
| 17.0504 | 26.0 | 364 | 0.4817 | 67.6332 |
| 16.9627 | 27.0 | 378 | 0.4852 | 77.5776 |
| 16.4976 | 28.0 | 392 | 0.4684 | 76.7091 |
| 16.0761 | 29.0 | 406 | 0.5068 | 55.3819 |
| 15.9957 | 30.0 | 420 | 0.4749 | 72.6874 |
| 15.9871 | 31.0 | 434 | 0.5005 | 77.6782 |
| 15.8228 | 32.0 | 448 | 0.4775 | 75.3537 |
| 15.4886 | 33.0 | 462 | 0.4907 | 76.7841 |
| 15.1934 | 34.0 | 476 | 0.4583 | 78.3322 |
| 15.3006 | 35.0 | 490 | 0.4911 | 69.4165 |
| 15.0434 | 36.0 | 504 | 0.5020 | 56.2806 |
| 14.9141 | 37.0 | 518 | 0.4672 | 71.5419 |
| 14.8351 | 38.0 | 532 | 0.5014 | 60.7283 |
| 14.5805 | 39.0 | 546 | 0.4910 | 65.7086 |
| 14.702 | 40.0 | 560 | 0.4830 | 70.8683 |
| 14.4871 | 41.0 | 574 | 0.4875 | 62.4533 |
| 14.3313 | 42.0 | 588 | 0.4789 | 69.2328 |
| 14.0295 | 43.0 | 602 | 0.4843 | 69.5139 |
| 13.996 | 44.0 | 616 | 0.4807 | 76.6653 |
| 14.0995 | 45.0 | 630 | 0.4840 | 67.6556 |
| 13.9247 | 46.0 | 644 | 0.4818 | 71.6413 |
| 14.0587 | 47.0 | 658 | 0.4804 | 79.2972 |
| 13.894 | 48.0 | 672 | 0.4801 | 76.0411 |
### Framework versions
- Transformers 4.50.0.dev0
- Pytorch 2.4.1
- Datasets 3.3.2
- Tokenizers 0.21.0
|
xingqiang/Medical-NER-Qwen-4B-Thinking-merged
|
xingqiang
| 2025-08-29T02:08:33Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"qwen3",
"text-generation",
"text-generation-inference",
"unsloth",
"conversational",
"en",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-08-29T01:49:03Z |
---
base_model: unsloth/qwen3-4b-thinking-2507-unsloth-bnb-4bit
tags:
- text-generation-inference
- transformers
- unsloth
- qwen3
license: apache-2.0
language:
- en
---
# Uploaded finetuned model
- **Developed by:** xingqiang
- **License:** apache-2.0
- **Finetuned from model :** unsloth/qwen3-4b-thinking-2507-unsloth-bnb-4bit
This qwen3 model was trained 2x faster with [Unsloth](https://github.com/unslothai/unsloth) and Huggingface's TRL library.
[<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png" width="200"/>](https://github.com/unslothai/unsloth)
|
mradermacher/Impish_LLAMA_4B-i1-GGUF
|
mradermacher
| 2025-08-29T02:08:25Z | 0 | 0 |
transformers
|
[
"transformers",
"gguf",
"en",
"dataset:SicariusSicariiStuff/UBW_Tapestries",
"base_model:SicariusSicariiStuff/Impish_LLAMA_4B",
"base_model:quantized:SicariusSicariiStuff/Impish_LLAMA_4B",
"license:llama3.1",
"endpoints_compatible",
"region:us",
"imatrix",
"conversational"
] | null | 2025-08-29T00:14:22Z |
---
base_model: SicariusSicariiStuff/Impish_LLAMA_4B
datasets:
- SicariusSicariiStuff/UBW_Tapestries
language:
- en
library_name: transformers
license: llama3.1
mradermacher:
readme_rev: 1
quantized_by: mradermacher
---
## About
<!-- ### quantize_version: 2 -->
<!-- ### output_tensor_quantised: 1 -->
<!-- ### convert_type: hf -->
<!-- ### vocab_type: -->
<!-- ### tags: nicoboss -->
<!-- ### quants: Q2_K IQ3_M Q4_K_S IQ3_XXS Q3_K_M small-IQ4_NL Q4_K_M IQ2_M Q6_K IQ4_XS Q2_K_S IQ1_M Q3_K_S IQ2_XXS Q3_K_L IQ2_XS Q5_K_S IQ2_S IQ1_S Q5_K_M Q4_0 IQ3_XS Q4_1 IQ3_S -->
<!-- ### quants_skip: -->
<!-- ### skip_mmproj: -->
weighted/imatrix quants of https://huggingface.co/SicariusSicariiStuff/Impish_LLAMA_4B
<!-- provided-files -->
***For a convenient overview and download list, visit our [model page for this model](https://hf.tst.eu/model#Impish_LLAMA_4B-i1-GGUF).***
static quants are available at https://huggingface.co/mradermacher/Impish_LLAMA_4B-GGUF
## Usage
If you are unsure how to use GGUF files, refer to one of [TheBloke's
READMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) for
more details, including on how to concatenate multi-part files.
## Provided Quants
(sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants)
| Link | Type | Size/GB | Notes |
|:-----|:-----|--------:|:------|
| [GGUF](https://huggingface.co/mradermacher/Impish_LLAMA_4B-i1-GGUF/resolve/main/Impish_LLAMA_4B.imatrix.gguf) | imatrix | 0.1 | imatrix file (for creating your own qwuants) |
| [GGUF](https://huggingface.co/mradermacher/Impish_LLAMA_4B-i1-GGUF/resolve/main/Impish_LLAMA_4B.i1-IQ1_S.gguf) | i1-IQ1_S | 1.3 | for the desperate |
| [GGUF](https://huggingface.co/mradermacher/Impish_LLAMA_4B-i1-GGUF/resolve/main/Impish_LLAMA_4B.i1-IQ1_M.gguf) | i1-IQ1_M | 1.4 | mostly desperate |
| [GGUF](https://huggingface.co/mradermacher/Impish_LLAMA_4B-i1-GGUF/resolve/main/Impish_LLAMA_4B.i1-IQ2_XXS.gguf) | i1-IQ2_XXS | 1.5 | |
| [GGUF](https://huggingface.co/mradermacher/Impish_LLAMA_4B-i1-GGUF/resolve/main/Impish_LLAMA_4B.i1-IQ2_XS.gguf) | i1-IQ2_XS | 1.6 | |
| [GGUF](https://huggingface.co/mradermacher/Impish_LLAMA_4B-i1-GGUF/resolve/main/Impish_LLAMA_4B.i1-IQ2_S.gguf) | i1-IQ2_S | 1.7 | |
| [GGUF](https://huggingface.co/mradermacher/Impish_LLAMA_4B-i1-GGUF/resolve/main/Impish_LLAMA_4B.i1-IQ2_M.gguf) | i1-IQ2_M | 1.8 | |
| [GGUF](https://huggingface.co/mradermacher/Impish_LLAMA_4B-i1-GGUF/resolve/main/Impish_LLAMA_4B.i1-Q2_K_S.gguf) | i1-Q2_K_S | 1.8 | very low quality |
| [GGUF](https://huggingface.co/mradermacher/Impish_LLAMA_4B-i1-GGUF/resolve/main/Impish_LLAMA_4B.i1-Q2_K.gguf) | i1-Q2_K | 1.9 | IQ3_XXS probably better |
| [GGUF](https://huggingface.co/mradermacher/Impish_LLAMA_4B-i1-GGUF/resolve/main/Impish_LLAMA_4B.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 2.0 | lower quality |
| [GGUF](https://huggingface.co/mradermacher/Impish_LLAMA_4B-i1-GGUF/resolve/main/Impish_LLAMA_4B.i1-IQ3_XS.gguf) | i1-IQ3_XS | 2.1 | |
| [GGUF](https://huggingface.co/mradermacher/Impish_LLAMA_4B-i1-GGUF/resolve/main/Impish_LLAMA_4B.i1-Q3_K_S.gguf) | i1-Q3_K_S | 2.2 | IQ3_XS probably better |
| [GGUF](https://huggingface.co/mradermacher/Impish_LLAMA_4B-i1-GGUF/resolve/main/Impish_LLAMA_4B.i1-IQ3_S.gguf) | i1-IQ3_S | 2.2 | beats Q3_K* |
| [GGUF](https://huggingface.co/mradermacher/Impish_LLAMA_4B-i1-GGUF/resolve/main/Impish_LLAMA_4B.i1-IQ3_M.gguf) | i1-IQ3_M | 2.3 | |
| [GGUF](https://huggingface.co/mradermacher/Impish_LLAMA_4B-i1-GGUF/resolve/main/Impish_LLAMA_4B.i1-Q3_K_M.gguf) | i1-Q3_K_M | 2.4 | IQ3_S probably better |
| [GGUF](https://huggingface.co/mradermacher/Impish_LLAMA_4B-i1-GGUF/resolve/main/Impish_LLAMA_4B.i1-Q3_K_L.gguf) | i1-Q3_K_L | 2.6 | IQ3_M probably better |
| [GGUF](https://huggingface.co/mradermacher/Impish_LLAMA_4B-i1-GGUF/resolve/main/Impish_LLAMA_4B.i1-IQ4_XS.gguf) | i1-IQ4_XS | 2.6 | |
| [GGUF](https://huggingface.co/mradermacher/Impish_LLAMA_4B-i1-GGUF/resolve/main/Impish_LLAMA_4B.i1-Q4_0.gguf) | i1-Q4_0 | 2.8 | fast, low quality |
| [GGUF](https://huggingface.co/mradermacher/Impish_LLAMA_4B-i1-GGUF/resolve/main/Impish_LLAMA_4B.i1-IQ4_NL.gguf) | i1-IQ4_NL | 2.8 | prefer IQ4_XS |
| [GGUF](https://huggingface.co/mradermacher/Impish_LLAMA_4B-i1-GGUF/resolve/main/Impish_LLAMA_4B.i1-Q4_K_S.gguf) | i1-Q4_K_S | 2.8 | optimal size/speed/quality |
| [GGUF](https://huggingface.co/mradermacher/Impish_LLAMA_4B-i1-GGUF/resolve/main/Impish_LLAMA_4B.i1-Q4_K_M.gguf) | i1-Q4_K_M | 2.9 | fast, recommended |
| [GGUF](https://huggingface.co/mradermacher/Impish_LLAMA_4B-i1-GGUF/resolve/main/Impish_LLAMA_4B.i1-Q4_1.gguf) | i1-Q4_1 | 3.0 | |
| [GGUF](https://huggingface.co/mradermacher/Impish_LLAMA_4B-i1-GGUF/resolve/main/Impish_LLAMA_4B.i1-Q5_K_S.gguf) | i1-Q5_K_S | 3.3 | |
| [GGUF](https://huggingface.co/mradermacher/Impish_LLAMA_4B-i1-GGUF/resolve/main/Impish_LLAMA_4B.i1-Q5_K_M.gguf) | i1-Q5_K_M | 3.3 | |
| [GGUF](https://huggingface.co/mradermacher/Impish_LLAMA_4B-i1-GGUF/resolve/main/Impish_LLAMA_4B.i1-Q6_K.gguf) | i1-Q6_K | 3.8 | practically like static Q6_K |
Here is a handy graph by ikawrakow comparing some lower-quality quant
types (lower is better):
And here are Artefact2's thoughts on the matter:
https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9
## FAQ / Model Request
See https://huggingface.co/mradermacher/model_requests for some answers to
questions you might have and/or if you want some other model quantized.
## Thanks
I thank my company, [nethype GmbH](https://www.nethype.de/), for letting
me use its servers and providing upgrades to my workstation to enable
this work in my free time. Additional thanks to [@nicoboss](https://huggingface.co/nicoboss) for giving me access to his private supercomputer, enabling me to provide many more imatrix quants, at much higher quality, than I would otherwise be able to.
<!-- end -->
|
kojeklollipop/blockassist-bc-spotted_amphibious_stork_1756431548
|
kojeklollipop
| 2025-08-29T02:07:07Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"spotted amphibious stork",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T02:07:04Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- spotted amphibious stork
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
motza0025/blockassist-bc-sturdy_leaping_jaguar_1756431383
|
motza0025
| 2025-08-29T02:01:04Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"sturdy leaping jaguar",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T02:00:48Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- sturdy leaping jaguar
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
dilip025/RoBERTa-mini
|
dilip025
| 2025-08-29T01:58:33Z | 6 | 0 | null |
[
"pytorch",
"safetensors",
"roberta-sentiment",
"text-classification",
"sentiment-classification",
"BERT",
"Roberta",
"mini-roberta",
"custom_code",
"en",
"dataset:yassiracharki/Amazon_Reviews_for_Sentiment_Analysis_fine_grained_5_classes",
"license:mit",
"region:us"
] |
text-classification
| 2025-06-02T18:38:06Z |
---
license: mit
datasets:
- yassiracharki/Amazon_Reviews_for_Sentiment_Analysis_fine_grained_5_classes
language:
- en
metrics:
- accuracy
tags:
- text-classification
- sentiment-classification
- BERT
- Roberta
- mini-roberta
---
# RoBERTa-mini: Sentiment Classifier
**Model Name**: `dilip025/RoBERTa-mini`
**Task**: Sentiment Classification
**Labels**: Very Negative, Negative, Neutral, Positive, Very Positive
A compact RoBERTa like model trained from scratch for sentiment classification.
## Example Usage
```python
from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch
tokenizer = AutoTokenizer.from_pretrained("dilip025/RoBERTa-mini")
model = AutoModelForSequenceClassification.from_pretrained("dilip025/RoBERTa-mini", trust_remote_code=True)
id2label = {
0: "Very Negative",
1: "Negative",
2: "Neutral",
3: "Positive",
4: "Very Positive"
}
def predict_sentiment(text):
inputs = tokenizer(text, return_tensors="pt", truncation=True, padding=True, max_length=128)
with torch.no_grad():
outputs = model(**inputs)
probs = torch.softmax(outputs["logits"], dim=1)
pred_class = torch.argmax(probs, dim=1).item()
return {
"text": text,
"class_id": pred_class,
"label": id2label[pred_class],
"probabilities": probs.tolist()[0]
}
# Example
result = predict_sentiment("I absolutely hate this product.")
print(result)
```
## Model Card
- **Architecture**: RoBERTa (custom small version)
- **Training Dataset**: [Amazon Reviews Dataset](https://huggingface.co/datasets/yassiracharki/Amazon_Reviews_for_Sentiment_Analysis_fine_grained_5_classes/viewer)
- **Use Case**: Sentiment classification for customer feedback, reviews, etc.
- **Input Format**: Plain text (string)
- **Output**: Dictionary with class ID, label, and class probabilities
## License
This model is licensed under the MIT License. You are free to use, modify, and distribute it with attribution.
## Author
Developed and Trained by [Dilip Pokhrel](https://huggingface.co/dilip025)
|
Jgdshkovi/csm-1b-lora-ft
|
Jgdshkovi
| 2025-08-29T01:54:22Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"text-generation-inference",
"unsloth",
"csm",
"trl",
"en",
"base_model:unsloth/csm-1b",
"base_model:finetune:unsloth/csm-1b",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] | null | 2025-08-29T01:54:02Z |
---
base_model: unsloth/csm-1b
tags:
- text-generation-inference
- transformers
- unsloth
- csm
- trl
license: apache-2.0
language:
- en
---
# Uploaded model
- **Developed by:** Jgdshkovi
- **License:** apache-2.0
- **Finetuned from model :** unsloth/csm-1b
This csm model was trained 2x faster with [Unsloth](https://github.com/unslothai/unsloth) and Huggingface's TRL library.
[<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png" width="200"/>](https://github.com/unslothai/unsloth)
|
PurplelinkPL/FinBERT_Test
|
PurplelinkPL
| 2025-08-29T01:49:01Z | 58 | 0 | null |
[
"safetensors",
"modernbert",
"finance",
"text-classification",
"en",
"dataset:HuggingFaceFW/fineweb",
"license:mit",
"region:us"
] |
text-classification
| 2025-08-01T21:23:54Z |
---
license: mit
datasets:
- HuggingFaceFW/fineweb
language:
- en
tags:
- finance
metrics:
- f1
pipeline_tag: text-classification
---
|
Vardis/greek-mt5-medical
|
Vardis
| 2025-08-29T01:45:24Z | 13 | 0 | null |
[
"safetensors",
"mt5",
"region:us"
] | null | 2025-05-01T16:02:00Z |
# Greek mT5 Text Correction Model
## Model Details
- **Model type:** Transformer-based encoder-decoder (mT5-base)
- **Language:** Greek (el)
- **License:** Same as [google/mt5-base](https://huggingface.co/google/mt5-base)
- **Task:** Grammatical and orthographic text correction
## Model Description
This model is a fine-tuned version of the multilingual T5 (mT5) model, specifically adapted for correcting noisy Greek text generated by ASR systems. The fine-tuning framed the task as a supervised sequence-to-sequence problem: the model takes a corrupted sentence as input and produces the corrected sentence as output.
To guide the model, an **instruction-style prompting approach** was employed. Input sentences were formatted as:
correct: "corrupted sentence"
Example:
| Input | Model Output |
|-------|--------------|
| correct:Ποσο θα διαρκέσει η αποκατάσταση μιτα από το χιρουργειο. | Πόσο θα διαρκέσει η αποκατάσταση μετά από το χειρουργείο; |
The model captures long-range dependencies, allowing it to correct misspellings, punctuation errors, grammatical mistakes, and occasional missing or repeated words.
## Training Data
- **Synthetic error-augmented dataset:** 56,000 corrupted–clean sentence pairs
- **Data sources:** Medical corpus and Wikipedia
- **Error types introduced:** Vowel swaps, duplicated letters, punctuation mistakes, grammatical issues
- **Input-output format:** Sequence-to-sequence (corrupted → correct)
Example corrupted–clean pairs:
| Corrupted Sentence | Clean Sentence |
|------------------|----------------|
| Ιδανηκο για προληψοη αλλαα και θεραπεια των κατακλισεων. | Ιδανικό για πρόληψη αλλα και θεραπεία των κατακλίσεων. |
| H πιεση του αερα ρυθμιζετα ανυλιογα με το βαρος του ασθενη. | H πίεση του αέρα ρυθμίζεται ανάλογα με το βάρος του ασθενή. |
## How to Use
```python
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("Vardis/greek-mt5-medical")
model = AutoModelForSeq2SeqLM.from_pretrained("Vardis/greek-mt5-medical")
input_text = "correct:Ποσο θα διαρκέσει η αποκατάσταση μιτα από το χιρουργειο."
inputs = tokenizer(input_text, return_tensors="pt")
outputs = model.generate(**inputs)
corrected_text = tokenizer.decode(outputs[0], skip_special_tokens=True)
print(corrected_text)
|
luckycanucky/impish-llama-2
|
luckycanucky
| 2025-08-29T01:39:55Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"gguf",
"llama",
"unsloth",
"arxiv:1910.09700",
"text-generation-inference",
"endpoints_compatible",
"region:us",
"conversational"
] | null | 2025-08-29T01:04:05Z |
---
library_name: transformers
tags:
- unsloth
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
|
hobson123/blockassist-bc-mammalian_dense_gibbon_1756431396
|
hobson123
| 2025-08-29T01:37:50Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"mammalian dense gibbon",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T01:37:35Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- mammalian dense gibbon
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
apple/mobileclip2_coca_dfn2b_s13b_recap-coco-30k_s12m_context128
|
apple
| 2025-08-29T01:37:30Z | 0 | 0 |
mobileclip
|
[
"mobileclip",
"arxiv:2508.20691",
"arxiv:2103.00020",
"arxiv:2303.15343",
"arxiv:2309.17425",
"license:apple-amlr",
"region:us"
] | null | 2025-08-25T19:11:24Z |
---
license: apple-amlr
license_name: apple-ascl
license_link: https://github.com/apple/ml-mobileclip/blob/main/LICENSE_weights_data
library_name: mobileclip
---
# MobileCLIP2: Improving Multi-Modal Reinforced Training
MobileCLIP2 was introduced in [MobileCLIP2: Improving Multi-Modal Reinforced Training](http://arxiv.org/abs/2508.20691) (TMLR August 2025 <mark>Featured</mark>), by Fartash Faghri, Pavan Kumar Anasosalu Vasu, Cem Koc, Vaishaal Shankar, Alexander T Toshev, Oncel Tuzel, Hadi Pouransari.
This repository contains the **CoCa** checkpoint pretrained on DFN-2B dataset and fine-tuned on varying datasets.
### Highlights
* `MobileCLIP2-S4` matches the accuracy of SigLIP-SO400M/14 with 2x fewer parameters and surpasses DFN ViT-L/14 at 2.5x lower latency measured on iPhone12 Pro Max.
* `MobileCLIP-S3/S4` are our new architectures trained on MobileCLIP’s training dataset, DataCompDR-1B (dashed lines).
* Our smallest variant `MobileCLIP-S0` obtains similar zero-shot performance as [OpenAI](https://arxiv.org/abs/2103.00020)'s ViT-B/16 model while being 4.8x faster and 2.8x smaller.
* `MobileCLIP-S2` obtains better avg zero-shot performance than [SigLIP](https://arxiv.org/abs/2303.15343)'s ViT-B/16 model while being 2.3x faster and 2.1x smaller, and trained with 3x less seen samples.
