modelId
stringlengths 5
139
| author
stringlengths 2
42
| last_modified
timestamp[us, tz=UTC]date 2020-02-15 11:33:14
2025-08-30 06:27:36
| downloads
int64 0
223M
| likes
int64 0
11.7k
| library_name
stringclasses 527
values | tags
listlengths 1
4.05k
| pipeline_tag
stringclasses 55
values | createdAt
timestamp[us, tz=UTC]date 2022-03-02 23:29:04
2025-08-30 06:27:12
| card
stringlengths 11
1.01M
|
|---|---|---|---|---|---|---|---|---|---|
jd0g/chessgpt-board-probes
|
jd0g
| 2025-08-29T04:00:55Z | 0 | 0 | null |
[
"region:us"
] | null | 2025-08-29T02:17:52Z |
# ChessGPT Board Probes
This repository contains linear probes trained to predict chess piece positions from the internal representations of various chess language models.
## Overview
These probes were trained as part of interpretability research on chess LLMs, investigating how board-state representations develop across different model architectures and layers.
## Models Analyzed
- **Small-16** (512 dim, 16 layers): All layers 0-15
- **Small-24** (512 dim, 24 layers): All layers 0-23
- **Small-36** (512 dim, 36 layers): Layers 0-23 (layers 24-35 pending)
- **Medium-16** (768 dim, 16 layers): All layers 0-15
- **Large-16** (1024 dim, 16 layers): All layers 0-15
## Probe Types
### Trained Model Probes
Linear classifiers trained on activations from models trained on chess games.
- Format: `tf_lens_{model_name}_chess_piece_probe_layer_{N}.pth`
- Example: `tf_lens_large-16-600K_iters_chess_piece_probe_layer_8.pth`
### Random Baseline Probes
Linear classifiers trained on activations from models with randomized weights, used as experimental controls.
- Format: `tf_lens_{model_name}_RANDOM_chess_piece_probe_layer_{N}.pth`
- Example: `tf_lens_large-16_RANDOM_chess_piece_probe_layer_8.pth`
## Probe Details
- **Task**: Predict the piece type on each of the 64 chess board squares
- **Input**: Model activations at specific sequence positions (after move notation dots)
- **Output**: 13-class classification per square (empty, 6 white pieces, 6 black pieces)
- **Architecture**: Single linear layer (no hidden layers)
- **Training**: Cross-entropy loss, trained on Stockfish games
## Key Findings
- **Trained models**: Show clear learning progression, with later layers achieving 75-99% accuracy
- **Random baselines**: Consistently lower performance (65-71%), validating experimental design
- **Layer progression**: Earlier layers show lower accuracy, later layers show higher accuracy
- **Model scaling**: Larger models tend to develop better board representations
## File Naming Convention
```
tf_lens_{model_size}-{layers}[-{training_iters}][_RANDOM]_chess_piece_probe_layer_{layer_num}.pth
```
Where:
- `model_size`: small, medium, large
- `layers`: 16, 24, 36
- `training_iters`: 600K_iters, 600k_iters
- `RANDOM`: Present for randomized baseline models
- `layer_num`: 0 to (layers-1)
## Usage
Load probes using PyTorch:
```python
import torch
# Load a trained probe
probe = torch.load('tf_lens_large-16-600K_iters_chess_piece_probe_layer_8.pth')
# The probe is a linear layer: torch.nn.Linear(d_model, 64*13)
# where d_model depends on the model (512/768/1024)
# and 64*13 represents 64 squares × 13 piece classes
```
## Research Context
This work is part of mechanistic interpretability research on chess language models, investigating:
- How board-state representations emerge during training
- Scaling laws for internal representations
- Layer-wise development of chess understanding
- Comparison between trained and random baselines
## Citation
If you use these probes in your research, please cite the original work:
```
@misc{chessgpt-board-probes-2024,
title={ChessGPT Board State Probes},
author={[Author Name]},
year={2024},
url={https://huggingface.co/jd0g/chessgpt-board-probes}
}
```
## License
MIT License - See LICENSE file for details.
|
nightmedia/Seed-OSS-36B-Instruct-qx6-mlx
|
nightmedia
| 2025-08-29T04:00:44Z | 157 | 0 |
mlx
|
[
"mlx",
"safetensors",
"seed_oss",
"vllm",
"text-generation",
"conversational",
"base_model:ByteDance-Seed/Seed-OSS-36B-Instruct",
"base_model:quantized:ByteDance-Seed/Seed-OSS-36B-Instruct",
"license:apache-2.0",
"6-bit",
"region:us"
] |
text-generation
| 2025-08-22T07:07:04Z |
---
license: apache-2.0
pipeline_tag: text-generation
library_name: mlx
tags:
- vllm
- mlx
base_model: ByteDance-Seed/Seed-OSS-36B-Instruct
---
# Seed-OSS-36B-Instruct-qx6-mlx
This model [Seed-OSS-36B-Instruct-qx6-mlx](https://huggingface.co/Seed-OSS-36B-Instruct-qx6-mlx) was
converted to MLX format from [ByteDance-Seed/Seed-OSS-36B-Instruct](https://huggingface.co/ByteDance-Seed/Seed-OSS-36B-Instruct)
using mlx-lm version **0.26.4**.
## Use with mlx
```bash
pip install mlx-lm
```
```python
from mlx_lm import load, generate
model, tokenizer = load("Seed-OSS-36B-Instruct-qx6-mlx")
prompt = "hello"
if tokenizer.chat_template is not None:
messages = [{"role": "user", "content": prompt}]
prompt = tokenizer.apply_chat_template(
messages, add_generation_prompt=True
)
response = generate(model, tokenizer, prompt=prompt, verbose=True)
```
|
bodigardehotma1/blockassist-bc-spotted_mimic_giraffe_1756438123
|
bodigardehotma1
| 2025-08-29T03:57:33Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"spotted mimic giraffe",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T03:57:21Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- spotted mimic giraffe
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
yanolja/EEVE-Korean-Instruct-7B-v2.0-Preview
|
yanolja
| 2025-08-29T03:51:32Z | 1,486 | 34 | null |
[
"safetensors",
"qwen2",
"generated_from_trainer",
"arxiv:2310.01377",
"arxiv:2501.12948",
"arxiv:2502.02737",
"arxiv:2402.14714",
"license:apache-2.0",
"region:us"
] | null | 2025-04-23T18:52:01Z |
---
license: apache-2.0
tags:
- generated_from_trainer
base_model: yanolja/EEVE-Korean-7B-v2.0-Preview
model-index:
- name: yanolja/EEVE-Korean-Instruct-7B-v2.0-Preview