* `MobileCLIP-B (LT)` attains zero-shot ImageNet performance of **77.2%** which is significantly better than recent works like [DFN](https://arxiv.org/abs/2309.17425) and [SigLIP](https://arxiv.org/abs/2303.15343) with similar architectures or even [OpenAI's ViT-L/14@336](https://arxiv.org/abs/2103.00020).
## Checkpoints
| Model | # Seen <BR>Samples (B) | # Params (M) <BR> (img + txt) | Latency (ms) <BR> (img + txt) | IN-1k Zero-Shot <BR> Top-1 Acc. (%) | Avg. Perf. (%) <BR> on 38 datasets |
|:----------------------------------------------------------|:----------------------:|:-----------------------------:|:-----------------------------:|:-----------------------------------:|:----------------------------------:|
| [MobileCLIP2-S0](https://hf.co/apple/MobileCLIP2-S0) | 13 | 11.4 + 42.4 | 1.5 + 1.6 | 71.5 | 59.7 |
| [MobileCLIP2-S2](https://hf.co/apple/MobileCLIP2-S2) | 13 | 35.7 + 63.4 | 3.6 + 3.3 | 77.2 | 64.1 |
| [MobileCLIP2-B](https://hf.co/apple/MobileCLIP2-B) | 13 | 86.3 + 63.4 | 10.4 + 3.3 | 79.4 | 65.8 |
| [MobileCLIP2-S3](https://hf.co/apple/MobileCLIP2-S3) | 13 | 125.1 + 123.6 | 8.0 + 6.6 | 80.7 | 66.8 |
| [MobileCLIP2-L/14](https://hf.co/apple/MobileCLIP2-L-14) | 13 | 304.3 + 123.6 | 57.9 + 6.6 | 81.9 | 67.8 |
| [MobileCLIP2-S4](https://hf.co/apple/MobileCLIP2-S4) | 13 | 321.6 + 123.6 | 19.6 + 6.6 | 81.9 | 67.5 |
| [MobileCLIP-S0](https://hf.co/apple/MobileCLIP-S0) | 13 | 11.4 + 42.4 | 1.5 + 1.6 | 67.8 | 58.1 |
| [MobileCLIP-S1](https://hf.co/apple/MobileCLIP-S1) | 13 | 21.5 + 63.4 | 2.5 + 3.3 | 72.6 | 61.3 |
| [MobileCLIP-S2](https://hf.co/apple/MobileCLIP-S2) | 13 | 35.7 + 63.4 | 3.6 + 3.3 | 74.4 | 63.7 |
| [MobileCLIP-B](https://hf.co/apple/MobileCLIP-B) | 13 | 86.3 + 63.4 | 10.4 + 3.3 | 76.8 | 65.2 |
| [MobileCLIP-B (LT)](https://hf.co/apple/MobileCLIP-B-LT) | 36 | 86.3 + 63.4 | 10.4 + 3.3 | 77.2 | 65.8 |
| [MobileCLIP-S3](https://hf.co/apple/MobileCLIP-S3) | 13 | 125.1 + 123.6 | 8.0 + 6.6 | 78.3 | 66.3 |
| [MobileCLIP-L/14](https://hf.co/apple/MobileCLIP-L-14) | 13 | 304.3 + 123.6 | 57.9 + 6.6 | 79.5 | 66.9 |
| [MobileCLIP-S4](https://hf.co/apple/MobileCLIP-S4) | 13 | 321.6 + 123.6 | 19.6 + 6.6 | 79.4 | 68.1 |
## How to Use
First, download the desired checkpoint visiting one of the links in the table above, then click the `Files and versions` tab, and download the PyTorch checkpoint.
For programmatic downloading, if you have `huggingface_hub` installed, you can also run:
```
hf download apple/mobileclip2_coca_dfn2b_s13b_<finetune-dataset>_context<length>
```
For models length with context lengths 128/256, copy `config.json` to `src/open_clip/model_configs/coca_ViT-L-14-context$len.json` and change the model name in below example to `coca_ViT-L-14-context$len`.
```py
import torch
import open_clip
from PIL import Image
model, _, preprocess = open_clip.create_model_and_transforms('coca_ViT-L-14', pretrained='/path/to/mobileclip2_coca.pt')
model.eval()
image = preprocess(Image.open("docs/fig_accuracy_latency.png").convert('RGB')).unsqueeze(0)
with torch.no_grad(), torch.cuda.amp.autocast():
syn_text = model.generate(
image,
generation_type="top_p",
top_p=0.9,
fixed_output_length=True
)[0]
syn_text = open_clip.decode(syn_text).split("<end_of_text>")[0].split("<start_of_text>")[-1].split(".")[0].rstrip()
print("Caption:", syn_text)
```
|
apple/mobileclip2_coca_dfn2b_s13b_gbc1m-long_context256
|
apple
| 2025-08-29T01:37:25Z | 0 | 0 |
mobileclip
|
[
"mobileclip",
"arxiv:2508.20691",
"arxiv:2103.00020",
"arxiv:2303.15343",
"arxiv:2309.17425",
"license:apple-amlr",
"region:us"
] | null | 2025-08-25T19:07:57Z |
---
license: apple-amlr
license_name: apple-ascl
license_link: https://github.com/apple/ml-mobileclip/blob/main/LICENSE_weights_data
library_name: mobileclip
---
# MobileCLIP2: Improving Multi-Modal Reinforced Training
MobileCLIP2 was introduced in [MobileCLIP2: Improving Multi-Modal Reinforced Training](http://arxiv.org/abs/2508.20691) (TMLR August 2025 <mark>Featured</mark>), by Fartash Faghri, Pavan Kumar Anasosalu Vasu, Cem Koc, Vaishaal Shankar, Alexander T Toshev, Oncel Tuzel, Hadi Pouransari.
This repository contains the **CoCa** checkpoint pretrained on DFN-2B dataset and fine-tuned on varying datasets.
### Highlights
* `MobileCLIP2-S4` matches the accuracy of SigLIP-SO400M/14 with 2x fewer parameters and surpasses DFN ViT-L/14 at 2.5x lower latency measured on iPhone12 Pro Max.
* `MobileCLIP-S3/S4` are our new architectures trained on MobileCLIP’s training dataset, DataCompDR-1B (dashed lines).
* Our smallest variant `MobileCLIP-S0` obtains similar zero-shot performance as [OpenAI](https://arxiv.org/abs/2103.00020)'s ViT-B/16 model while being 4.8x faster and 2.8x smaller.
* `MobileCLIP-S2` obtains better avg zero-shot performance than [SigLIP](https://arxiv.org/abs/2303.15343)'s ViT-B/16 model while being 2.3x faster and 2.1x smaller, and trained with 3x less seen samples.
* `MobileCLIP-B (LT)` attains zero-shot ImageNet performance of **77.2%** which is significantly better than recent works like [DFN](https://arxiv.org/abs/2309.17425) and [SigLIP](https://arxiv.org/abs/2303.15343) with similar architectures or even [OpenAI's ViT-L/14@336](https://arxiv.org/abs/2103.00020).
## Checkpoints
| Model | # Seen <BR>Samples (B) | # Params (M) <BR> (img + txt) | Latency (ms) <BR> (img + txt) | IN-1k Zero-Shot <BR> Top-1 Acc. (%) | Avg. Perf. (%) <BR> on 38 datasets |
|:----------------------------------------------------------|:----------------------:|:-----------------------------:|:-----------------------------:|:-----------------------------------:|:----------------------------------:|
| [MobileCLIP2-S0](https://hf.co/apple/MobileCLIP2-S0) | 13 | 11.4 + 42.4 | 1.5 + 1.6 | 71.5 | 59.7 |
| [MobileCLIP2-S2](https://hf.co/apple/MobileCLIP2-S2) | 13 | 35.7 + 63.4 | 3.6 + 3.3 | 77.2 | 64.1 |
| [MobileCLIP2-B](https://hf.co/apple/MobileCLIP2-B) | 13 | 86.3 + 63.4 | 10.4 + 3.3 | 79.4 | 65.8 |
| [MobileCLIP2-S3](https://hf.co/apple/MobileCLIP2-S3) | 13 | 125.1 + 123.6 | 8.0 + 6.6 | 80.7 | 66.8 |
| [MobileCLIP2-L/14](https://hf.co/apple/MobileCLIP2-L-14) | 13 | 304.3 + 123.6 | 57.9 + 6.6 | 81.9 | 67.8 |
| [MobileCLIP2-S4](https://hf.co/apple/MobileCLIP2-S4) | 13 | 321.6 + 123.6 | 19.6 + 6.6 | 81.9 | 67.5 |
| [MobileCLIP-S0](https://hf.co/apple/MobileCLIP-S0) | 13 | 11.4 + 42.4 | 1.5 + 1.6 | 67.8 | 58.1 |
| [MobileCLIP-S1](https://hf.co/apple/MobileCLIP-S1) | 13 | 21.5 + 63.4 | 2.5 + 3.3 | 72.6 | 61.3 |
| [MobileCLIP-S2](https://hf.co/apple/MobileCLIP-S2) | 13 | 35.7 + 63.4 | 3.6 + 3.3 | 74.4 | 63.7 |
| [MobileCLIP-B](https://hf.co/apple/MobileCLIP-B) | 13 | 86.3 + 63.4 | 10.4 + 3.3 | 76.8 | 65.2 |
| [MobileCLIP-B (LT)](https://hf.co/apple/MobileCLIP-B-LT) | 36 | 86.3 + 63.4 | 10.4 + 3.3 | 77.2 | 65.8 |
| [MobileCLIP-S3](https://hf.co/apple/MobileCLIP-S3) | 13 | 125.1 + 123.6 | 8.0 + 6.6 | 78.3 | 66.3 |
| [MobileCLIP-L/14](https://hf.co/apple/MobileCLIP-L-14) | 13 | 304.3 + 123.6 | 57.9 + 6.6 | 79.5 | 66.9 |
| [MobileCLIP-S4](https://hf.co/apple/MobileCLIP-S4) | 13 | 321.6 + 123.6 | 19.6 + 6.6 | 79.4 | 68.1 |
## How to Use
First, download the desired checkpoint visiting one of the links in the table above, then click the `Files and versions` tab, and download the PyTorch checkpoint.
For programmatic downloading, if you have `huggingface_hub` installed, you can also run:
```
hf download apple/mobileclip2_coca_dfn2b_s13b_<finetune-dataset>_context<length>
```
For models length with context lengths 128/256, copy `config.json` to `src/open_clip/model_configs/coca_ViT-L-14-context$len.json` and change the model name in below example to `coca_ViT-L-14-context$len`.
```py
import torch
import open_clip
from PIL import Image
model, _, preprocess = open_clip.create_model_and_transforms('coca_ViT-L-14', pretrained='/path/to/mobileclip2_coca.pt')
model.eval()
image = preprocess(Image.open("docs/fig_accuracy_latency.png").convert('RGB')).unsqueeze(0)
with torch.no_grad(), torch.cuda.amp.autocast():
syn_text = model.generate(
image,
generation_type="top_p",
top_p=0.9,
fixed_output_length=True
)[0]
syn_text = open_clip.decode(syn_text).split("<end_of_text>")[0].split("<start_of_text>")[-1].split(".")[0].rstrip()
print("Caption:", syn_text)
```
|
apple/mobileclip2_coca_dfn2b_s13b_dci-extended_s12m_context128
|
apple
| 2025-08-29T01:37:09Z | 0 | 0 |
mobileclip
|
[
"mobileclip",
"arxiv:2508.20691",
"arxiv:2103.00020",
"arxiv:2303.15343",
"arxiv:2309.17425",
"license:apple-amlr",
"region:us"
] | null | 2025-08-25T18:55:18Z |
---
license: apple-amlr
license_name: apple-ascl
license_link: https://github.com/apple/ml-mobileclip/blob/main/LICENSE_weights_data
library_name: mobileclip
---
# MobileCLIP2: Improving Multi-Modal Reinforced Training
MobileCLIP2 was introduced in [MobileCLIP2: Improving Multi-Modal Reinforced Training](http://arxiv.org/abs/2508.20691) (TMLR August 2025 <mark>Featured</mark>), by Fartash Faghri, Pavan Kumar Anasosalu Vasu, Cem Koc, Vaishaal Shankar, Alexander T Toshev, Oncel Tuzel, Hadi Pouransari.
This repository contains the **CoCa** checkpoint pretrained on DFN-2B dataset and fine-tuned on varying datasets.
### Highlights
* `MobileCLIP2-S4` matches the accuracy of SigLIP-SO400M/14 with 2x fewer parameters and surpasses DFN ViT-L/14 at 2.5x lower latency measured on iPhone12 Pro Max.
* `MobileCLIP-S3/S4` are our new architectures trained on MobileCLIP’s training dataset, DataCompDR-1B (dashed lines).
* Our smallest variant `MobileCLIP-S0` obtains similar zero-shot performance as [OpenAI](https://arxiv.org/abs/2103.00020)'s ViT-B/16 model while being 4.8x faster and 2.8x smaller.
* `MobileCLIP-S2` obtains better avg zero-shot performance than [SigLIP](https://arxiv.org/abs/2303.15343)'s ViT-B/16 model while being 2.3x faster and 2.1x smaller, and trained with 3x less seen samples.
* `MobileCLIP-B (LT)` attains zero-shot ImageNet performance of **77.2%** which is significantly better than recent works like [DFN](https://arxiv.org/abs/2309.17425) and [SigLIP](https://arxiv.org/abs/2303.15343) with similar architectures or even [OpenAI's ViT-L/14@336](https://arxiv.org/abs/2103.00020).
## Checkpoints
| Model | # Seen <BR>Samples (B) | # Params (M) <BR> (img + txt) | Latency (ms) <BR> (img + txt) | IN-1k Zero-Shot <BR> Top-1 Acc. (%) | Avg. Perf. (%) <BR> on 38 datasets |
|:----------------------------------------------------------|:----------------------:|:-----------------------------:|:-----------------------------:|:-----------------------------------:|:----------------------------------:|
| [MobileCLIP2-S0](https://hf.co/apple/MobileCLIP2-S0) | 13 | 11.4 + 42.4 | 1.5 + 1.6 | 71.5 | 59.7 |
| [MobileCLIP2-S2](https://hf.co/apple/MobileCLIP2-S2) | 13 | 35.7 + 63.4 | 3.6 + 3.3 | 77.2 | 64.1 |
| [MobileCLIP2-B](https://hf.co/apple/MobileCLIP2-B) | 13 | 86.3 + 63.4 | 10.4 + 3.3 | 79.4 | 65.8 |
| [MobileCLIP2-S3](https://hf.co/apple/MobileCLIP2-S3) | 13 | 125.1 + 123.6 | 8.0 + 6.6 | 80.7 | 66.8 |
| [MobileCLIP2-L/14](https://hf.co/apple/MobileCLIP2-L-14) | 13 | 304.3 + 123.6 | 57.9 + 6.6 | 81.9 | 67.8 |
| [MobileCLIP2-S4](https://hf.co/apple/MobileCLIP2-S4) | 13 | 321.6 + 123.6 | 19.6 + 6.6 | 81.9 | 67.5 |
| [MobileCLIP-S0](https://hf.co/apple/MobileCLIP-S0) | 13 | 11.4 + 42.4 | 1.5 + 1.6 | 67.8 | 58.1 |
| [MobileCLIP-S1](https://hf.co/apple/MobileCLIP-S1) | 13 | 21.5 + 63.4 | 2.5 + 3.3 | 72.6 | 61.3 |
| [MobileCLIP-S2](https://hf.co/apple/MobileCLIP-S2) | 13 | 35.7 + 63.4 | 3.6 + 3.3 | 74.4 | 63.7 |
| [MobileCLIP-B](https://hf.co/apple/MobileCLIP-B) | 13 | 86.3 + 63.4 | 10.4 + 3.3 | 76.8 | 65.2 |
| [MobileCLIP-B (LT)](https://hf.co/apple/MobileCLIP-B-LT) | 36 | 86.3 + 63.4 | 10.4 + 3.3 | 77.2 | 65.8 |
| [MobileCLIP-S3](https://hf.co/apple/MobileCLIP-S3) | 13 | 125.1 + 123.6 | 8.0 + 6.6 | 78.3 | 66.3 |
| [MobileCLIP-L/14](https://hf.co/apple/MobileCLIP-L-14) | 13 | 304.3 + 123.6 | 57.9 + 6.6 | 79.5 | 66.9 |
| [MobileCLIP-S4](https://hf.co/apple/MobileCLIP-S4) | 13 | 321.6 + 123.6 | 19.6 + 6.6 | 79.4 | 68.1 |
## How to Use
First, download the desired checkpoint visiting one of the links in the table above, then click the `Files and versions` tab, and download the PyTorch checkpoint.
For programmatic downloading, if you have `huggingface_hub` installed, you can also run:
```
hf download apple/mobileclip2_coca_dfn2b_s13b_<finetune-dataset>_context<length>
```
For models length with context lengths 128/256, copy `config.json` to `src/open_clip/model_configs/coca_ViT-L-14-context$len.json` and change the model name in below example to `coca_ViT-L-14-context$len`.
```py
import torch
import open_clip
from PIL import Image
model, _, preprocess = open_clip.create_model_and_transforms('coca_ViT-L-14', pretrained='/path/to/mobileclip2_coca.pt')
model.eval()
image = preprocess(Image.open("docs/fig_accuracy_latency.png").convert('RGB')).unsqueeze(0)
with torch.no_grad(), torch.cuda.amp.autocast():
syn_text = model.generate(
image,
generation_type="top_p",
top_p=0.9,
fixed_output_length=True
)[0]
syn_text = open_clip.decode(syn_text).split("<end_of_text>")[0].split("<start_of_text>")[-1].split(".")[0].rstrip()
print("Caption:", syn_text)
```
|
apple/mobileclip2_coca_dfn2b_s13b_dci-complete_s12m_context128
|
apple
| 2025-08-29T01:37:04Z | 0 | 0 |
mobileclip
|
[
"mobileclip",
"arxiv:2508.20691",
"arxiv:2103.00020",
"arxiv:2303.15343",
"arxiv:2309.17425",
"license:apple-amlr",
"region:us"
] | null | 2025-08-25T18:52:05Z |
---
license: apple-amlr
license_name: apple-ascl
license_link: https://github.com/apple/ml-mobileclip/blob/main/LICENSE_weights_data
library_name: mobileclip
---
# MobileCLIP2: Improving Multi-Modal Reinforced Training
MobileCLIP2 was introduced in [MobileCLIP2: Improving Multi-Modal Reinforced Training](http://arxiv.org/abs/2508.20691) (TMLR August 2025 <mark>Featured</mark>), by Fartash Faghri, Pavan Kumar Anasosalu Vasu, Cem Koc, Vaishaal Shankar, Alexander T Toshev, Oncel Tuzel, Hadi Pouransari.
This repository contains the **CoCa** checkpoint pretrained on DFN-2B dataset and fine-tuned on varying datasets.
### Highlights
* `MobileCLIP2-S4` matches the accuracy of SigLIP-SO400M/14 with 2x fewer parameters and surpasses DFN ViT-L/14 at 2.5x lower latency measured on iPhone12 Pro Max.
* `MobileCLIP-S3/S4` are our new architectures trained on MobileCLIP’s training dataset, DataCompDR-1B (dashed lines).
* Our smallest variant `MobileCLIP-S0` obtains similar zero-shot performance as [OpenAI](https://arxiv.org/abs/2103.00020)'s ViT-B/16 model while being 4.8x faster and 2.8x smaller.
* `MobileCLIP-S2` obtains better avg zero-shot performance than [SigLIP](https://arxiv.org/abs/2303.15343)'s ViT-B/16 model while being 2.3x faster and 2.1x smaller, and trained with 3x less seen samples.
* `MobileCLIP-B (LT)` attains zero-shot ImageNet performance of **77.2%** which is significantly better than recent works like [DFN](https://arxiv.org/abs/2309.17425) and [SigLIP](https://arxiv.org/abs/2303.15343) with similar architectures or even [OpenAI's ViT-L/14@336](https://arxiv.org/abs/2103.00020).
## Checkpoints
| Model | # Seen <BR>Samples (B) | # Params (M) <BR> (img + txt) | Latency (ms) <BR> (img + txt) | IN-1k Zero-Shot <BR> Top-1 Acc. (%) | Avg. Perf. (%) <BR> on 38 datasets |
|:----------------------------------------------------------|:----------------------:|:-----------------------------:|:-----------------------------:|:-----------------------------------:|:----------------------------------:|
| [MobileCLIP2-S0](https://hf.co/apple/MobileCLIP2-S0) | 13 | 11.4 + 42.4 | 1.5 + 1.6 | 71.5 | 59.7 |
| [MobileCLIP2-S2](https://hf.co/apple/MobileCLIP2-S2) | 13 | 35.7 + 63.4 | 3.6 + 3.3 | 77.2 | 64.1 |
| [MobileCLIP2-B](https://hf.co/apple/MobileCLIP2-B) | 13 | 86.3 + 63.4 | 10.4 + 3.3 | 79.4 | 65.8 |
| [MobileCLIP2-S3](https://hf.co/apple/MobileCLIP2-S3) | 13 | 125.1 + 123.6 | 8.0 + 6.6 | 80.7 | 66.8 |
| [MobileCLIP2-L/14](https://hf.co/apple/MobileCLIP2-L-14) | 13 | 304.3 + 123.6 | 57.9 + 6.6 | 81.9 | 67.8 |
| [MobileCLIP2-S4](https://hf.co/apple/MobileCLIP2-S4) | 13 | 321.6 + 123.6 | 19.6 + 6.6 | 81.9 | 67.5 |
| [MobileCLIP-S0](https://hf.co/apple/MobileCLIP-S0) | 13 | 11.4 + 42.4 | 1.5 + 1.6 | 67.8 | 58.1 |
| [MobileCLIP-S1](https://hf.co/apple/MobileCLIP-S1) | 13 | 21.5 + 63.4 | 2.5 + 3.3 | 72.6 | 61.3 |
| [MobileCLIP-S2](https://hf.co/apple/MobileCLIP-S2) | 13 | 35.7 + 63.4 | 3.6 + 3.3 | 74.4 | 63.7 |
| [MobileCLIP-B](https://hf.co/apple/MobileCLIP-B) | 13 | 86.3 + 63.4 | 10.4 + 3.3 | 76.8 | 65.2 |
| [MobileCLIP-B (LT)](https://hf.co/apple/MobileCLIP-B-LT) | 36 | 86.3 + 63.4 | 10.4 + 3.3 | 77.2 | 65.8 |
| [MobileCLIP-S3](https://hf.co/apple/MobileCLIP-S3) | 13 | 125.1 + 123.6 | 8.0 + 6.6 | 78.3 | 66.3 |
| [MobileCLIP-L/14](https://hf.co/apple/MobileCLIP-L-14) | 13 | 304.3 + 123.6 | 57.9 + 6.6 | 79.5 | 66.9 |
| [MobileCLIP-S4](https://hf.co/apple/MobileCLIP-S4) | 13 | 321.6 + 123.6 | 19.6 + 6.6 | 79.4 | 68.1 |
## How to Use
First, download the desired checkpoint visiting one of the links in the table above, then click the `Files and versions` tab, and download the PyTorch checkpoint.
For programmatic downloading, if you have `huggingface_hub` installed, you can also run:
```
hf download apple/mobileclip2_coca_dfn2b_s13b_<finetune-dataset>_context<length>
```
For models length with context lengths 128/256, copy `config.json` to `src/open_clip/model_configs/coca_ViT-L-14-context$len.json` and change the model name in below example to `coca_ViT-L-14-context$len`.
```py
import torch
import open_clip
from PIL import Image
model, _, preprocess = open_clip.create_model_and_transforms('coca_ViT-L-14', pretrained='/path/to/mobileclip2_coca.pt')
model.eval()
image = preprocess(Image.open("docs/fig_accuracy_latency.png").convert('RGB')).unsqueeze(0)
with torch.no_grad(), torch.cuda.amp.autocast():
syn_text = model.generate(
image,
generation_type="top_p",
top_p=0.9,
fixed_output_length=True
)[0]
syn_text = open_clip.decode(syn_text).split("<end_of_text>")[0].split("<start_of_text>")[-1].split(".")[0].rstrip()
print("Caption:", syn_text)
```
|
Maori999/Qwen3-0.6B-Gensyn-Swarm-tiny_shrewd_alligator
|
Maori999
| 2025-08-29T01:31:55Z | 77 | 0 |
transformers
|
[
"transformers",
"safetensors",
"qwen3",
"text-generation",
"rl-swarm",
"genrl-swarm",
"grpo",
"gensyn",
"I am tiny_shrewd_alligator",
"arxiv:1910.09700",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-06-27T12:29:26Z |
---
library_name: transformers
tags:
- rl-swarm
- genrl-swarm
- grpo
- gensyn
- I am tiny_shrewd_alligator
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
|
badfriend221/Qwen2.5-0.5B-Instruct-Gensyn-Swarm-rabid_amphibious_donkey
|
badfriend221
| 2025-08-29T01:31:31Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"qwen2",
"text-generation",
"rl-swarm",
"genrl-swarm",
"grpo",
"gensyn",
"I am rabid_amphibious_donkey",
"arxiv:1910.09700",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-08-28T17:07:16Z |
---
library_name: transformers
tags:
- rl-swarm
- genrl-swarm
- grpo
- gensyn
- I am rabid_amphibious_donkey
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
|
apple/MobileCLIP2-S2
|
apple
| 2025-08-29T01:30:20Z | 0 | 0 |
mobileclip
|
[
"mobileclip",
"arxiv:2508.20691",
"arxiv:2103.00020",
"arxiv:2303.15343",
"arxiv:2309.17425",
"license:apple-amlr",
"region:us"
] | null | 2025-08-25T16:51:25Z |
---
license: apple-amlr
license_name: apple-ascl
license_link: https://github.com/apple/ml-mobileclip/blob/main/LICENSE_weights_data
library_name: mobileclip
---
# MobileCLIP2: Improving Multi-Modal Reinforced Training
MobileCLIP2 was introduced in [MobileCLIP2: Improving Multi-Modal Reinforced Training](http://arxiv.org/abs/2508.20691) (TMLR August 2025 <mark>Featured</mark>), by Fartash Faghri, Pavan Kumar Anasosalu Vasu, Cem Koc, Vaishaal Shankar, Alexander T Toshev, Oncel Tuzel, Hadi Pouransari.
This repository contains the **MobileCLIP2-S2** checkpoint.
### Highlights
* `MobileCLIP2-S4` matches the accuracy of SigLIP-SO400M/14 with 2x fewer parameters and surpasses DFN ViT-L/14 at 2.5x lower latency measured on iPhone12 Pro Max.
* `MobileCLIP-S3/S4` are our new architectures trained on MobileCLIP’s training dataset, DataCompDR-1B (dashed lines).
* Our smallest variant `MobileCLIP-S0` obtains similar zero-shot performance as [OpenAI](https://arxiv.org/abs/2103.00020)'s ViT-B/16 model while being 4.8x faster and 2.8x smaller.
* `MobileCLIP-S2` obtains better avg zero-shot performance than [SigLIP](https://arxiv.org/abs/2303.15343)'s ViT-B/16 model while being 2.3x faster and 2.1x smaller, and trained with 3x less seen samples.
* `MobileCLIP-B (LT)` attains zero-shot ImageNet performance of **77.2%** which is significantly better than recent works like [DFN](https://arxiv.org/abs/2309.17425) and [SigLIP](https://arxiv.org/abs/2303.15343) with similar architectures or even [OpenAI's ViT-L/14@336](https://arxiv.org/abs/2103.00020).
## Checkpoints
| Model | # Seen <BR>Samples (B) | # Params (M) <BR> (img + txt) | Latency (ms) <BR> (img + txt) | IN-1k Zero-Shot <BR> Top-1 Acc. (%) | Avg. Perf. (%) <BR> on 38 datasets |
|:----------------------------------------------------------|:----------------------:|:-----------------------------:|:-----------------------------:|:-----------------------------------:|:----------------------------------:|
| [MobileCLIP2-S0](https://hf.co/apple/MobileCLIP2-S0) | 13 | 11.4 + 42.4 | 1.5 + 1.6 | 71.5 | 59.7 |
| [MobileCLIP2-S2](https://hf.co/apple/MobileCLIP2-S2) | 13 | 35.7 + 63.4 | 3.6 + 3.3 | 77.2 | 64.1 |
| [MobileCLIP2-B](https://hf.co/apple/MobileCLIP2-B) | 13 | 86.3 + 63.4 | 10.4 + 3.3 | 79.4 | 65.8 |
| [MobileCLIP2-S3](https://hf.co/apple/MobileCLIP2-S3) | 13 | 125.1 + 123.6 | 8.0 + 6.6 | 80.7 | 66.8 |
| [MobileCLIP2-L/14](https://hf.co/apple/MobileCLIP2-L-14) | 13 | 304.3 + 123.6 | 57.9 + 6.6 | 81.9 | 67.8 |
| [MobileCLIP2-S4](https://hf.co/apple/MobileCLIP2-S4) | 13 | 321.6 + 123.6 | 19.6 + 6.6 | 81.9 | 67.5 |
| [MobileCLIP-S0](https://hf.co/apple/MobileCLIP-S0) | 13 | 11.4 + 42.4 | 1.5 + 1.6 | 67.8 | 58.1 |
| [MobileCLIP-S1](https://hf.co/apple/MobileCLIP-S1) | 13 | 21.5 + 63.4 | 2.5 + 3.3 | 72.6 | 61.3 |
| [MobileCLIP-S2](https://hf.co/apple/MobileCLIP-S2) | 13 | 35.7 + 63.4 | 3.6 + 3.3 | 74.4 | 63.7 |
| [MobileCLIP-B](https://hf.co/apple/MobileCLIP-B) | 13 | 86.3 + 63.4 | 10.4 + 3.3 | 76.8 | 65.2 |
| [MobileCLIP-B (LT)](https://hf.co/apple/MobileCLIP-B-LT) | 36 | 86.3 + 63.4 | 10.4 + 3.3 | 77.2 | 65.8 |
| [MobileCLIP-S3](https://hf.co/apple/MobileCLIP-S3) | 13 | 125.1 + 123.6 | 8.0 + 6.6 | 78.3 | 66.3 |
| [MobileCLIP-L/14](https://hf.co/apple/MobileCLIP-L-14) | 13 | 304.3 + 123.6 | 57.9 + 6.6 | 79.5 | 66.9 |
| [MobileCLIP-S4](https://hf.co/apple/MobileCLIP-S4) | 13 | 321.6 + 123.6 | 19.6 + 6.6 | 79.4 | 68.1 |
## How to Use
First, download the desired checkpoint visiting one of the links in the table above, then click the `Files and versions` tab, and download the PyTorch checkpoint.
For programmatic downloading, if you have `huggingface_hub` installed, you can also run:
```
hf download apple/MobileCLIP2-S2
```
Then, install [`ml-mobileclip`](https://github.com/apple/ml-mobileclip) by following the instructions in the repo. It uses an API similar to [`open_clip`'s](https://github.com/mlfoundations/open_clip).
You can run inference with a code snippet like the following:
```py
import torch
import open_clip
from PIL import Image
from mobileclip.modules.common.mobileone import reparameterize_model
model, _, preprocess = open_clip.create_model_and_transforms('MobileCLIP-S2', pretrained='/path/to/mobileclip2_s2.pt')
tokenizer = open_clip.get_tokenizer('MobileCLIP-S2')
# For inference/model exporting purposes, please reparameterize first
model = reparameterize_model(model.eval())
image = preprocess(Image.open("docs/fig_accuracy_latency.png").convert('RGB')).unsqueeze(0)
text = tokenizer(["a diagram", "a dog", "a cat"])
with torch.no_grad(), torch.cuda.amp.autocast():
image_features = model.encode_image(image)
text_features = model.encode_text(text)
image_features /= image_features.norm(dim=-1, keepdim=True)
text_features /= text_features.norm(dim=-1, keepdim=True)
text_probs = (100.0 * image_features @ text_features.T).softmax(dim=-1)
print("Label probs:", text_probs)
```
|
apple/MobileCLIP2-S0
|
apple
| 2025-08-29T01:30:18Z | 0 | 0 |
mobileclip
|
[
"mobileclip",
"arxiv:2508.20691",
"arxiv:2103.00020",
"arxiv:2303.15343",
"arxiv:2309.17425",
"license:apple-amlr",
"region:us"
] | null | 2025-08-25T16:51:15Z |
---
license: apple-amlr
license_name: apple-ascl
license_link: https://github.com/apple/ml-mobileclip/blob/main/LICENSE_weights_data
library_name: mobileclip
---
# MobileCLIP2: Improving Multi-Modal Reinforced Training
MobileCLIP2 was introduced in [MobileCLIP2: Improving Multi-Modal Reinforced Training](http://arxiv.org/abs/2508.20691) (TMLR August 2025 <mark>Featured</mark>), by Fartash Faghri, Pavan Kumar Anasosalu Vasu, Cem Koc, Vaishaal Shankar, Alexander T Toshev, Oncel Tuzel, Hadi Pouransari.
This repository contains the **MobileCLIP2-S0** checkpoint.
### Highlights
* `MobileCLIP2-S4` matches the accuracy of SigLIP-SO400M/14 with 2x fewer parameters and surpasses DFN ViT-L/14 at 2.5x lower latency measured on iPhone12 Pro Max.
* `MobileCLIP-S3/S4` are our new architectures trained on MobileCLIP’s training dataset, DataCompDR-1B (dashed lines).
* Our smallest variant `MobileCLIP-S0` obtains similar zero-shot performance as [OpenAI](https://arxiv.org/abs/2103.00020)'s ViT-B/16 model while being 4.8x faster and 2.8x smaller.
* `MobileCLIP-S2` obtains better avg zero-shot performance than [SigLIP](https://arxiv.org/abs/2303.15343)'s ViT-B/16 model while being 2.3x faster and 2.1x smaller, and trained with 3x less seen samples.
* `MobileCLIP-B (LT)` attains zero-shot ImageNet performance of **77.2%** which is significantly better than recent works like [DFN](https://arxiv.org/abs/2309.17425) and [SigLIP](https://arxiv.org/abs/2303.15343) with similar architectures or even [OpenAI's ViT-L/14@336](https://arxiv.org/abs/2103.00020).
## Checkpoints
| Model | # Seen <BR>Samples (B) | # Params (M) <BR> (img + txt) | Latency (ms) <BR> (img + txt) | IN-1k Zero-Shot <BR> Top-1 Acc. (%) | Avg. Perf. (%) <BR> on 38 datasets |
|:----------------------------------------------------------|:----------------------:|:-----------------------------:|:-----------------------------:|:-----------------------------------:|:----------------------------------:|
| [MobileCLIP2-S0](https://hf.co/apple/MobileCLIP2-S0) | 13 | 11.4 + 42.4 | 1.5 + 1.6 | 71.5 | 59.7 |
| [MobileCLIP2-S2](https://hf.co/apple/MobileCLIP2-S2) | 13 | 35.7 + 63.4 | 3.6 + 3.3 | 77.2 | 64.1 |
| [MobileCLIP2-B](https://hf.co/apple/MobileCLIP2-B) | 13 | 86.3 + 63.4 | 10.4 + 3.3 | 79.4 | 65.8 |
| [MobileCLIP2-S3](https://hf.co/apple/MobileCLIP2-S3) | 13 | 125.1 + 123.6 | 8.0 + 6.6 | 80.7 | 66.8 |
| [MobileCLIP2-L/14](https://hf.co/apple/MobileCLIP2-L-14) | 13 | 304.3 + 123.6 | 57.9 + 6.6 | 81.9 | 67.8 |
| [MobileCLIP2-S4](https://hf.co/apple/MobileCLIP2-S4) | 13 | 321.6 + 123.6 | 19.6 + 6.6 | 81.9 | 67.5 |
| [MobileCLIP-S0](https://hf.co/apple/MobileCLIP-S0) | 13 | 11.4 + 42.4 | 1.5 + 1.6 | 67.8 | 58.1 |
| [MobileCLIP-S1](https://hf.co/apple/MobileCLIP-S1) | 13 | 21.5 + 63.4 | 2.5 + 3.3 | 72.6 | 61.3 |
| [MobileCLIP-S2](https://hf.co/apple/MobileCLIP-S2) | 13 | 35.7 + 63.4 | 3.6 + 3.3 | 74.4 | 63.7 |
| [MobileCLIP-B](https://hf.co/apple/MobileCLIP-B) | 13 | 86.3 + 63.4 | 10.4 + 3.3 | 76.8 | 65.2 |
| [MobileCLIP-B (LT)](https://hf.co/apple/MobileCLIP-B-LT) | 36 | 86.3 + 63.4 | 10.4 + 3.3 | 77.2 | 65.8 |
| [MobileCLIP-S3](https://hf.co/apple/MobileCLIP-S3) | 13 | 125.1 + 123.6 | 8.0 + 6.6 | 78.3 | 66.3 |
| [MobileCLIP-L/14](https://hf.co/apple/MobileCLIP-L-14) | 13 | 304.3 + 123.6 | 57.9 + 6.6 | 79.5 | 66.9 |
| [MobileCLIP-S4](https://hf.co/apple/MobileCLIP-S4) | 13 | 321.6 + 123.6 | 19.6 + 6.6 | 79.4 | 68.1 |
## How to Use
First, download the desired checkpoint visiting one of the links in the table above, then click the `Files and versions` tab, and download the PyTorch checkpoint.
For programmatic downloading, if you have `huggingface_hub` installed, you can also run:
```
hf download apple/MobileCLIP2-S0
```
Then, install [`ml-mobileclip`](https://github.com/apple/ml-mobileclip) by following the instructions in the repo. It uses an API similar to [`open_clip`'s](https://github.com/mlfoundations/open_clip).
You can run inference with a code snippet like the following:
```py
import torch
import open_clip
from PIL import Image
from mobileclip.modules.common.mobileone import reparameterize_model
model, _, preprocess = open_clip.create_model_and_transforms('MobileCLIP2-S0', pretrained='/path/to/mobileclip2_s0.pt')
tokenizer = open_clip.get_tokenizer('MobileCLIP2-S0')
# For inference/model exporting purposes, please reparameterize first
model = reparameterize_model(model.eval())
image = preprocess(Image.open("docs/fig_accuracy_latency.png").convert('RGB')).unsqueeze(0)
text = tokenizer(["a diagram", "a dog", "a cat"])
with torch.no_grad(), torch.cuda.amp.autocast():
image_features = model.encode_image(image)
text_features = model.encode_text(text)
image_features /= image_features.norm(dim=-1, keepdim=True)
text_features /= text_features.norm(dim=-1, keepdim=True)
text_probs = (100.0 * image_features @ text_features.T).softmax(dim=-1)
print("Label probs:", text_probs)
```
|
haritsondavid/blockassist-bc-padded_howling_anteater_1756430745
|
haritsondavid
| 2025-08-29T01:26:50Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"padded howling anteater",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T01:26:10Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- padded howling anteater
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
apple/MobileCLIP-S4
|
apple
| 2025-08-29T01:26:19Z | 0 | 0 |
mobileclip
|
[
"mobileclip",
"arxiv:2508.20691",
"arxiv:2103.00020",
"arxiv:2303.15343",
"arxiv:2309.17425",
"license:apple-amlr",
"region:us"
] | null | 2025-08-25T16:49:54Z |
---
license: apple-amlr
license_name: apple-ascl
license_link: https://github.com/apple/ml-mobileclip/blob/main/LICENSE_weights_data
library_name: mobileclip
---
# MobileCLIP2: Improving Multi-Modal Reinforced Training
MobileCLIP2 was introduced in [MobileCLIP2: Improving Multi-Modal Reinforced Training](http://arxiv.org/abs/2508.20691) (TMLR August 2025 <mark>Featured</mark>), by Fartash Faghri, Pavan Kumar Anasosalu Vasu, Cem Koc, Vaishaal Shankar, Alexander T Toshev, Oncel Tuzel, Hadi Pouransari.
This repository contains the **MobileCLIP-S4** checkpoint.
### Highlights
* `MobileCLIP2-S4` matches the accuracy of SigLIP-SO400M/14 with 2x fewer parameters and surpasses DFN ViT-L/14 at 2.5x lower latency measured on iPhone12 Pro Max.
* `MobileCLIP-S3/S4` are our new architectures trained on MobileCLIP’s training dataset, DataCompDR-1B (dashed lines).
* Our smallest variant `MobileCLIP-S0` obtains similar zero-shot performance as [OpenAI](https://arxiv.org/abs/2103.00020)'s ViT-B/16 model while being 4.8x faster and 2.8x smaller.
* `MobileCLIP-S2` obtains better avg zero-shot performance than [SigLIP](https://arxiv.org/abs/2303.15343)'s ViT-B/16 model while being 2.3x faster and 2.1x smaller, and trained with 3x less seen samples.
* `MobileCLIP-B (LT)` attains zero-shot ImageNet performance of **77.2%** which is significantly better than recent works like [DFN](https://arxiv.org/abs/2309.17425) and [SigLIP](https://arxiv.org/abs/2303.15343) with similar architectures or even [OpenAI's ViT-L/14@336](https://arxiv.org/abs/2103.00020).
## Checkpoints
| Model | # Seen <BR>Samples (B) | # Params (M) <BR> (img + txt) | Latency (ms) <BR> (img + txt) | IN-1k Zero-Shot <BR> Top-1 Acc. (%) | Avg. Perf. (%) <BR> on 38 datasets |
|:----------------------------------------------------------|:----------------------:|:-----------------------------:|:-----------------------------:|:-----------------------------------:|:----------------------------------:|
| [MobileCLIP2-S0](https://hf.co/apple/MobileCLIP2-S0) | 13 | 11.4 + 42.4 | 1.5 + 1.6 | 71.5 | 59.7 |
| [MobileCLIP2-S2](https://hf.co/apple/MobileCLIP2-S2) | 13 | 35.7 + 63.4 | 3.6 + 3.3 | 77.2 | 64.1 |
| [MobileCLIP2-B](https://hf.co/apple/MobileCLIP2-B) | 13 | 86.3 + 63.4 | 10.4 + 3.3 | 79.4 | 65.8 |
| [MobileCLIP2-S3](https://hf.co/apple/MobileCLIP2-S3) | 13 | 125.1 + 123.6 | 8.0 + 6.6 | 80.7 | 66.8 |
| [MobileCLIP2-L/14](https://hf.co/apple/MobileCLIP2-L-14) | 13 | 304.3 + 123.6 | 57.9 + 6.6 | 81.9 | 67.8 |
| [MobileCLIP2-S4](https://hf.co/apple/MobileCLIP2-S4) | 13 | 321.6 + 123.6 | 19.6 + 6.6 | 81.9 | 67.5 |
| [MobileCLIP-S0](https://hf.co/apple/MobileCLIP-S0) | 13 | 11.4 + 42.4 | 1.5 + 1.6 | 67.8 | 58.1 |
| [MobileCLIP-S1](https://hf.co/apple/MobileCLIP-S1) | 13 | 21.5 + 63.4 | 2.5 + 3.3 | 72.6 | 61.3 |
| [MobileCLIP-S2](https://hf.co/apple/MobileCLIP-S2) | 13 | 35.7 + 63.4 | 3.6 + 3.3 | 74.4 | 63.7 |
| [MobileCLIP-B](https://hf.co/apple/MobileCLIP-B) | 13 | 86.3 + 63.4 | 10.4 + 3.3 | 76.8 | 65.2 |
| [MobileCLIP-B (LT)](https://hf.co/apple/MobileCLIP-B-LT) | 36 | 86.3 + 63.4 | 10.4 + 3.3 | 77.2 | 65.8 |
| [MobileCLIP-S3](https://hf.co/apple/MobileCLIP-S3) | 13 | 125.1 + 123.6 | 8.0 + 6.6 | 78.3 | 66.3 |
| [MobileCLIP-L/14](https://hf.co/apple/MobileCLIP-L-14) | 13 | 304.3 + 123.6 | 57.9 + 6.6 | 79.5 | 66.9 |
| [MobileCLIP-S4](https://hf.co/apple/MobileCLIP-S4) | 13 | 321.6 + 123.6 | 19.6 + 6.6 | 79.4 | 68.1 |
## How to Use
First, download the desired checkpoint visiting one of the links in the table above, then click the `Files and versions` tab, and download the PyTorch checkpoint.
For programmatic downloading, if you have `huggingface_hub` installed, you can also run:
```
hf download apple/MobileCLIP-S4
```
Then, install [`ml-mobileclip`](https://github.com/apple/ml-mobileclip) by following the instructions in the repo. It uses an API similar to [`open_clip`'s](https://github.com/mlfoundations/open_clip).
You can run inference with a code snippet like the following:
```py
import torch
import open_clip
from PIL import Image
from mobileclip.modules.common.mobileone import reparameterize_model
model, _, preprocess = open_clip.create_model_and_transforms('MobileCLIP2-S4', pretrained='/path/to/mobileclip_s4.pt')
tokenizer = open_clip.get_tokenizer('MobileCLIP2-S4')
# For inference/model exporting purposes, please reparameterize first
model = reparameterize_model(model.eval())
image = preprocess(Image.open("docs/fig_accuracy_latency.png").convert('RGB')).unsqueeze(0)
text = tokenizer(["a diagram", "a dog", "a cat"])
with torch.no_grad(), torch.cuda.amp.autocast():
image_features = model.encode_image(image)
text_features = model.encode_text(text)
image_features /= image_features.norm(dim=-1, keepdim=True)
text_features /= text_features.norm(dim=-1, keepdim=True)
text_probs = (100.0 * image_features @ text_features.T).softmax(dim=-1)
print("Label probs:", text_probs)
```
|
apple/MobileCLIP-S3
|
apple
| 2025-08-29T01:26:18Z | 0 | 0 |
mobileclip
|
[
"mobileclip",
"arxiv:2508.20691",
"arxiv:2103.00020",
"arxiv:2303.15343",
"arxiv:2309.17425",
"license:apple-amlr",
"region:us"
] | null | 2025-08-25T16:49:52Z |
---
license: apple-amlr
license_name: apple-ascl
license_link: https://github.com/apple/ml-mobileclip/blob/main/LICENSE_weights_data
library_name: mobileclip
---
# MobileCLIP2: Improving Multi-Modal Reinforced Training
MobileCLIP2 was introduced in [MobileCLIP2: Improving Multi-Modal Reinforced Training](http://arxiv.org/abs/2508.20691) (TMLR August 2025 <mark>Featured</mark>), by Fartash Faghri, Pavan Kumar Anasosalu Vasu, Cem Koc, Vaishaal Shankar, Alexander T Toshev, Oncel Tuzel, Hadi Pouransari.
This repository contains the **MobileCLIP-S3** checkpoint.
### Highlights
* `MobileCLIP2-S4` matches the accuracy of SigLIP-SO400M/14 with 2x fewer parameters and surpasses DFN ViT-L/14 at 2.5x lower latency measured on iPhone12 Pro Max.
* `MobileCLIP-S3/S4` are our new architectures trained on MobileCLIP’s training dataset, DataCompDR-1B (dashed lines).
* Our smallest variant `MobileCLIP-S0` obtains similar zero-shot performance as [OpenAI](https://arxiv.org/abs/2103.00020)'s ViT-B/16 model while being 4.8x faster and 2.8x smaller.
* `MobileCLIP-S2` obtains better avg zero-shot performance than [SigLIP](https://arxiv.org/abs/2303.15343)'s ViT-B/16 model while being 2.3x faster and 2.1x smaller, and trained with 3x less seen samples.
* `MobileCLIP-B (LT)` attains zero-shot ImageNet performance of **77.2%** which is significantly better than recent works like [DFN](https://arxiv.org/abs/2309.17425) and [SigLIP](https://arxiv.org/abs/2303.15343) with similar architectures or even [OpenAI's ViT-L/14@336](https://arxiv.org/abs/2103.00020).
## Checkpoints
| Model | # Seen <BR>Samples (B) | # Params (M) <BR> (img + txt) | Latency (ms) <BR> (img + txt) | IN-1k Zero-Shot <BR> Top-1 Acc. (%) | Avg. Perf. (%) <BR> on 38 datasets |
|:----------------------------------------------------------|:----------------------:|:-----------------------------:|:-----------------------------:|:-----------------------------------:|:----------------------------------:|
| [MobileCLIP2-S0](https://hf.co/apple/MobileCLIP2-S0) | 13 | 11.4 + 42.4 | 1.5 + 1.6 | 71.5 | 59.7 |
| [MobileCLIP2-S2](https://hf.co/apple/MobileCLIP2-S2) | 13 | 35.7 + 63.4 | 3.6 + 3.3 | 77.2 | 64.1 |
| [MobileCLIP2-B](https://hf.co/apple/MobileCLIP2-B) | 13 | 86.3 + 63.4 | 10.4 + 3.3 | 79.4 | 65.8 |
| [MobileCLIP2-S3](https://hf.co/apple/MobileCLIP2-S3) | 13 | 125.1 + 123.6 | 8.0 + 6.6 | 80.7 | 66.8 |
| [MobileCLIP2-L/14](https://hf.co/apple/MobileCLIP2-L-14) | 13 | 304.3 + 123.6 | 57.9 + 6.6 | 81.9 | 67.8 |
| [MobileCLIP2-S4](https://hf.co/apple/MobileCLIP2-S4) | 13 | 321.6 + 123.6 | 19.6 + 6.6 | 81.9 | 67.5 |
| [MobileCLIP-S0](https://hf.co/apple/MobileCLIP-S0) | 13 | 11.4 + 42.4 | 1.5 + 1.6 | 67.8 | 58.1 |
| [MobileCLIP-S1](https://hf.co/apple/MobileCLIP-S1) | 13 | 21.5 + 63.4 | 2.5 + 3.3 | 72.6 | 61.3 |
| [MobileCLIP-S2](https://hf.co/apple/MobileCLIP-S2) | 13 | 35.7 + 63.4 | 3.6 + 3.3 | 74.4 | 63.7 |
| [MobileCLIP-B](https://hf.co/apple/MobileCLIP-B) | 13 | 86.3 + 63.4 | 10.4 + 3.3 | 76.8 | 65.2 |
| [MobileCLIP-B (LT)](https://hf.co/apple/MobileCLIP-B-LT) | 36 | 86.3 + 63.4 | 10.4 + 3.3 | 77.2 | 65.8 |
| [MobileCLIP-S3](https://hf.co/apple/MobileCLIP-S3) | 13 | 125.1 + 123.6 | 8.0 + 6.6 | 78.3 | 66.3 |
| [MobileCLIP-L/14](https://hf.co/apple/MobileCLIP-L-14) | 13 | 304.3 + 123.6 | 57.9 + 6.6 | 79.5 | 66.9 |
| [MobileCLIP-S4](https://hf.co/apple/MobileCLIP-S4) | 13 | 321.6 + 123.6 | 19.6 + 6.6 | 79.4 | 68.1 |
## How to Use
First, download the desired checkpoint visiting one of the links in the table above, then click the `Files and versions` tab, and download the PyTorch checkpoint.
For programmatic downloading, if you have `huggingface_hub` installed, you can also run:
```
hf download apple/MobileCLIP-S3
```
Then, install [`ml-mobileclip`](https://github.com/apple/ml-mobileclip) by following the instructions in the repo. It uses an API similar to [`open_clip`'s](https://github.com/mlfoundations/open_clip).
You can run inference with a code snippet like the following:
```py
import torch
import open_clip
from PIL import Image
from mobileclip.modules.common.mobileone import reparameterize_model
model, _, preprocess = open_clip.create_model_and_transforms('MobileCLIP2-S3', pretrained='/path/to/mobileclip_s3.pt')
tokenizer = open_clip.get_tokenizer('MobileCLIP2-S3')
# For inference/model exporting purposes, please reparameterize first
model = reparameterize_model(model.eval())
image = preprocess(Image.open("docs/fig_accuracy_latency.png").convert('RGB')).unsqueeze(0)
text = tokenizer(["a diagram", "a dog", "a cat"])
with torch.no_grad(), torch.cuda.amp.autocast():
image_features = model.encode_image(image)
text_features = model.encode_text(text)
image_features /= image_features.norm(dim=-1, keepdim=True)
text_features /= text_features.norm(dim=-1, keepdim=True)
text_probs = (100.0 * image_features @ text_features.T).softmax(dim=-1)
print("Label probs:", text_probs)
```
|
Sail1013/my-awesome-model
|
Sail1013
| 2025-08-29T01:25:04Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"bert",
"feature-extraction",
"arxiv:1910.09700",
"endpoints_compatible",
"region:us"
] |
feature-extraction
| 2025-08-29T01:24:58Z |
---
library_name: transformers
tags: []
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
|
unileon-robotics/Trasgu-1.5B
|
unileon-robotics
| 2025-08-29T01:24:11Z | 9 | 0 | null |
[
"pytorch",
"qwen2",
"es",
"dataset:unileon-robotics/lliones-dict-tr",
"base_model:Qwen/Qwen2.5-1.5B",
"base_model:finetune:Qwen/Qwen2.5-1.5B",
"license:apache-2.0",
"region:us"
] | null | 2025-08-26T18:41:17Z |
---
license: apache-2.0
datasets:
- unileon-robotics/lliones-dict-tr
language:
- es
base_model:
- Qwen/Qwen2.5-1.5B
---
<p align="center">
<img src="Utils/Banner-Trasgu.PNG" alt="Trasgu Banner">
</p>
<h1 align="center">Trasgu-1.5B — Llionés (Qwen2.5 + Unsloth, FP16 · PyTorch)</h1>
<p align="center">
<img src="Utils/Logo-Trasgu.png" alt="Trasgu Logo" width="90">
</p>
> Ajuste fino de **Qwen2.5-1.5B** con **Unsloth** en **Google Colab**, entrenado y evaluado con
> el dataset **[`unileon-robotics/lliones-dict-tr`](https://huggingface.co/datasets/unileon-robotics/lliones-dict-tr)** (pares *input→output*).
> Este repositorio publica los **pesos FP16 en formato Transformers** (compatibles con **PyTorch**).
> La **versión GGUF (Q5_K_M)** está disponible en: **[Trasgu-1.5B-GGUF](https://huggingface.co/unileon-robotics/Trasgu-1.5B-GGUF)**.
---
## 📌 Descripción
**Trasgu-1.5B** es un modelo conversacional en **llionés (leonés)** afinado con **aprendizaje supervisado** a partir de pares *entrada→salida* del dataset `lliones-dict-tr`.
Al tratarse de datos **estructurados**, la señal de entrenamiento es directa y permite una **evaluación objetiva** sobre el propio conjunto.
### 🧰 Archivos de utilidad
- Banner: `Utils/Banner-Trasgu.PNG`
- Logo: `Utils/Logo-Trasgu.png`
- Notebook de entrenamiento: `Utils/Training_Notebook.ipynb`
- Configuración (hiperparámetros y formato): `Utils/config_entrenamiento.json`
---
## 🗂️ Dataset `lliones-dict-tr`
Dataset **estructurado** de pares *input–output* (traducciones, vocabulario y significados llionés↔español).
Para el *fine-tuning* se adapta cada ejemplo al **formato chat** con la plantilla **Qwen 2.5** (instrucción → respuesta), alineando el comportamiento conversacional.
---
## 🧪 Metodología de evaluación
Para cada muestra del dataset (N = 3000):
1. **Se consulta el modelo** con la **entrada** (pregunta/instrucción).
2. **Se compara la salida** del modelo con la **respuesta de referencia** mediante dos criterios:
**Métricas por muestra (escala 0–100):**
- **Fuzzy** (`mean_fuzzy`): *fuzzy ratio* “normal” (coincidencia textual aproximada).
- **Semántica** (`mean_semantic`): **cosine similarity** entre *embeddings* de
[`sentence-transformers/all-MiniLM-L6-v2`](https://huggingface.co/sentence-transformers/all-MiniLM-L6-v2).
**Puntuación final por muestra:**
`score_muestra = max(fuzzy, semantic)`
**Agregados del experimento:**
- `mean_fuzzy` = media de *fuzzy ratio*.
- `mean_semantic` = media de similaridad semántica.
- **`mean_similarity`** = media de `score_muestra` → **métrica final**.
> Este esquema valida tanto coincidencias **literales** (fuzzy alto) como **paráfrasis correctas** (semántica alta).
---
## 📊 Resultados (1.5B · 3.000 ejemplos)
> La fila **FP16 (Transformers)** corresponde a **este repositorio**.
> La fila **GGUF Q5_K_M** corresponde a la versión en **[Trasgu-1.5B-GGUF](https://huggingface.co/unileon-robotics/Trasgu-1.5B-GGUF)**.
| Modelo / Formato | mean_similarity (final) | mean_fuzzy | mean_semantic |
|---|---:|---:|---:|
| **Trasgu-1.5B — FP16 (Transformers)** | **79.50** | 58.48 | 79.39 |
| **Trasgu-1.5B — GGUF Q5_K_M** ([repo GGUF](https://huggingface.co/unileon-robotics/Trasgu-1.5B-GGUF)) | **79.51** | 58.51 | 79.37 |
**Lectura rápida:** la cuantización **Q5_K_M** mantiene resultados **prácticamente idénticos** a FP16 en la métrica final y en la semántica.
---
## 🔧 Entrenamiento (resumen)
- **Base:** Qwen2.5-1.5B (vía **Unsloth**)
- **Épocas:** 3 · **max_seq_length:** 4096
- **Optimización:** `paged_adamw_8bit` · **LR:** 2e-4 (scheduler lineal, *warmup_ratio* 0.1, *weight_decay* 0.01)
- **Precisión:** **FP16**
- **LoRA:** r=128, α=256, *dropout*=0.1, sin *bias*, aplicado a proyecciones y `lm_head`
- **Exportación a GGUF (en repo aparte):** artefacto **Q5_K_M**
- Referencias: `Utils/Training_Notebook.ipynb` · `Utils/config_entrenamiento.json`
---
## ⚠️ Notas
- El modelo está **especializado** en los tipos de pares presentes en `lliones-dict-tr`.
- Las métricas automáticas (fuzzy y semántica) son **indicativas**; la evaluación humana puede diferir.
|
cybertruck32489/gemma-3-1b-it-pretrain
|
cybertruck32489
| 2025-08-29T01:21:46Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"gemma3_text",
"text-generation",
"text-generation-inference",
"unsloth",
"conversational",
"en",
"base_model:unsloth/gemma-3-1b-it-unsloth-bnb-4bit",
"base_model:finetune:unsloth/gemma-3-1b-it-unsloth-bnb-4bit",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-08-28T21:03:09Z |
---
base_model: unsloth/gemma-3-1b-it-unsloth-bnb-4bit
tags:
- text-generation-inference
- transformers
- unsloth
- gemma3_text
license: apache-2.0
language:
- en
---
# Uploaded finetuned model
- **Developed by:** cybertruck32489
- **License:** apache-2.0
- **Finetuned from model :** unsloth/gemma-3-1b-it-unsloth-bnb-4bit
This gemma3_text model was trained 2x faster with [Unsloth](https://github.com/unslothai/unsloth) and Huggingface's TRL library.
[<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png" width="200"/>](https://github.com/unslothai/unsloth)
|
hyunjong7/gemma-fire-finetun-4b-test
|
hyunjong7
| 2025-08-29T01:20:06Z | 0 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"generated_from_trainer",
"trl",
"sft",
"base_model:google/gemma-3-4b-it",
"base_model:finetune:google/gemma-3-4b-it",
"endpoints_compatible",
"region:us"
] | null | 2025-07-03T02:17:16Z |
---
base_model: google/gemma-3-4b-it
library_name: transformers
model_name: gemma-fire-finetun-4b-test
tags:
- generated_from_trainer
- trl
- sft
licence: license
---
# Model Card for gemma-fire-finetun-4b-test
This model is a fine-tuned version of [google/gemma-3-4b-it](https://huggingface.co/google/gemma-3-4b-it).
It has been trained using [TRL](https://github.com/huggingface/trl).
## Quick start
```python
from transformers import pipeline
question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
generator = pipeline("text-generation", model="hyunjong7/gemma-fire-finetun-4b-test", device="cuda")
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
print(output["generated_text"])
```
## Training procedure
This model was trained with SFT.
### Framework versions
- TRL: 0.21.0
- Transformers: 4.55.4
- Pytorch: 2.8.0
- Datasets: 3.0.1
- Tokenizers: 0.21.1
## Citations
Cite TRL as:
```bibtex
@misc{vonwerra2022trl,
title = {{TRL: Transformer Reinforcement Learning}},
author = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallou{\'e}dec},
year = 2020,
journal = {GitHub repository},
publisher = {GitHub},
howpublished = {\url{https://github.com/huggingface/trl}}
}
```
|
thejaminator/qwen-hook-layer-9-step-500
|
thejaminator
| 2025-08-29T01:18:03Z | 0 | 0 |
peft
|
[
"peft",
"safetensors",
"qwen3",
"base_model:Qwen/Qwen3-8B",
"base_model:adapter:Qwen/Qwen3-8B",
"region:us"
] | null | 2025-08-29T01:17:41Z |
---
base_model: Qwen/Qwen3-8B
library_name: peft
---
# LoRA Adapter for SAE Introspection
This is a LoRA (Low-Rank Adaptation) adapter trained for SAE (Sparse Autoencoder) introspection tasks.
## Base Model
- **Base Model**: `Qwen/Qwen3-8B`
- **Adapter Type**: LoRA
- **Task**: SAE Feature Introspection
## Usage
```python
from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel
# Load base model and tokenizer
base_model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen3-8B")
tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen3-8B")
# Load LoRA adapter
model = PeftModel.from_pretrained(base_model, "thejaminator/qwen-hook-layer-9-step-500")
```
## Training Details
This adapter was trained using the lightweight SAE introspection training script to help the model understand and explain SAE features through activation steering.
|
bah63843/blockassist-bc-plump_fast_antelope_1756430153
|
bah63843
| 2025-08-29T01:16:51Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"plump fast antelope",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T01:16:41Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- plump fast antelope
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
haritsondavid/blockassist-bc-padded_howling_anteater_1756429945
|
haritsondavid
| 2025-08-29T01:13:16Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"padded howling anteater",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T01:12:50Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- padded howling anteater
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
Rustamshry/Llama-electronic-radiology-TR
|
Rustamshry
| 2025-08-29T01:10:43Z | 38 | 1 |
peft
|
[
"peft",
"safetensors",
"radiology",
"transformers",
"pretraining",
"trl",
"text-generation",
"tr",
"dataset:hazal/electronic-radiology-phd-thesis-trR",
"base_model:unsloth/Llama-3.2-1B",
"base_model:adapter:unsloth/Llama-3.2-1B",
"license:llama3.2",
"region:us"
] |
text-generation
| 2025-06-10T01:50:33Z |
---
base_model: unsloth/Llama-3.2-1B
library_name: peft
license: llama3.2
datasets:
- hazal/electronic-radiology-phd-thesis-trR
language:
- tr
pipeline_tag: text-generation
tags:
- radiology
- transformers
- pretraining
- trl
---
# Model Card for Llama-electronic-radiology-TR
## Model Details
## Model Summary
This model is a domain-adapted version of Llama-3.2-1B, fine-tuned via continued pretraining on Turkish-language electronic radiology PhD theses.
The model was trained in an autoregressive (causal language modeling) setup using the [hazal/electronic-radiology-phd-thesis-trR](https://huggingface.co/datasets/hazal/electronic-radiology-phd-thesis-trR) dataset.
Unlike instruction-tuned models, this version focuses on improving the model’s fluency, vocabulary, and semantic consistency in highly technical medical and radiological contexts.
It is intended for downstream applications such as domain-specific generation, summarization, and potential fine-tuning for clinical tasks in Turkish.
### Model Description
- **Language(s) (NLP):** Turkish
- **License:** Llama 3.2
- **Finetuned from model:** Llama-3.2-1B
## Uses
### Direct Use
The primary intended uses include:
- **Domain-specific generation**: Generating fluent, semantically rich Turkish text in radiological contexts, e.g., imaging protocols, research summaries, or academic abstracts.
- **Medical document summarization**: Summarizing long Turkish-language radiological texts, such as reports or thesis chapters.
- **Language modeling for downstream tasks**: Serving as a base model for fine-tuning into instruction-tuned clinical models or QA systems in radiology.
- **Research applications**: Assisting in the development of Turkish-language models for clinical NLP, especially in low-resource and domain-specific contexts.
This model **is not instruction-tuned** and does not perform well in prompt-based Q&A or dialogue setups without additional supervised fine-tuning.
## Bias, Risks, and Limitations
### 🔬 Domain Bias
The model has been trained exclusively on Turkish PhD-level academic texts in radiology. As such, its knowledge and language patterns are narrowly focused on:
- Formal, academic Turkish
- Medical terminology in radiology and imaging
- Structured dissertation-like content
It may underperform or produce awkward completions when applied to:
- Conversational Turkish
- Non-medical or non-radiological topics
- Informal writing styles or dialectal Turkish
### ❌ Medical Safety
This model **should not be used for clinical decision-making, diagnosis, or treatment recommendations**. Despite being trained on medical content, it lacks factual grounding, context awareness, and real-time clinical judgment.
Any outputs generated by this model must be verified by licensed medical professionals.
### 🧠 Memorization Risk
Continued pretraining on a small or repetitive corpus can lead to memorization of phrases, potentially exposing:
- Patient case formats
- Study identifiers
- Sections of dissertations
Although this dataset appears anonymized and academic, **you should not use the model for data anonymization, patient privacy protection, or regulatory compliance tasks**.
### 🧪 Limitations
- The model does not have grounding in real-world imaging data or structured knowledge bases.
- Outputs may hallucinate plausible-sounding but incorrect medical facts.
- Limited to Turkish; does not generalize to multilingual or English medical contexts.
- Repetition or looping in generation may still occur in long sequences if decoding is not configured properly (`repetition_penalty`, `eos_token_id`, etc.).
## How to Get Started with the Model
Use the code below to get started with the model.
```python
from huggingface_hub import login
from transformers import AutoTokenizer, AutoModelForCausalLM
import torch
from peft import PeftModel
login(token="")
tokenizer = AutoTokenizer.from_pretrained("unsloth/Llama-3.2-1B",)
base_model = AutoModelForCausalLM.from_pretrained(
"unsloth/Llama-3.2-1B",
device_map={"":0}, token=""
)
model = PeftModel.from_pretrained(base_model,"Rustamshry/Llama-electronic-radiology-TR")
input_text = "Bulgular: Gruplar arası yaş ve cinsiyet dağılımı açısından istatiksel olarakanlamlı farklılık saptanmadı."
inputs = tokenizer(input_text, return_tensors="pt", add_special_tokens=True).to(model.device)
outputs = model.generate(
**inputs,
max_new_tokens=512,
do_sample=True,
temperature=1.0,
top_p=0.95,
repetition_penalty = 1.2,
eos_token_id=tokenizer.eos_token_id
)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))
```
## Training Details
### Training Data
- **Hours used:** 10 hours
## Dataset: hazal/electronic-radiology-phd-thesis-trR
This dataset contains Turkish-language PhD theses focused on electronic and diagnostic radiology. It was curated for the purpose of training language models in the Turkish medical domain. The documents are academic in tone, rich in domain-specific vocabulary, and structured into medical sections (e.g., materials & methods, results, discussion).
- **Language**: Turkish
- **Domain**: Electronic Radiology
- **Type**: Academic dissertations
- **Preprocessing**: The dataset was tokenized and truncated to a maximum sequence length suitable for LLM training. No instruction-style formatting was applied.
Dataset link: [hazal/electronic-radiology-phd-thesis-trR](https://huggingface.co/datasets/hazal/electronic-radiology-phd-thesis-trR)
### Framework versions
- PEFT 0.14.0
|
bah63843/blockassist-bc-plump_fast_antelope_1756429759
|
bah63843
| 2025-08-29T01:10:10Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"plump fast antelope",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T01:09:59Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- plump fast antelope
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
Rustamshry/QwenAzFT-v1
|
Rustamshry
| 2025-08-29T01:06:53Z | 6 | 1 |
peft
|
[
"peft",
"safetensors",
"alpaca",
"unsloth",
"transformers",
"trl",
"sft",
"text-generation",
"conversational",
"az",
"dataset:omar07ibrahim/Alpaca_Stanford_Azerbaijan",
"base_model:unsloth/Qwen3-1.7B",
"base_model:adapter:unsloth/Qwen3-1.7B",
"license:mit",
"region:us"
] |
text-generation
| 2025-05-24T12:22:38Z |
---
base_model: unsloth/Qwen3-1.7B
library_name: peft
license: mit
datasets:
- omar07ibrahim/Alpaca_Stanford_Azerbaijan
language:
- az
pipeline_tag: text-generation
tags:
- alpaca
- unsloth
- transformers
- trl
- sft
---
# Model Card for QwenAzFT-v1
## Model Details
This model is a fine-tuned version of [`unsloth/Qwen3-1.7B`](https://huggingface.co/unsloth/Qwen3-1.7B) on a translated version of the **Alpaca Stanford dataset** in **Azerbaijani language**.
The model is instruction-tuned to better follow prompts and generate relevant responses in Azerbaijani.
### Model Description
- **Language(s) (NLP):** Azerbaijani
- **License:** MIT
- **Finetuned from model:** unsloth/Qwen3-1.7B
## Uses
### Direct Use
- Instruction following in Azerbaijani
- Education, research, and experimentation with low-resource language LLMs
- Chatbots, task-oriented systems, language agents
## How to Get Started with the Model
Use the code below to get started with the model.
```python
from huggingface_hub import login
from transformers import AutoTokenizer, AutoModelForCausalLM
from peft import PeftModel
login(token="")
tokenizer = AutoTokenizer.from_pretrained("unsloth/Qwen3-1.7B",)
base_model = AutoModelForCausalLM.from_pretrained(
"unsloth/Qwen3-1.7B",
device_map={"": 0}, token=""
)
model = PeftModel.from_pretrained(base_model,"Rustamshry/QwenAzFT-v1")
question = "Bir sifət əlavə edərək aşağıdakı cümləni yenidən yazın. Tələbə mürəkkəb anlayışları anlaya bildi. "
messages = [
{"role" : "user", "content" : question}
]
text = tokenizer.apply_chat_template(
messages,
tokenize = False,
add_generation_prompt = True,
enable_thinking = False,
)
from transformers import TextStreamer
_ = model.generate(
**tokenizer(text, return_tensors = "pt").to("cuda"),
max_new_tokens = 2048,
temperature = 0.7,
top_p = 0.8,
top_k = 20,
streamer = TextStreamer(tokenizer, skip_prompt = True),
)
```
## Training Data
- **Dataset:** omar07ibrahim/Alpaca_Stanford_Azerbaijan
### Framework versions
- PEFT 0.14.0
|
chainway9/blockassist-bc-untamed_quick_eel_1756427947
|
chainway9
| 2025-08-29T01:04:55Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"untamed quick eel",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T01:04:51Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- untamed quick eel
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
coelacanthxyz/blockassist-bc-finicky_thriving_grouse_1756427542
|
coelacanthxyz
| 2025-08-29T00:59:33Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"finicky thriving grouse",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T00:59:26Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- finicky thriving grouse
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
memphiskol/blockassist-bc-deadly_hardy_flea_1756429046
|
memphiskol
| 2025-08-29T00:58:13Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"deadly hardy flea",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T00:57:53Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- deadly hardy flea
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
gsjang/ko-llama-3-korean-bllossom-8b-x-meta-llama-3-8b-instruct-arcee_fusion-50_50
|
gsjang
| 2025-08-29T00:57:58Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"llama",
"text-generation",
"mergekit",
"merge",
"conversational",
"base_model:MLP-KTLim/llama-3-Korean-Bllossom-8B",
"base_model:merge:MLP-KTLim/llama-3-Korean-Bllossom-8B",
"base_model:meta-llama/Meta-Llama-3-8B-Instruct",
"base_model:merge:meta-llama/Meta-Llama-3-8B-Instruct",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-08-29T00:52:05Z |
---
base_model:
- MLP-KTLim/llama-3-Korean-Bllossom-8B
- meta-llama/Meta-Llama-3-8B-Instruct
library_name: transformers
tags:
- mergekit
- merge
---
# ko-llama-3-korean-bllossom-8b-x-meta-llama-3-8b-instruct-arcee_fusion-50_50
This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).
## Merge Details
### Merge Method
This model was merged using the [Arcee Fusion](https://arcee.ai) merge method using [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) as a base.
### Models Merged
The following models were included in the merge:
* [MLP-KTLim/llama-3-Korean-Bllossom-8B](https://huggingface.co/MLP-KTLim/llama-3-Korean-Bllossom-8B)
### Configuration
The following YAML configuration was used to produce this model:
```yaml
merge_method: arcee_fusion
models:
- model: MLP-KTLim/llama-3-Korean-Bllossom-8B
parameters:
weight: 0.5
- model: meta-llama/Meta-Llama-3-8B-Instruct
parameters:
weight: 0.5
parameters: {}
dtype: bfloat16
tokenizer:
source: union
base_model: meta-llama/Meta-Llama-3-8B-Instruct
write_readme: README.md
```
|
bah63843/blockassist-bc-plump_fast_antelope_1756428441
|
bah63843
| 2025-08-29T00:48:20Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"plump fast antelope",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T00:48:11Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- plump fast antelope
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
calexander0614/output
|
calexander0614
| 2025-08-29T00:40:31Z | 0 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"generated_from_trainer",
"sft",
"trl",
"base_model:mistralai/Mistral-Small-3.1-24B-Instruct-2503",
"base_model:finetune:mistralai/Mistral-Small-3.1-24B-Instruct-2503",
"endpoints_compatible",
"region:us"
] | null | 2025-08-28T14:30:51Z |
---
base_model: mistralai/Mistral-Small-3.1-24B-Instruct-2503
library_name: transformers
model_name: output
tags:
- generated_from_trainer
- sft
- trl
licence: license
---
# Model Card for output
This model is a fine-tuned version of [mistralai/Mistral-Small-3.1-24B-Instruct-2503](https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503).
It has been trained using [TRL](https://github.com/huggingface/trl).
## Quick start
```python
from transformers import pipeline
question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
generator = pipeline("text-generation", model="calexander0614/output", device="cuda")
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
print(output["generated_text"])
```
## Training procedure
This model was trained with SFT.
### Framework versions
- TRL: 0.21.0
- Transformers: 4.55.4
- Pytorch: 2.5.1+cu124
- Datasets: 3.6.0
- Tokenizers: 0.21.2
## Citations
Cite TRL as:
```bibtex
@misc{vonwerra2022trl,
title = {{TRL: Transformer Reinforcement Learning}},
author = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallou{\'e}dec},
year = 2020,
journal = {GitHub repository},
publisher = {GitHub},
howpublished = {\url{https://github.com/huggingface/trl}}
}
```
|
goppetoyu/blockassist-bc-barky_wiry_chicken_1756427908
|
goppetoyu
| 2025-08-29T00:38:28Z | 0 | 0 | null |
[
"region:us"
] | null | 2025-08-29T00:38:28Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- barky wiry chicken
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
Pomni/whisper-large-v2-ggml-allquants
|
Pomni
| 2025-08-29T00:38:10Z | 0 | 0 | null |
[
"whisper.cpp",
"ggml",
"whisper",
"audio",
"speech",
"voice",
"automatic-speech-recognition",
"en",
"zh",
"de",
"es",
"ru",
"ko",
"fr",
"ja",
"pt",
"tr",
"pl",
"ca",
"nl",
"ar",
"sv",
"it",
"id",
"hi",
"fi",
"vi",
"he",
"uk",
"el",
"ms",
"cs",
"ro",
"da",
"hu",
"ta",
"no",
"th",
"ur",
"hr",
"bg",
"lt",
"la",
"mi",
"ml",
"cy",
"sk",
"te",
"fa",
"lv",
"bn",
"sr",
"az",
"sl",
"kn",
"et",
"mk",
"br",
"eu",
"is",
"hy",
"ne",
"mn",
"bs",
"kk",
"sq",
"sw",
"gl",
"mr",
"pa",
"si",
"km",
"sn",
"yo",
"so",
"af",
"oc",
"ka",
"be",
"tg",
"sd",
"gu",
"am",
"yi",
"lo",
"uz",
"fo",
"ht",
"ps",
"tk",
"nn",
"mt",
"sa",
"lb",
"my",
"bo",
"tl",
"mg",
"as",
"tt",
"haw",
"ln",
"ha",
"ba",
"jw",
"su",
"base_model:openai/whisper-large-v2",
"base_model:finetune:openai/whisper-large-v2",
"license:apache-2.0",
"region:us"
] |
automatic-speech-recognition
| 2025-08-26T05:54:34Z |
---
license: apache-2.0
quantized_by: Pomni
language:
- en
- zh
- de
- es
- ru
- ko
- fr
- ja
- pt
- tr
- pl
- ca
- nl
- ar
- sv
- it
- id
- hi
- fi
- vi
- he
- uk
- el
- ms
- cs
- ro
- da
- hu
- ta
- 'no'
- th
- ur
- hr
- bg
- lt
- la
- mi
- ml
- cy
- sk
- te
- fa
- lv
- bn
- sr
- az
- sl
- kn
- et
- mk
- br
- eu
- is
- hy
- ne
- mn
- bs
- kk
- sq
- sw
- gl
- mr
- pa
- si
- km
- sn
- yo
- so
- af
- oc
- ka
- be
- tg
- sd
- gu
- am
- yi
- lo
- uz
- fo
- ht
- ps
- tk
- nn
- mt
- sa
- lb
- my
- bo
- tl
- mg
- as
- tt
- haw
- ln
- ha
- ba
- jw
- su
base_model:
- openai/whisper-large-v2
pipeline_tag: automatic-speech-recognition
tags:
- whisper.cpp
- ggml
- whisper
- audio
- speech
- voice
new_version: Pomni/whisper-large-v3-ggml-allquants
---
# Whisper-Large-v2 quants
This is a repository of **GGML quants for [whisper-large-v2](https://huggingface.co/openai/whisper-large-v2)**, for use with [whisper.cpp](https://github.com/ggml-org/whisper.cpp).
If you are looking for a program to run this model with, then I would recommend [EasyWhisper UI](https://github.com/mehtabmahir/easy-whisper-ui), as it is user-friendly, has a GUI, and will automate a lot of the hard stuff for you.
## List of Quants
Clicking on a link will download the corresponding quant instantly.
| Link | Quant | Size | Notes
|:-----|:-----|--------:|:------|
| [GGML](https://huggingface.co/Pomni/whisper-large-v2-ggml-allquants/resolve/main/ggml-large-v2-f32.bin) | F32 | 6.17 GB | Likely overkill. |
| [GGML](https://huggingface.co/Pomni/whisper-large-v2-ggml-allquants/resolve/main/ggml-large-v2-f16.bin) | F16 | 3.09 GB | Performs better than Q8_0 for noisy audio and music. |
| [GGML](https://huggingface.co/Pomni/whisper-large-v2-ggml-allquants/resolve/main/ggml-large-v2-q8_0.bin) | Q8_0 | 1.66 GB | Sweet spot; superficial quality loss at nearly double the speed. |
| [GGML](https://huggingface.co/Pomni/whisper-large-v2-ggml-allquants/resolve/main/ggml-large-v2-q6_k.bin) | Q6_K | 1.28 GB | |
| [GGML](https://huggingface.co/Pomni/whisper-large-v2-ggml-allquants/resolve/main/ggml-large-v2-q5_k.bin) | Q5_K | 1.08 GB | |
| [GGML](https://huggingface.co/Pomni/whisper-large-v2-ggml-allquants/resolve/main/ggml-large-v2-q5_1.bin) | Q5_1 | 1.18 GB | |
| [GGML](https://huggingface.co/Pomni/whisper-large-v2-ggml-allquants/resolve/main/ggml-large-v2-q5_0.bin) | Q5_0 | 1.08 GB | Last "good" quant; anything below loses quality rapidly. |
| [GGML](https://huggingface.co/Pomni/whisper-large-v2-ggml-allquants/resolve/main/ggml-large-v2-q4_k.bin) | Q4_K | 889 MB | *Might* not have lost too much quality, but I'm not sure. |
| [GGML](https://huggingface.co/Pomni/whisper-large-v2-ggml-allquants/resolve/main/ggml-large-v2-q4_1.bin) | Q4_1 | 985 MB | |
| [GGML](https://huggingface.co/Pomni/whisper-large-v2-ggml-allquants/resolve/main/ggml-large-v2-q4_0.bin) | Q4_0 | 889 MB | |
| [GGML](https://huggingface.co/Pomni/whisper-large-v2-ggml-allquants/resolve/main/ggml-large-v2-q3_k.bin) | Q3_K | 685 MB | |
| [GGML](https://huggingface.co/Pomni/whisper-large-v2-ggml-allquants/resolve/main/ggml-large-v2-q2_k.bin) | Q2_K | 529 MB | Completely non-sensical outputs. |
The F16 quant was taken from [ggerganov/whisper.cpp/ggml-large-v2.bin](https://huggingface.co/ggerganov/whisper.cpp/blob/main/ggml-large-v2.bin).
## Questions you may have
### Why do the "K-quants" not work for me?
My guess is that your GPU might be too old to recognize them, considering that I have gotten the same error on my GTX 1080. If you would like to run them regardless, you can try switching to CPU inference.
### Are the K-quants "S", "M", or "L"?
The quantizer I was using was not specific about this, so I do not know about this either.
### What program did you use to make these quants?
I used [whisper.cpp v1.7.6](https://github.com/ggml-org/whisper.cpp/releases/tag/v1.7.6) on Windows x64, leveraging CUDA 12.4.0. For the F32 quant, I converted the original Hugging Face (H5) format model to a GGML using the `models/convert-h5-to-ggml.py` script.
### One or multiple of the quants are not working for me.
[Open a new discussion](https://huggingface.co/Pomni/whisper-large-v2-ggml-allquants/discussions/new) in the community tab about this, and I will look into the issue.
|
iatop65/Civitai_Backup
|
iatop65
| 2025-08-29T00:35:48Z | 0 | 0 | null |
[
"Backup",
"CivArchive",
"en",
"region:us"
] | null | 2025-08-29T00:27:07Z |
---
language:
- en
tags:
- Backup
- CivArchive
---
Lora Backup Upload From Civitai to be use on CivArchive for Mirrors
Made by iatop65/iatop68304 On Civitai
|
Loder-S/blockassist-bc-sprightly_knobby_tiger_1756425851
|
Loder-S
| 2025-08-29T00:32:29Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"sprightly knobby tiger",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T00:32:24Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- sprightly knobby tiger
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
memphiskol/blockassist-bc-deadly_hardy_flea_1756427439
|
memphiskol
| 2025-08-29T00:31:23Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"deadly hardy flea",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T00:31:05Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- deadly hardy flea
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
pmcheng/leon_bin_picking_new
|
pmcheng
| 2025-08-29T00:29:49Z | 0 | 0 |
lerobot
|
[
"lerobot",
"safetensors",
"robotics",
"smolvla",
"dataset:pmcheng/leon_bin_picking_new",
"arxiv:2506.01844",
"base_model:lerobot/smolvla_base",
"base_model:finetune:lerobot/smolvla_base",
"license:apache-2.0",
"region:us"
] |
robotics
| 2025-08-29T00:19:20Z |
---
base_model: lerobot/smolvla_base
datasets: pmcheng/leon_bin_picking_new
library_name: lerobot
license: apache-2.0
model_name: smolvla
pipeline_tag: robotics
tags:
- lerobot
- robotics
- smolvla
---
# Model Card for smolvla
<!-- Provide a quick summary of what the model is/does. -->
[SmolVLA](https://huggingface.co/papers/2506.01844) is a compact, efficient vision-language-action model that achieves competitive performance at reduced computational costs and can be deployed on consumer-grade hardware.
This policy has been trained and pushed to the Hub using [LeRobot](https://github.com/huggingface/lerobot).
See the full documentation at [LeRobot Docs](https://huggingface.co/docs/lerobot/index).
---
## How to Get Started with the Model
For a complete walkthrough, see the [training guide](https://huggingface.co/docs/lerobot/il_robots#train-a-policy).
Below is the short version on how to train and run inference/eval:
### Train from scratch
```bash
lerobot-train \
--dataset.repo_id=${HF_USER}/<dataset> \
--policy.type=act \
--output_dir=outputs/train/<desired_policy_repo_id> \
--job_name=lerobot_training \
--policy.device=cuda \
--policy.repo_id=${HF_USER}/<desired_policy_repo_id>
--wandb.enable=true
```
_Writes checkpoints to `outputs/train/<desired_policy_repo_id>/checkpoints/`._
### Evaluate the policy/run inference
```bash
lerobot-record \
--robot.type=so100_follower \
--dataset.repo_id=<hf_user>/eval_<dataset> \
--policy.path=<hf_user>/<desired_policy_repo_id> \
--episodes=10
```
Prefix the dataset repo with **eval\_** and supply `--policy.path` pointing to a local or hub checkpoint.
---
## Model Details
- **License:** apache-2.0
|
nvidia/FasterViT
|
nvidia
| 2025-08-29T00:29:36Z | 0 | 19 | null |
[
"dataset:imagenet-1k",
"arxiv:2306.06189",
"license:other",
"region:us"
] | null | 2023-07-19T17:53:05Z |
---
license: other
datasets:
- imagenet-1k
---
[**FasterViT: Fast Vision Transformers with Hierarchical Attention**](https://arxiv.org/abs/2306.06189).
FasterViT achieves a new SOTA Pareto-front in
terms of accuracy vs. image throughput without extra training data !
<p align="center">
<img src="https://github.com/NVlabs/FasterViT/assets/26806394/253d1a2e-b5f5-4a9b-a362-6cdd16bfccc1" width=62% height=62%
class="center">
</p>
Note: Please use the [**latest NVIDIA TensorRT release**](https://docs.nvidia.com/deeplearning/tensorrt/container-release-notes/index.html) to enjoy the benefits of optimized FasterViT ops.
## Quick Start
We can import pre-trained FasterViT models with **1 line of code**. First, FasterViT can be simply installed by:
```bash
pip install fastervit
```
A pretrained FasterViT model with default hyper-parameters can be created as in the following:
```python
>>> from fastervit import create_model
# Define fastervit-0 model with 224 x 224 resolution
>>> model = create_model('faster_vit_0_224',
pretrained=True,
model_path="/tmp/faster_vit_0.pth.tar")
```
`model_path` is used to set the directory to download the model.
We can also simply test the model by passing a dummy input image. The output is the logits:
```python
>>> import torch
>>> image = torch.rand(1, 3, 224, 224)
>>> output = model(image) # torch.Size([1, 1000])
```
We can also use the any-resolution FasterViT model to accommodate arbitrary image resolutions. In the following, we define an any-resolution FasterViT-0
model with input resolution of 576 x 960, window sizes of 12 and 6 in 3rd and 4th stages, carrier token size of 2 and embedding dimension of
64:
```python
>>> from fastervit import create_model
# Define any-resolution FasterViT-0 model with 576 x 960 resolution
>>> model = create_model('faster_vit_0_any_res',
resolution=[576, 960],
window_size=[7, 7, 12, 6],
ct_size=2,
dim=64,
pretrained=True)
```
Note that the above model is intiliazed from the original ImageNet pre-trained FasterViT with original resolution of 224 x 224. As a result, missing keys and mis-matches could be expected since we are addign new layers (e.g. addition of new carrier tokens, etc.)
We can simply test the model by passing a dummy input image. The output is the logits:
```python
>>> import torch
>>> image = torch.rand(1, 3, 576, 960)
>>> output = model(image) # torch.Size([1, 1000])
```
---
## Results + Pretrained Models
### ImageNet-1K
**FasterViT ImageNet-1K Pretrained Models**
<table>
<tr>
<th>Name</th>
<th>Acc@1(%)</th>
<th>Acc@5(%)</th>
<th>Throughput(Img/Sec)</th>
<th>Resolution</th>
<th>#Params(M)</th>
<th>FLOPs(G)</th>
<th>Download</th>
</tr>
<tr>
<td>FasterViT-0</td>
<td>82.1</td>
<td>95.9</td>
<td>5802</td>
<td>224x224</td>
<td>31.4</td>
<td>3.3</td>
<td><a href="https://drive.google.com/uc?export=download&id=1twI2LFJs391Yrj8MR4Ui9PfrvWqjE1iB">model</a></td>
</tr>
<tr>
<td>FasterViT-1</td>
<td>83.2</td>
<td>96.5</td>
<td>4188</td>
<td>224x224</td>
<td>53.4</td>
<td>5.3</td>
<td><a href="https://drive.google.com/uc?export=download&id=1r7W10n5-bFtM3sz4bmaLrowN2gYPkLGT">model</a></td>
</tr>
<tr>
<td>FasterViT-2</td>
<td>84.2</td>
<td>96.8</td>
<td>3161</td>
<td>224x224</td>
<td>75.9</td>
<td>8.7</td>
<td><a href="https://drive.google.com/uc?export=download&id=1n_a6s0pgi0jVZOGmDei2vXHU5E6RH5wU">model</a></td>
</tr>
<tr>
<td>FasterViT-3</td>
<td>84.9</td>
<td>97.2</td>
<td>1780</td>
<td>224x224</td>
<td>159.5</td>
<td>18.2</td>
<td><a href="https://drive.google.com/uc?export=download&id=1tvWElZ91Sia2SsXYXFMNYQwfipCxtI7X">model</a></td>
</tr>
<tr>
<td>FasterViT-4</td>
<td>85.4</td>
<td>97.3</td>
<td>849</td>
<td>224x224</td>
<td>424.6</td>
<td>36.6</td>
<td><a href="https://drive.google.com/uc?export=download&id=1gYhXA32Q-_9C5DXel17avV_ZLoaHwdgz">model</a></td>
</tr>
<tr>
<td>FasterViT-5</td>
<td>85.6</td>
<td>97.4</td>
<td>449</td>
<td>224x224</td>
<td>975.5</td>
<td>113.0</td>
<td><a href="https://drive.google.com/uc?export=download&id=1mqpai7XiHLr_n1tjxjzT8q369xTCq_z-">model</a></td>
</tr>
<tr>
<td>FasterViT-6</td>
<td>85.8</td>
<td>97.4</td>
<td>352</td>
<td>224x224</td>
<td>1360.0</td>
<td>142.0</td>
<td><a href="https://drive.google.com/uc?export=download&id=12jtavR2QxmMzcKwPzWe7kw-oy34IYi59">model</a></td>
</tr>
</table>
### ImageNet-21K
**FasterViT ImageNet-21K Pretrained Models (ImageNet-1K Fine-tuned)**
<table>
<tr>
<th>Name</th>
<th>Acc@1(%)</th>
<th>Acc@5(%)</th>
<th>Resolution</th>
<th>#Params(M)</th>
<th>FLOPs(G)</th>
<th>Download</th>
</tr>
<tr>
<td>FasterViT-4-21K-224</td>
<td>86.6</td>
<td>97.8</td>
<td>224x224</td>
<td>271.9</td>
<td>40.8</td>
<td><a href="https://huggingface.co/ahatamiz/FasterViT/resolve/main/fastervit_4_21k_224_w14.pth.tar">model</a></td>
</tr>
<tr>
<td>FasterViT-4-21K-384</td>
<td>87.6</td>
<td>98.3</td>
<td>384x384</td>
<td>271.9</td>
<td>120.1</td>
<td><a href="https://huggingface.co/ahatamiz/FasterViT/resolve/main/fastervit_4_21k_384_w24.pth.tar">model</a></td>
</tr>
<tr>
<td>FasterViT-4-21K-512</td>
<td>87.8</td>
<td>98.4</td>
<td>512x512</td>
<td>271.9</td>
<td>213.5</td>
<td><a href="https://huggingface.co/ahatamiz/FasterViT/resolve/main/fastervit_4_21k_512_w32.pth.tar">model</a></td>
</tr>
<tr>
<td>FasterViT-4-21K-768</td>
<td>87.9</td>
<td>98.5</td>
<td>768x768</td>
<td>271.9</td>
<td>480.4</td>
<td><a href="https://huggingface.co/ahatamiz/FasterViT/resolve/main/fastervit_4_21k_768_w48.pth.tar">model</a></td>
</tr>
</table>
### Robustness (ImageNet-A - ImageNet-R - ImageNet-V2)
All models use `crop_pct=0.875`. Results are obtained by running inference on ImageNet-1K pretrained models without finetuning.
<table>
<tr>
<th>Name</th>
<th>A-Acc@1(%)</th>
<th>A-Acc@5(%)</th>
<th>R-Acc@1(%)</th>
<th>R-Acc@5(%)</th>
<th>V2-Acc@1(%)</th>
<th>V2-Acc@5(%)</th>
</tr>
<tr>
<td>FasterViT-0</td>
<td>23.9</td>
<td>57.6</td>
<td>45.9</td>
<td>60.4</td>
<td>70.9</td>
<td>90.0</td>
</tr>
<tr>
<td>FasterViT-1</td>
<td>31.2</td>
<td>63.3</td>
<td>47.5</td>
<td>61.9</td>
<td>72.6</td>
<td>91.0</td>
</tr>
<tr>
<td>FasterViT-2</td>
<td>38.2</td>
<td>68.9</td>
<td>49.6</td>
<td>63.4</td>
<td>73.7</td>
<td>91.6</td>
</tr>
<tr>
<td>FasterViT-3</td>
<td>44.2</td>
<td>73.0</td>
<td>51.9</td>
<td>65.6</td>
<td>75.0</td>
<td>92.2</td>
</tr>
<tr>
<td>FasterViT-4</td>
<td>49.0</td>
<td>75.4</td>
<td>56.0</td>
<td>69.6</td>
<td>75.7</td>
<td>92.7</td>
</tr>
<tr>
<td>FasterViT-5</td>
<td>52.7</td>
<td>77.6</td>
<td>56.9</td>
<td>70.0</td>
<td>76.0</td>
<td>93.0</td>
</tr>
<tr>
<td>FasterViT-6</td>
<td>53.7</td>
<td>78.4</td>
<td>57.1</td>
<td>70.1</td>
<td>76.1</td>
<td>93.0</td>
</tr>
</table>
A, R and V2 denote ImageNet-A, ImageNet-R and ImageNet-V2 respectively.
## Citation
Please consider citing FasterViT if this repository is useful for your work.
```
@article{hatamizadeh2023fastervit,
title={FasterViT: Fast Vision Transformers with Hierarchical Attention},
author={Hatamizadeh, Ali and Heinrich, Greg and Yin, Hongxu and Tao, Andrew and Alvarez, Jose M and Kautz, Jan and Molchanov, Pavlo},
journal={arXiv preprint arXiv:2306.06189},
year={2023}
}
```
## Licenses
Copyright © 2023, NVIDIA Corporation. All rights reserved.
This work is made available under the NVIDIA Source Code License-NC. Click [here](https://huggingface.co/nvidia/FasterViT/blob/main/LICENSE) to view a copy of this license.
For license information regarding the timm repository, please refer to its [repository](https://github.com/rwightman/pytorch-image-models).
For license information regarding the ImageNet dataset, please see the [ImageNet official website](https://www.image-net.org/).
## Acknowledgement
This repository is built on top of the [timm](https://github.com/huggingface/pytorch-image-models) repository. We thank [Ross Wrightman](https://rwightman.com/) for creating and maintaining this high-quality library.
|
Jackrong/llama-3.2-3B-Chinese-Elite
|
Jackrong
| 2025-08-29T00:26:48Z | 0 | 0 | null |
[
"gguf",
"text-generation",
"zh",
"en",
"dataset:Jackrong/Qwen3-235B-A22B-Instruct-2507-Distilled-chat",
"base_model:meta-llama/Llama-3.2-3B-Instruct",
"base_model:quantized:meta-llama/Llama-3.2-3B-Instruct",
"license:apache-2.0",
"endpoints_compatible",
"region:us",
"conversational"
] |
text-generation
| 2025-08-28T23:41:08Z |
---
license: apache-2.0
datasets:
- Jackrong/Qwen3-235B-A22B-Instruct-2507-Distilled-chat
language:
- zh
- en
base_model:
- meta-llama/Llama-3.2-3B-Instruct
new_version: Jackrong/llama-3.2-3B-Chinese-Elite
pipeline_tag: text-generation
---
---
# 🦙 llama-3.2-3B-Chinese-Elite 🔥
## 🌟 模型亮点
这是一个基于 **Meta-Llama-3.2-3B-Instruct** 的微调模型，使用 **Qwen3-235B 蒸馏数据** + **监督微调 (SFT)** 训练而成。
在实际使用中，我发现它不仅在 **中文对话、输出内容格式、简单推理、科学问题回答、聊天对话、创意写作** 等任务上表现远优于原始 Llama3.2-3B。同时 **基于中文环境的综合能力，远超越基座模型**，而且风格更接近 **2025年最新 Qwen3系列模型**，输出更自然、更贴近人类偏好，也会有表情符号使文章更生动活泼。
🔥 得益于仅 **3B 的参数规模**，该模型在 **响应速度** 上表现非常突出，交互体验 **流畅自然**。
它能够轻松处理 **日常对话、文本总结、翻译、学习资料讲解** 等常见任务，并且在 **资源受限的环境（如轻量 GPU、本地 CPU 或个人电脑）** 下依然能够 **高效运行**。
与此同时，模型支持 **离线部署**，在保障 **数据安全** 的同时大幅 **节省计算与能源开销**，非常适合 **教育、研究以及个人学习** 场景使用。
---
## 🔧 训练详情
- **基座模型**: `meta-llama/Meta-Llama-3.2-3B-Instruct`
- **训练方法**: 监督微调 (SFT)
- **教师模型**: Qwen-3-235B-A22B-Instruct-2507
- **框架**: [Unsloth]官方 Notebook
- **GPU**: 单卡 A100 (40GB)
- **数据量**: ~50k 高质量蒸馏样本
- **量化支持**: q8_0 / gguf
---
## 📊 效果对比 (Before vs After)
> 表格长度有限，对比图片中只展示部分回答内容。实际模型效果可以看底下的部分问题回答的截屏。
### 1️⃣ 数学推理
| 原始 Llama3.2-3B | 微调后模型 |
|---|---|
| **答案**：235.18 千克/小时 ❌<br>（计算逻辑混乱，结果错误） | **答案**：37.53 千克/小时 ✅<br>（逐步推理，验证正确） |
---
### 2️⃣ 雅思词汇讲解 (constitution)
| 原始 Llama3.2-3B | 微调后模型 |
|---|---|
| 简单定义 + 基础例句<br>输出略显单薄 | 系统讲解：定义、词源、政治语境、雅思常见题型、高频搭配<br>➡️ 更“懂考试”，输出能直接用来备考 ✅ |
---
### 3️⃣ 故事生成
| 原始 Llama3.2-3B | 微调后模型 |
|---|---|
| 故事夹杂英文单词（graduated, sad），情节简单，逻辑不完整 ❌ | 讲述心理学家 John Bleck 的哲理故事：情节完整、叙事自然、结尾升华主题 ✅ |
---
### 4️⃣ 学术推理任务（Contradict a Theory）
| 原始 Llama3.2-3B | 微调后模型 |
|---|---|
| 给出“地心说 vs 日心说”的例子，解释地球并非宇宙中心。逻辑清晰，但局限在单一实例，缺乏方法论与抽象总结。❌ | 输出分层次：**科学方法（实验可证伪）→ 哲学视角（矛盾促使修正）→ 总结（可证伪性与理论完善）**。同时举例进化论与牛顿力学，结构清晰、排版美观，内容更丰富。<br><br>✅ 优点：逻辑清楚、学术性强、对齐 Qwen3 风格。（表格长度问题，所以只展示了部分回答摘要） |
### 实测输出展示：
---
## ⚖️ 局限性
- 训练数据量仅 50k，虽然效果明显提升，但对开放领域问题仍可能不足。
- 模型主要优化了 **聊天 / 语言 / 叙事/ ** 场景，专业领域可能不如更大模型强。
- 基座模型限制：Llama-3.2-3B 的基础能力和通用性能相对有限，微调虽能改善表现，但无法突破基座模型本身的上限。数学能力与复杂问题解决能力孱弱。
- 尚未使用 RLHF / DPO，个别输出在“人类偏好对齐”上还有限。
---
## 🙏 致谢
- **Meta** 提供的基座模型 Llama3.2-3B
- **Qwen 团队** 提供的强大师模型 Qwen-3-235B
- **Unsloth** 高效的微调工具链
---
## 📥 下载 & 试用
👉 这是一个轻量、实用、推理速度快的 **中英双语小模型**。
它继承了 Qwen3 的风格，又保持了 3B 模型的高效，欢迎大家下载体验，并在社区反馈效果！
|
Dejiat/blockassist-bc-savage_unseen_bobcat_1756427157
|
Dejiat
| 2025-08-29T00:26:21Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"savage unseen bobcat",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T00:26:19Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- savage unseen bobcat
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
yinita/mg-ppo-4o-4b-mix-0828-v1-200step
|
yinita
| 2025-08-29T00:25:11Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"qwen3",
"text-generation",
"conversational",
"arxiv:1910.09700",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-08-29T00:24:33Z |
---
library_name: transformers
tags: []
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
|
Rajat1327/Qwen2.5-Coder-7B-1E_LoRA
|
Rajat1327
| 2025-08-29T00:25:09Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"text-generation-inference",
"unsloth",
"qwen2",
"trl",
"en",
"base_model:unsloth/Qwen2.5-Coder-7B-Instruct-bnb-4bit",
"base_model:finetune:unsloth/Qwen2.5-Coder-7B-Instruct-bnb-4bit",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] | null | 2025-08-29T00:24:46Z |
---
base_model: unsloth/Qwen2.5-Coder-7B-Instruct-bnb-4bit
tags:
- text-generation-inference
- transformers
- unsloth
- qwen2
- trl
license: apache-2.0
language:
- en
---
# Uploaded model
- **Developed by:** Rajat1327
- **License:** apache-2.0
- **Finetuned from model :** unsloth/Qwen2.5-Coder-7B-Instruct-bnb-4bit
This qwen2 model was trained 2x faster with [Unsloth](https://github.com/unslothai/unsloth) and Huggingface's TRL library.
[<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png" width="200"/>](https://github.com/unslothai/unsloth)
|
mano-ktk-lake-video-viral/New.full.videos.mano.ktk.lake.viral.video.Official.Tutorial
|
mano-ktk-lake-video-viral
| 2025-08-29T00:24:34Z | 0 | 0 | null |
[
"region:us"
] | null | 2025-08-28T09:29:31Z |
<animated-image data-catalyst=""><a href="https://tinyurl.com/5ye5v3bc?dfhgKasbonStudiosdfg" rel="nofollow" data-target="animated-image.originalLink"><img src="https://static.wixstatic.com/media/b249f9_adac8f70fb3f45b88691696c77de18f3~mv2.gif" alt="Foo" data-canonical-src="https://static.wixstatic.com/media/b249f9_adac8f70fb3f45b88691696c77de18f3~mv2.gif" style="max-width: 100%; display: inline-block;" data-target="animated-image.originalImage"></a>
|
yinita/mg-ppo-4o-4b-0828-mix-v1-382step
|
yinita
| 2025-08-29T00:24:09Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"qwen3",
"text-generation",
"conversational",
"arxiv:1910.09700",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-08-29T00:23:28Z |
---
library_name: transformers
tags: []
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
|
weathermanj/NVIDIA-Nemotron-Nano-9B-v2-gguf
|
weathermanj
| 2025-08-29T00:20:12Z | 0 | 1 |
llama.cpp
|
[
"llama.cpp",
"gguf",
"nemotron_h",
"text-generation",
"quantized",
"nvidia",
"nemotron",
"mamba2",
"transformer",
"en",
"base_model:nvidia/NVIDIA-Nemotron-Nano-9B-v2",
"base_model:quantized:nvidia/NVIDIA-Nemotron-Nano-9B-v2",
"license:other",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-08-28T19:28:08Z |
---
tags:
- gguf
- llama.cpp
- text-generation
- quantized
- nvidia
- nemotron
- mamba2
- transformer
language:
- en
license: other
license_name: nvidia-open-model-license
license_link: https://www.nvidia.com/en-us/agreements/enterprise-software/nvidia-open-model-license/
base_model: nvidia/NVIDIA-Nemotron-Nano-9B-v2
library_name: llama.cpp
pipeline_tag: text-generation
model_type: nemotron_h
quantized: true
quantization_type: gguf
quantization_config:
quantized: true
format: gguf
variants:
- filename: NVIDIA-Nemotron-Nano-9B-v2-gguf-Q2_K.gguf
size: 4.7GB
bits_per_weight: "~2.0"
description: "2-bit K-quantization, maximum compression"
- filename: NVIDIA-Nemotron-Nano-9B-v2-gguf-Q8_0.gguf
size: 8.9GB
bits_per_weight: "~8.0"
description: "Near-lossless, reference quality"
- filename: NVIDIA-Nemotron-Nano-9B-v2-gguf-Q6_K.gguf
size: 8.6GB
bits_per_weight: "~6.0"
description: "High quality, recommended"
- filename: NVIDIA-Nemotron-Nano-9B-v2-gguf-Q5_K_M.gguf
size: 6.6GB
bits_per_weight: "~5.0"
description: "Good quality, balanced"
- filename: NVIDIA-Nemotron-Nano-9B-v2-gguf-Q4_K_M.gguf
size: 6.1GB
bits_per_weight: "~4.0"
description: "Standard choice, good compression"
- filename: NVIDIA-Nemotron-Nano-9B-v2-gguf-Q4_1.gguf
size: 5.5GB
bits_per_weight: "~4.0"
description: "Legacy 4-bit (Q4_1), slightly better quality than Q4_0"
- filename: NVIDIA-Nemotron-Nano-9B-v2-gguf-Q4_0.gguf
size: 5.0GB
bits_per_weight: "~4.0"
description: "Legacy 4-bit (Q4_0), smaller, lower quality"
- filename: NVIDIA-Nemotron-Nano-9B-v2-gguf-Q4_K_S.gguf
size: "~5.8GB"
bits_per_weight: "~4.0"
description: "4-bit K (small), smaller than Q4_K_M"
- filename: NVIDIA-Nemotron-Nano-9B-v2-gguf-IQ4_XS.gguf
size: 5.0GB
bits_per_weight: "4.25"
description: "Integer quantization, excellent compression"
- filename: NVIDIA-Nemotron-Nano-9B-v2-gguf-IQ3_M.gguf
size: 4.9GB
bits_per_weight: "3.66"
description: "Ultra-small, mobile/edge"
---
# NVIDIA-Nemotron-Nano-9B-v2-gguf
GGUF quantizations of NVIDIA’s [NVIDIA-Nemotron-Nano-9B-v2](https://huggingface.co/nvidia/NVIDIA-Nemotron-Nano-9B-v2). These files target llama.cpp-compatible runtimes.
## Available Models
| Model | Size | Bits/Weight | Description |
|-------|------|-------------|-------------|
| `NVIDIA-Nemotron-Nano-9B-v2-gguf-Q8_0.gguf` | 8.9GB | ~8.0 | Near-lossless, reference quality |
| `NVIDIA-Nemotron-Nano-9B-v2-gguf-Q6_K.gguf` | 8.6GB | ~6.0 | High quality, recommended for most users |
| `NVIDIA-Nemotron-Nano-9B-v2-gguf-Q5_K_M.gguf` | 6.6GB | ~5.0 | Good quality, balanced |
| `NVIDIA-Nemotron-Nano-9B-v2-gguf-Q4_K_M.gguf` | 6.1GB | ~4.0 | Standard choice, good compression |
| `NVIDIA-Nemotron-Nano-9B-v2-gguf-Q4_1.gguf` | 5.5GB | ~4.0 | Legacy 4-bit (Q4_1), better than Q4_0 |
| `NVIDIA-Nemotron-Nano-9B-v2-gguf-Q4_0.gguf` | 5.0GB | ~4.0 | Legacy 4-bit (Q4_0), smaller |
| `NVIDIA-Nemotron-Nano-9B-v2-gguf-IQ4_XS.gguf` | 5.0GB | 4.25 | Integer quantization, excellent compression |
| `NVIDIA-Nemotron-Nano-9B-v2-gguf-IQ3_M.gguf` | 4.9GB | 3.66 | Ultra-small, mobile/edge deployment |
| `NVIDIA-Nemotron-Nano-9B-v2-gguf-Q4_K_S.gguf` | 5.8GB | ~4.0 | 4-bit K (small), smaller than Q4_K_M |
| `NVIDIA-Nemotron-Nano-9B-v2-gguf-Q2_K.gguf` | 4.7GB | ~2.0 | 2-bit K, maximum compression |
| `NVIDIA-Nemotron-Nano-9B-v2-gguf-f16.gguf` | 17GB | 16.0 | Full precision reference (optional) |
## Usage
- Download a quantization
- `huggingface-cli download weathermanj/NVIDIA-Nemotron-Nano-9B-v2-gguf NVIDIA-Nemotron-Nano-9B-v2-gguf-Q4_K_M.gguf --local-dir ./`
- Run with llama.cpp
- `./llama-server -m NVIDIA-Nemotron-Nano-9B-v2-gguf-Q4_K_M.gguf -c 4096`
## Performance (tokens/s)
CPU vs CUDA vs CUDA+FlashAttn on a 24GB RTX 3090, n_predict=64, temp=0.7, top_p=0.95.
| Model | CPU Factoid | CPU Code | CPU Reasoning | CUDA Factoid | CUDA Code | CUDA Reasoning | CUDA+FA Factoid | CUDA+FA Code | CUDA+FA Reasoning |
|--------|------------:|---------:|--------------:|-------------:|----------:|---------------:|----------------:|-------------:|------------------:|
| IQ3_M | 10.96 | 9.83 | 9.84 | 59.51 | 48.83 | 51.22 | 49.46 | 51.48 | 51.54 |
| Q4_K_M | 8.59 | 8.03 | 8.02 | 48.28 | 48.72 | 48.70 | 53.48 | 48.73 | 47.97 |
| Q5_K_M | 7.54 | 7.54 | 7.52 | 49.09 | 46.00 | 46.87 | 51.25 | 50.58 | 47.00 |
| Q6_K | 6.65 | 6.19 | 5.89 | 52.77 | 41.84 | 42.06 | 47.59 | 41.48 | 42.85 |
| Q8_0 | 6.95 | 5.79 | 5.93 | 45.99 | 40.81 | 41.51 | 48.32 | 41.21 | 41.54 |
Notes:
- IQ3_M is fastest on this setup; Q4_K_M offers stronger quality with close speed.
- Flash Attention helps variably; larger micro-batches (e.g., `--ubatch-size 1024`) can improve throughput.
## Notes
- Base model: [nvidia/NVIDIA-Nemotron-Nano-9B-v2](https://huggingface.co/nvidia/NVIDIA-Nemotron-Nano-9B-v2)
- These are GGUF files suitable for llama.cpp and compatible backends.
- Choose a quantization based on your resource/quality needs (see table).
## License
- NVIDIA Open Model License: https://www.nvidia.com/en-us/agreements/enterprise-software/nvidia-open-model-license/
|
gsjang/ko-llama-3-korean-bllossom-8b-x-meta-llama-3-8b-instruct-dare_ties-50_50
|
gsjang
| 2025-08-29T00:20:08Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"llama",
"text-generation",
"mergekit",
"merge",
"conversational",
"arxiv:2311.03099",
"base_model:MLP-KTLim/llama-3-Korean-Bllossom-8B",
"base_model:merge:MLP-KTLim/llama-3-Korean-Bllossom-8B",
"base_model:meta-llama/Meta-Llama-3-8B-Instruct",
"base_model:merge:meta-llama/Meta-Llama-3-8B-Instruct",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-08-29T00:17:02Z |
---
base_model:
- MLP-KTLim/llama-3-Korean-Bllossom-8B
- meta-llama/Meta-Llama-3-8B-Instruct
library_name: transformers
tags:
- mergekit
- merge
---
# ko-llama-3-korean-bllossom-8b-x-meta-llama-3-8b-instruct-dare_ties-50_50
This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).
## Merge Details
### Merge Method
This model was merged using the [DARE TIES](https://arxiv.org/abs/2311.03099) merge method using [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) as a base.
### Models Merged
The following models were included in the merge:
* [MLP-KTLim/llama-3-Korean-Bllossom-8B](https://huggingface.co/MLP-KTLim/llama-3-Korean-Bllossom-8B)
### Configuration
The following YAML configuration was used to produce this model:
```yaml
merge_method: dare_ties
models:
- model: MLP-KTLim/llama-3-Korean-Bllossom-8B
parameters:
weight: 0.5
density: 0.4
- model: meta-llama/Meta-Llama-3-8B-Instruct
parameters:
weight: 0.5
density: 0.4
parameters:
int8_mask: true
dtype: bfloat16
tokenizer:
source: union
base_model: meta-llama/Meta-Llama-3-8B-Instruct
write_readme: README.md
```
|
maboranomet/blockassist-bc-lumbering_soft_macaw_1756426721
|
maboranomet
| 2025-08-29T00:19:29Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"lumbering soft macaw",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T00:19:06Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- lumbering soft macaw
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
deadman44/Wan2.2_Workflow_for_myxx_series_LoRA
|
deadman44
| 2025-08-29T00:16:16Z | 0 | 11 | null |
[
"text-to-image",
"t2i",
"wan video",
"safetensors",
"text-to-video",
"en",
"license:apache-2.0",
"region:us"
] |
text-to-video
| 2025-08-04T13:08:07Z |
---
license: apache-2.0
pipeline_tag: text-to-video
language:
- en
tags:
- text-to-image
- t2i
- wan video
- safetensors
---
<style>
.title{
font-size: 2.5em;
letter-spacing: 0.01em;
padding: 0.5em 0;
}
.thumbwidth{
max-width: 180px;
}
.font_red{
color:red;
}
.font_blue{
color:blue;
}
.font_grey{
color: #aaaaaa;
}
</style>
# Workflow for myxx series LoRA
<br />
- [myxx Lora](https://huggingface.co/deadman44/WAN_T2i_LoRA)
<br />
# Recent
- Fix [I2v Triple High Workflow](#i2v_t) (wrong node): 2025-08-29<br />
- Add [T2v Triple High Workflow](#t2v_t) / I2v Triple High (experimental): 2025-08-21<br />
- flf2v has been integrated into I2v_better: 2025-08-16<br />
- Added Video Extend function to [I2v_better Workflow](#i2v): 2025-08-15<br />
- Implement "my series LoRA select switch" on each node: 2025-08-14<br />
- Add Interpolation Workflow / Update some nodes: 2025-08-13<br />
---
<a id="workflow"></a>
<h1 class="title">
<span>Sample workflow</span>
</h1>
## - Wan2.2
- [T2i / T2v Workflow](#t2i)
- [I2v Workflow](#i2v)
- [Highres Upscale Workflow](#highres)
- [Speed Mult Workflow](#speedmult)
- [Concat Workflow](#concat)
- [Interpolation Workflow](#inter)
- [Reccomended Models](#reccomended)
---
<a id="t2i"></a>
# T2i / T2v - Better Quality Workflow
- <span class="font_blue">Recommended for T2i</span>
<br/><br/>
[Download: T2i/T2v better](https://huggingface.co/deadman44/Wan2.2_Workflow_for_myxx_series_LoRA/raw/main/Wan2.2_T2i_better.json)
<br/><br/>
<div style="display: flex; flex-direction: column; align-items: flex-start; gap: 12px; margin-bottom: 32px;">
<strong>T2i</strong>
<a href="https://img1.pixhost.to/images/7710/628818240_20250804115623_t2i_00001_.jpg" target="_blank">
<img src="https://t1.pixhost.to/thumbs/7710/628818240_20250804115623_t2i_00001_.jpg"
alt="T2I"
style="width: 240px; height: auto; object-fit: contain; border: 1px solid #ccc;">
</a>
<pre style="background: #f4f4f4; padding: 12px; border-radius: 6px; overflow-x: auto;">
<code>15yo, myjc, japanese, photorealistic,
An upper body portrait of a girl in white dress at library.
She has a long black hair with bangs.
She is holding books, and looking at camera.</code>
</pre>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-start; gap: 12px;">
<strong>T2v</strong>
<video controls loop style="width: 480px; height: auto; object-fit: contain; border: 1px solid #ccc;">
<source src="https://files.catbox.moe/k71i17.mp4" type="video/mp4">
Your browser cannot play the video.
</video>
<pre style="background: #f4f4f4; padding: 12px; border-radius: 6px; overflow-x: auto;">
<code>A girl taking a selfie. The camera circles around the girl.
She has black hair, a ponytail, and is wearing a school uniform.
The background is a daytime classroom with several students.</code>
</pre>
</div>
<br/>
---
# T2i / T2v (kijai node)
- <span class="font_blue">Recommended for T2v</span>
<br/><br/>
[Download: T2i/T2v (kijai)](https://huggingface.co/deadman44/Wan2.2_Workflow_for_myxx_series_LoRA/raw/main/Wan2.2_T2i.json)
<br/><br/>
<div style="display: flex; flex-direction: column; align-items: flex-start; gap: 24px;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<strong>T2i</strong>
<a href="https://img1.pixhost.to/images/7844/630528571_20250810092108_t2i_00001_.png" target="_blank">
<img src="https://t1.pixhost.to/thumbs/7844/630528571_20250810092108_t2i_00001_.png"
alt="kijai T2I"
style="width: 240px; height: auto; object-fit: contain; border: 1px solid #ccc;">
</a>
</div>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<strong>T2v</strong>
<video controls loop
style="width: 320px; height: auto; object-fit: contain; border: 1px solid #ccc;">
<source src="https://files.catbox.moe/m7lxqs.mp4" type="video/mp4">
Your browser cannot play the video.
</video>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<pre style="background: #f4f4f4; padding: 12px; border-radius: 6px; overflow-x: auto;">
<code>15yo, myjc, japanese, photorealistic,
portrait of a girl walking at street.
She has a long black ponytail with side bangs.
She is wearing short sleeves school uniform.</code></pre>
</div>
<br/>
---
<a id="t2v_t"></a>
# T2v Triple High
- <span class="font_blue">better motion quality, </span><span class="font_red">but very slowly</span>
<br/><br/>
[Download: T2v Triple High](https://huggingface.co/deadman44/Wan2.2_Workflow_for_myxx_series_LoRA/raw/main/Wan2.2_T2v_Triple_High.json)
<br/><br/>
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<strong>T2v</strong>
<video controls loop
style="width: 320px; height: auto; object-fit: contain; border: 1px solid #ccc;">
<source src="https://huggingface.co/deadman44/Wan2.2_Workflow_for_myxx_series_LoRA/resolve/main/samples/20250820180348_T2V_00001.mp4" type="video/mp4">
Your browser cannot play the video.
</video>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<pre style="background: #f4f4f4; padding: 12px; border-radius: 6px; overflow-x: auto;">
<code>15yo, myjc, japanese,
The camera approaches a girl in school uniform walking at street of Tokyo.
She has a black twintails with side bangs.
The girl smiles and waves at the camera, then turns around and runs away.
The camera moves away from the girl.</code></pre>
</div>
<br/>
---
<a id="i2v"></a>
# I2v - Better Quality Workflow
- <span class="font_blue">Recommended for I2v / flf2v / Extend</span>
<br/><br/>
[Download: I2v better](https://huggingface.co/deadman44/Wan2.2_Workflow_for_myxx_series_LoRA/raw/main/Wan2.2_I2v_better.json)
<br/><br/>
<div style="display: flex; flex-direction: column; align-items: flex-start; gap: 24px;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<strong>Start Image</strong>
<img src="https://img1.pixhost.to/images/7711/628837767_20250804203455_t2i_00001_.jpg" alt="I2v better" style="max-width: 240px; height: auto; border: 1px solid #ccc;">
</div>
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<video controls loop style="max-width: 320px; height: auto; border: 1px solid #ccc;">
<source src="https://files.catbox.moe/5c8jt9.mp4" type="video/mp4">
Your browser cannot play the video.
</video>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<pre style="background: #f4f4f4; padding: 12px; border-radius: 6px; overflow-x: auto;">
<code>18yo, myjk, japanese,
A nurse is walking down a dark hospital corridor when suddenly the doctors appear with a bouquet of flowers and she jumps up and down, smiling and overjoyed.</code>
</pre>
</div>
</div>
<br/>
### Cut Start Frames
- Reduces discomfort at edges, etc. (e.g., collage images) by removing the initial few frames.<br><br>
<div style="display: flex; flex-direction: column; align-items: flex-start; gap: 24px;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<strong>Collage Image</strong>
<a href="https://img1.pixhost.to/images/7726/629071198_20250805083230_t2i_00001_.jpg" target="_blank">
<img src="https://t1.pixhost.to/thumbs/7726/629071198_20250805083230_t2i_00001_.jpg"
alt="T2I"
style="width: 240px; height: auto; object-fit: contain; border: 1px solid #ccc;"></a>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<video controls loop style="max-width: 480px; height: auto; border: 1px solid #ccc;">
<source src="https://files.catbox.moe/xm8s0i.mp4" type="video/mp4">
Your browser cannot play the video.
</video>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-start;">
Cut the first 4 frames.
</div>
<br/>
### Video Extend
- Overlapping video frames ensures consistent extension<br/>
- Load Image a "webp video file".<br/>
- [Wan2.2-Fun-A14B-InP-GGUF](https://huggingface.co/QuantStack/Wan2.2-Fun-A14B-InP-GGUF/tree/main) required<br/>
<br/>
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<strong>Last Image + I2V</strong>
- T2V 5sec (81frames) + LastFrame I2V 5sec (81frames)
<span class="font_red">- The speed changes when switching videos.</span>
<video controls loop
style="width: 480px; height: auto; object-fit: contain; border: 1px solid #ccc;">
<source src="https://huggingface.co/deadman44/Wan2.2_Workflow_for_myxx_series_LoRA/resolve/main/samples/20250815180535_Concat_00002.mp4" type="video/mp4">
Your browser cannot play the video.
</video>
</div>
<br/>
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<strong>Video Extend</strong>
- (Base) T2V 5sec (81frames): 12frames overrap 10sec
<span class="font_blue">- The transition is smooth.</span>
<video controls loop
style="width: 480px; height: auto; object-fit: contain; border: 1px solid #ccc;">
<source src="https://huggingface.co/deadman44/Wan2.2_Workflow_for_myxx_series_LoRA/resolve/main/samples/20250815185414_I2V_00001.mp4" type="video/mp4">
Your browser cannot play the video.
</video>
</div>
<br/>
<strong>I2V StartEnd (flf2v)</strong>
- Start Image -> End Image<br>
<br/><br/>
<div style="display: flex; flex-direction: row; gap: 24px; align-items: flex-start; margin-bottom: 24px;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<strong>Start Image</strong>
<img src="https://img1.pixhost.to/images/7711/628826893_20250804200958_t2i_00001_.jpg" alt="start" style="max-width: 320px; height: auto; border: 1px solid #ccc;">
</div>
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<strong>End Image</strong>
<img src="https://img1.pixhost.to/images/7711/628826895_20250804190007_t2i_00001_.jpg" alt="end" style="max-width: 320px; height: auto; border: 1px solid #ccc;">
</div>
</div>
<div style="margin-bottom: 24px;">
<video controls loop style="max-width: 480px; height: auto; border: 1px solid #ccc;">
<source src="https://files.catbox.moe/s911sw.mp4" type="video/mp4">
Your browser cannot play the video.
</video>
</div>
<div>
<pre style="background: #f4f4f4; padding: 12px; border-radius: 6px; overflow-x: auto;">
<code>18yo, myjk, japanese,
Suddenly A girl in a yellow parachute suit lands with her parachute open on a street crowded with people.
She has a black ponytail with side bangs.
People cheer and applaud.</code>
</pre>
</div>
<strong>Only End Image</strong>
- empty start -> End Image<br>
<br/><br/>
<div style="display: flex; flex-direction: column; align-items: flex-start; gap: 24px;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<strong>End Image</strong>
<img src="https://img1.pixhost.to/images/7711/628839765_20250804205608_t2i_00001_.jpg" alt="randomstartend better" style="max-width: 320px; height: auto; border: 1px solid #ccc;">
</div>
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<video controls loop style="max-width: 480px; height: auto; border: 1px solid #ccc;">
<source src="https://files.catbox.moe/kal8vn.mp4" type="video/mp4">
Your browser cannot play the video.
</video>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<pre style="background: #f4f4f4; padding: 12px; border-radius: 6px; overflow-x: auto;">
<code>18yo, myjk, japanese,
A girl is running.
She wears black twin-tail hair and a school uniform.
In the background is a row of trees and a blue sky, her hair blowing in the wind.</code>
</pre>
</div>
<br/>
# I2v / StartEnd (kijai node)
- <span class="font_blue">Recommended for Last Image + I2v</span>
<br/><br/>
[Download: I2v (kijai)](https://huggingface.co/deadman44/Wan2.2_Workflow_for_myxx_series_LoRA/raw/main/Wan2.2_I2v.json)
<br/><br/>
<div style="display: flex; flex-direction: column; align-items: flex-start; gap: 24px;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<strong>Start Image</strong>
<a href="https://img1.pixhost.to/images/7846/630539614_20250810103859_t2i_00001_.png" target="_blank">
<img src="https://t1.pixhost.to/thumbs/7846/630539614_20250810103859_t2i_00001_.png"
alt="T2I"
style="width: 240px; height: auto; object-fit: contain; border: 1px solid #ccc;"></a>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<video controls loop style="max-width: 480px; height: auto; border: 1px solid #ccc;">
<source src="https://files.catbox.moe/1puol2.mp4" type="video/mp4">
Your browser cannot play the video.
</video>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<pre style="background: #f4f4f4; padding: 12px; border-radius: 6px; overflow-x: auto;">
<code>15yo, myjc, japanese, photorealistic,
Suddenly few students appear.
The girl is dancing with students.</code></pre>
</div>
</div>
<br/>
<a id="i2v_t"></a>
# I2v / flf2v / Extend Triple High
- <span class="font_blue">better motion quality, </span><span class="font_red">but very slowly</span>
<br/><br/>
[Download: I2v Triple High](https://huggingface.co/deadman44/Wan2.2_Workflow_for_myxx_series_LoRA/raw/main/Wan2.2_I2v_Triple_High.json)
<br/><br/>
---
# Util
<a id="highres"></a>
# Highres Upscale Workflow
- High-resolution upscaling of load images and videos<br>
- image file and movie (webp)<br>
- Low denoise shows less change, high denoise shows more change<br>
<br/><br/>
[Download: Highres_Upscale](https://huggingface.co/deadman44/Wan2.2_Workflow_for_myxx_series_LoRA/raw/main/Wan2.2_Highres_Upscale.json)
<br/><br/>
<div style="display: flex; flex-direction: column; align-items: flex-start; gap: 24px;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<strong>Base movie</strong>
<video controls loop style="max-width: 480px; height: auto; border: 1px solid #ccc;">
<source src="https://files.catbox.moe/55jo26.mp4" type="video/mp4">
Your browser cannot play the video.
</video>
<div style="margin-top: 8px;">512 × 768</div>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<video controls loop style="max-width: 480px; height: auto; border: 1px solid #ccc;">
<source src="https://files.catbox.moe/mdsbrh.mp4" type="video/mp4">
Your browser cannot play the video.
</video>
<div style="margin-top: 8px;">768 × 1024（1.5x）・denoise: 0.3</div>
</div>
</div>
---
<a id="speedmult"></a>
# Speed Mult Workflow
- Change video speed
- webp movie -> mp4
<br/><br/>
[Download: Speed_Mult](https://huggingface.co/deadman44/Wan2.2_Workflow_for_myxx_series_LoRA/raw/main/Wan2.2_Speed_Mult.json)
<br/><br/>
---
<a id="concat"></a>
# Concat Workflow
- Connecting videos
- <span class="font_red">(webp) Videos with different fps don't work well together.</span>
<br/><br/>
[Download: Concat](https://huggingface.co/deadman44/Wan2.2_Workflow_for_myxx_series_LoRA/raw/main/Wan2.2_Concat.json)
<br/><br/>
image -> Highres (webp) -> Speed_Mult (mp4) -> Concat (mp4)
<br/><br/>
---
<a id="inter"></a>
# Interpolation Workflow
- Frame Interpolation videos
- webp / mp4 -> fps:60
<br/><br/>
[Download: Interpolation](https://huggingface.co/deadman44/Wan2.2_Workflow_for_myxx_series_LoRA/raw/main/Wan2.2_Interpolation.json)
<br/><br/>
image -> Highres (webp) -> Speed_Mult (mp4) -> Concat (mp4)
<br/><br/>
<br/><br/>
---
<a id="reccomended"></a>
# Recommended model
### - Wan2.2
Q5, Q4, Q3... will further omit vram<br>
[T2i / T2v]<br>
- [wan2.2_t2v_high_noise_14B_Q8_0.gguf](https://huggingface.co/bullerwins/Wan2.2-T2V-A14B-GGUF/tree/main)<br>
- [wan2.2_t2v_low_noise_14B_Q8_0.gguf](https://huggingface.co/bullerwins/Wan2.2-T2V-A14B-GGUF/tree/main)<br>
[I2v]<br>
- [wan2.2_i2v_high_noise_14B_Q5_K_M.gguf](https://huggingface.co/bullerwins/Wan2.2-I2V-A14B-GGUF/tree/main)<br>
- [wan2.2_i2v_low_noise_14B_Q5_K_M.gguf](https://huggingface.co/bullerwins/Wan2.2-I2V-A14B-GGUF/tree/main)<br>
[I2v Video Extend]<br>
- [Wan2.2-Fun-A14B-InP_HighNoise-Q8_0.gguf](https://huggingface.co/QuantStack/Wan2.2-Fun-A14B-InP-GGUF/tree/main/HighNoise)<br>
- [Wan2.2-Fun-A14B-InP_LowNoise-Q8_0.gguf](https://huggingface.co/QuantStack/Wan2.2-Fun-A14B-InP-GGUF/tree/main/LowNoise)<br>
[etc]<br>
- [umt5_xxl_fp8_e4m3fn_scaled](https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/blob/main/split_files/text_encoders/umt5_xxl_fp8_e4m3fn_scaled.safetensors)<br>
- [umt5-xxl-enc-bf16.safetensors](https://huggingface.co/Kijai/WanVideo_comfy/blob/main/umt5-xxl-enc-bf16.safetensors)<br>
- [Wan2_1_VAE_bf16.safetensors](https://huggingface.co/Kijai/WanVideo_comfy/blob/main/Wan2_1_VAE_bf16.safetensors)<br>
[LoRA]<br>
- [Wan2.2-T/I2V-A14B-4steps-lora-rank64-Seko-Vx](https://huggingface.co/lightx2v/Wan2.2-Lightning/tree/main)
- [Wan21_T2V_14B_lightx2v_cfg_step_distill_lora_rank128.safetensors](https://huggingface.co/Kijai/WanVideo_comfy/tree/main/Lightx2v)<br>
- [Wan2.1_T2V_14B_FusionX_LoRA.safetensors](https://huggingface.co/Thelocallab/WAN-2.1-loras/blob/main/Wan2.1_T2V_14B_FusionX_LoRA.safetensors)<br>
- [WAN2.2-LowNoise_SmartphoneSnapshotPhotoReality_v2_by-AI_Characters.safetensors](https://civitai.com/models/1834338?modelVersionId=2075810)<br>
[upscale]<br>
- [4x-ClearRealityV1.pth](https://openmodeldb.info/models/4x-ClearRealityV1)<br>
---
|
bah63843/blockassist-bc-plump_fast_antelope_1756426503
|
bah63843
| 2025-08-29T00:15:56Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"plump fast antelope",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T00:15:46Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- plump fast antelope
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
kamizane/Llama-3.2-1B-Instruct-bnb-FINETUNING
|
kamizane
| 2025-08-29T00:15:16Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"llama",
"text-generation",
"conversational",
"arxiv:1910.09700",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"4-bit",
"bitsandbytes",
"region:us"
] |
text-generation
| 2025-08-28T18:34:49Z |
---
library_name: transformers
tags: []
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
|
Dejiat/blockassist-bc-savage_unseen_bobcat_1756426486
|
Dejiat
| 2025-08-29T00:15:09Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"savage unseen bobcat",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T00:15:07Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- savage unseen bobcat
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
gsjang/ko-llama-3-korean-bllossom-8b-x-meta-llama-3-8b-instruct-ties-50_50
|
gsjang
| 2025-08-29T00:11:44Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"llama",
"text-generation",
"mergekit",
"merge",
"conversational",
"arxiv:2306.01708",
"base_model:MLP-KTLim/llama-3-Korean-Bllossom-8B",
"base_model:merge:MLP-KTLim/llama-3-Korean-Bllossom-8B",
"base_model:meta-llama/Meta-Llama-3-8B-Instruct",
"base_model:merge:meta-llama/Meta-Llama-3-8B-Instruct",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-08-29T00:08:45Z |
---
base_model:
- meta-llama/Meta-Llama-3-8B-Instruct
- MLP-KTLim/llama-3-Korean-Bllossom-8B
library_name: transformers
tags:
- mergekit
- merge
---
# ko-llama-3-korean-bllossom-8b-x-meta-llama-3-8b-instruct-ties-50_50
This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).
## Merge Details
### Merge Method
This model was merged using the [TIES](https://arxiv.org/abs/2306.01708) merge method using [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) as a base.
### Models Merged
The following models were included in the merge:
* [MLP-KTLim/llama-3-Korean-Bllossom-8B](https://huggingface.co/MLP-KTLim/llama-3-Korean-Bllossom-8B)
### Configuration
The following YAML configuration was used to produce this model:
```yaml
merge_method: ties
models:
- model: MLP-KTLim/llama-3-Korean-Bllossom-8B
parameters:
weight: 0.5
density: 0.5
- model: meta-llama/Meta-Llama-3-8B-Instruct
parameters:
weight: 0.5
density: 0.5
parameters:
normalize: true
int8_mask: true
dtype: bfloat16
tokenizer:
source: union
base_model: meta-llama/Meta-Llama-3-8B-Instruct
write_readme: README.md
```
|
Dejiat/blockassist-bc-savage_unseen_bobcat_1756426194
|
Dejiat
| 2025-08-29T00:10:22Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"savage unseen bobcat",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T00:10:18Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- savage unseen bobcat
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
Datasmartly/smartly-tourjouman-tamazight-lora-multi
|
Datasmartly
| 2025-08-29T00:06:37Z | 0 | 0 |
peft
|
[
"peft",
"safetensors",
"trl",
"sft",
"generated_from_trainer",
"base_model:google/gemma-2-9b-it",
"base_model:adapter:google/gemma-2-9b-it",
"license:gemma",
"region:us"
] | null | 2025-08-27T18:15:07Z |
---
base_model: google/gemma-2-9b-it
library_name: peft
license: gemma
tags:
- trl
- sft
- generated_from_trainer
model-index:
- name: smartly-tourjouman-tamazight-lora-multi
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. -->
# smartly-tourjouman-tamazight-lora-multi
This model is a fine-tuned version of [google/gemma-2-9b-it](https://huggingface.co/google/gemma-2-9b-it) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: nan
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0002
- train_batch_size: 2
- eval_batch_size: 1
- seed: 42
- distributed_type: multi-GPU
- num_devices: 8
- gradient_accumulation_steps: 8
- total_train_batch_size: 128
- total_eval_batch_size: 8
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.03
- num_epochs: 3
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:------:|:----:|:---------------:|
| 1.3253 | 0.5006 | 218 | nan |
| 0.8912 | 1.0011 | 436 | nan |
| 0.4228 | 1.5017 | 654 | nan |
### Framework versions
- PEFT 0.12.0
- Transformers 4.44.2
- Pytorch 2.2.0
- Datasets 2.21.0
- Tokenizers 0.19.1
|
liukevin666/blockassist-bc-yawning_striped_cassowary_1756425930
|
liukevin666
| 2025-08-29T00:06:35Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"yawning striped cassowary",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T00:06:29Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- yawning striped cassowary
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
bah63843/blockassist-bc-plump_fast_antelope_1756425883
|
bah63843
| 2025-08-29T00:06:12Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"plump fast antelope",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T00:06:03Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- plump fast antelope
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
aleebaster/blockassist-bc-sly_eager_boar_1756424089
|
aleebaster
| 2025-08-29T00:02:13Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"sly eager boar",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T00:02:04Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- sly eager boar
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
koloni/blockassist-bc-deadly_graceful_stingray_1756424108
|
koloni
| 2025-08-29T00:01:22Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"deadly graceful stingray",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T00:01:17Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- deadly graceful stingray
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
onderyazici/blockassist-bc-lithe_pensive_falcon_1756424606
|
onderyazici
| 2025-08-28T23:59:55Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"lithe pensive falcon",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-28T23:59:46Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- lithe pensive falcon
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
gsjang/ko-llama-3-korean-bllossom-8b-x-meta-llama-3-8b-instruct-multislerp-50_50
|
gsjang
| 2025-08-28T23:59:36Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"llama",
"text-generation",
"mergekit",
"merge",
"conversational",
"base_model:MLP-KTLim/llama-3-Korean-Bllossom-8B",
"base_model:merge:MLP-KTLim/llama-3-Korean-Bllossom-8B",
"base_model:meta-llama/Meta-Llama-3-8B-Instruct",
"base_model:merge:meta-llama/Meta-Llama-3-8B-Instruct",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-08-28T23:56:40Z |
---
base_model:
- MLP-KTLim/llama-3-Korean-Bllossom-8B
- meta-llama/Meta-Llama-3-8B-Instruct
library_name: transformers
tags:
- mergekit
- merge
---
# ko-llama-3-korean-bllossom-8b-x-meta-llama-3-8b-instruct-multislerp-50_50
This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).
## Merge Details
### Merge Method
This model was merged using the [Multi-SLERP](https://goddard.blog/posts/multislerp-wow-what-a-cool-idea) merge method using [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) as a base.
### Models Merged
The following models were included in the merge:
* [MLP-KTLim/llama-3-Korean-Bllossom-8B](https://huggingface.co/MLP-KTLim/llama-3-Korean-Bllossom-8B)
### Configuration
The following YAML configuration was used to produce this model:
```yaml
merge_method: multislerp
models:
- model: MLP-KTLim/llama-3-Korean-Bllossom-8B
parameters:
weight: 0.5
- model: meta-llama/Meta-Llama-3-8B-Instruct
parameters:
weight: 0.5
parameters:
t: 0.5
dtype: bfloat16
tokenizer:
source: union
base_model: meta-llama/Meta-Llama-3-8B-Instruct
write_readme: README.md
```
|
Dejiat/blockassist-bc-savage_unseen_bobcat_1756425458
|
Dejiat
| 2025-08-28T23:58:25Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"savage unseen bobcat",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-28T23:58:22Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- savage unseen bobcat
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
cwayneconnor/blockassist-bc-mute_loud_lynx_1756425213
|
cwayneconnor
| 2025-08-28T23:57:32Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"mute loud lynx",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-28T23:54:58Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- mute loud lynx
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
netotewerp/blockassist-bc-nasty_stubby_weasel_1756425382
|
netotewerp
| 2025-08-28T23:57:13Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"nasty stubby weasel",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-28T23:56:46Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- nasty stubby weasel
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
Thireus/Qwen3-4B-Thinking-2507-THIREUS-IQ2_KS-SPECIAL_SPLIT
|
Thireus
| 2025-08-28T23:56:42Z | 0 | 0 | null |
[
"gguf",
"arxiv:2505.23786",
"license:mit",
"region:us"
] | null | 2025-08-28T23:55:46Z |
---
license: mit
---
# Qwen3-4B-Thinking-2507
## 🤔 What is this [HuggingFace repository](https://huggingface.co/Thireus/Qwen3-4B-Thinking-2507-THIREUS-BF16-SPECIAL_SPLIT/) about?
This repository provides **GGUF-quantized tensors** for the Qwen3-4B-Thinking-2507 model (official repo: https://huggingface.co/Qwen/Qwen3-4B-Thinking-2507). These GGUF shards are designed to be used with **Thireus’ GGUF Tool Suite** (https://gguf.thireus.com), a collection of tools that automatically finds the perplexity-optimal mix of quantizations for any given VRAM and RAM target. With the Tool Suite, you can generate and download custom quantization “recipes” effortlessly.
- 📖 Read more: https://github.com/Thireus/GGUF-Tool-Suite
- 🔍 Example quant mixes: https://github.com/Thireus/GGUF-Tool-Suite/tree/main/recipe_examples
- 🛠️ Create your own recipe: https://colab.research.google.com/github/Thireus/GGUF-Tool-Suite/blob/main/quant_recipe_pipeline.ipynb
- 📂 Browse available quant shards: https://huggingface.co/Thireus/collections
*tl;dr: Expand the details section below*
<details>
```
cd ~
# Make sure to install all ik_llama.cpp compilation dependencies...
apt install python3-dev python3-pip python3-venv python3-wheel python3-setuptools git acl netcat-openbsd cmake # pipx
# Obtain ik_llama's Thireus version - Windows builds available at https://github.com/Thireus/ik_llama.cpp/releases
git clone https://github.com/Thireus/ik_llama.cpp
cd ik_llama.cpp
git pull
# Build ik_llama.cpp
cmake -B build -DGGML_AVX=ON -DGGML_AVX2=ON -DLLAMA_CURL=OFF -DGGML_MAX_CONTEXTS=2048
cmake --build build --config Release -j16
cd ..
# Obtain Thireus' GGUF-Tool-Suite
git clone https://github.com/Thireus/GGUF-Tool-Suite
# Download model quant mix from recipe file:
cd GGUF-Tool-Suite
rm -f download.conf # Make sure to copy the relevant download.conf for the model before running quant_assign.py
cp -f models/Qwen3-4B-Thinking-2507/download.conf . # Use the download.conf of the chosen model
mkdir -p kitchen && cd kitchen
../quant_downloader.sh ../recipe_examples/ik_llama.cpp_recipes/Qwen3-4B-Thinking-2507.ROOT-4.2498bpw-10.9335ppl.1GB-GGUF_0GB-GPU_1GB-CPU.9888e4b_9193781.recipe
# Other recipe examples can be found at https://github.com/Thireus/GGUF-Tool-Suite/tree/main/recipe_examples
# Launch ik_llama's llama-cli:
ulimit -n 9999 # Lifts "too many open files" limitation on Linux
~/ik_llama.cpp/build/bin/llama-server \
-m Qwen3-4B-Thinking-2507-THIREUS-BF16-SPECIAL_TENSOR-00001-of-00399.gguf \
-fa -amb 1024 -ctk q8_0 -c 32768 -ngl 99 \
-b 4096 -ub 4096 --warmup-batch --no-mmap --threads 1 \
--main-gpu 0
```
</details>
---
## ❓ Why does this Tool Suite exist?
1. **Compatibility & Speed** – [unsloth](https://huggingface.co/unsloth)’s dynamic quants may not always work optimally with `ik_llama.cpp`.
2. **Custom Rig Fit** – No off-the-shelf GGUF model perfectly matched my VRAM/RAM setup, so I built a way to tailor models and leverage extra VRAM/RAM to reduce perplexity.
3. **Automated PPL-Optimal Quantization** – To my knowledge, there was no open source flexible, automated method to minimize perplexity for any bits-per-weight (bpw) target—so I created one with excellent results!
---
## 📊 How does it compare to other GGUFs?
Here’s how Qwen3-4B-Thinking-2507 quantized with **Thireus’ GGUF Tool Suite** stacks up against other quantizers (lower perplexity = better at equal or lower bpw):
> _Note: The `recipe_examples` files illustrate good recipes. The Tool Suite computes the optimal ppl/bpw curve for you — just specify your target RAM, VRAM, and quant types, and `quant_assign.py` finds the best mix._
More perplexity/bpw graphs for other supported models: https://github.com/Thireus/GGUF-Tool-Suite/tree/main/ppl_graphs
---
## 🚀 How do I get started?
Check out the [GGUF Tool Suite README](https://github.com/Thireus/GGUF-Tool-Suite) — focus on these sections:
1. ⚠️ **Requirements** – Which `ik_llama.cpp` (or `llama.cpp`) version to use and how to compile.
- Windows binaries (no patching needed) at: https://github.com/Thireus/ik_llama.cpp/releases
2. 📥 **Download Model Shards** – Use `quant_downloader.sh` to fetch GGUF shards from any recipe.
- Recipe examples: https://github.com/Thireus/GGUF-Tool-Suite/tree/main/recipe_examples
3. 🧠 **Run a Downloaded Model** – Sample usage with `llama-cli`.
4. 🛠️ **Generate a Custom Recipe** – Produce recipes tailored to your VRAM/RAM target usage for optimum perplexity.
---
## ✅ Supported Models
Supported models are listed under `models/` in the [Tool Suite Github repo](https://github.com/Thireus/GGUF-Tool-Suite/tree/main/models). Presence of `ppl_results.csv` indicates official support and compatibility with `quant_assign.py`.
---
## 🤷♂️ Will I release baked dynamic quant GGUFs?
No, because I believe in **tailored quantization** for each user’s hardware. If you prefer ready-made shards, you are welcome to merge them via `llama-gguf-split --merge`, or request someone to publish them, or rely on generic GGUF dynamic quants such as [unsloth](https://huggingface.co/unsloth)'s.
Instead, I prefer to share examples of recipes so users can see exactly how they were produced (command included inside these recipe files) and tweak them for their own rigs. The `quant_downloader.sh` script handles automatic fetching and verification of each shard. Note that recipes provided by [Ubergarm](https://huggingface.co/ubergarm) on his model cards are also compatible with `quant_downloader.sh`.
Users who don’t trust the GGUF shards on HuggingFace can also quantize their own by passing recipe lines to `llama-quantize --custom-q` ([see example](https://github.com/Thireus/GGUF-Tool-Suite/blob/main/models/DeepSeek-R1-0528/DeepSeek-R1-0528-THIREUS-ANY-SPECIAL.sh#L482-L486)). Run `llama-quantize --help` to list compatible quants for `quant_assign.py`. This approach is especially useful if you prefer `llama.cpp` over `ik_llama.cpp`.
---
## 📦 What’s in this repository?
- **00001 GGUF header shard** – Contains metadata (tokens, chat template, tensor count, etc.). This metadata can be explored directly from the HuggingFace web interface after clicking on that shard.
- **Tensor shards** – Each shard holds one tensor; see `tensors.map` for names, quant types, sizes, SHA-256 hash, shard IDs, etc.
- **GPG-signed files** – `tensors.map` and header shard are signed with the key in [trusted-keys.asc](https://github.com/Thireus/GGUF-Tool-Suite/blob/main/trusted-keys.asc) for tamper detection.
- **Security note** – Some papers about various ways to attack GGUFs and LLMs are available online, such as https://arxiv.org/abs/2505.23786, and there are also more classic security exploits like CVE-2024-23496 and CVE-2024-25664 through CVE-2024-25668. Only use GGUFs from reputable, trusted authors—or alternatively self-quantize—to avoid potential exploits.
---
## 💡 Pro Tips
You can easily download the BF16 model version to quantize your own shards:
```
mkdir kitchen
echo '.*=bf16' > kitchen/bf16.recipe
cd kitchen
../quant_downloader.sh bf16.recipe
```
Enjoy optimized quantization! 🎉
|
seraphimzzzz/603264
|
seraphimzzzz
| 2025-08-28T23:53:21Z | 0 | 0 | null |
[
"region:us"
] | null | 2025-08-28T23:53:15Z |
[View on Civ Archive](https://civarchive.com/models/615395?modelVersionId=687937)
|
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.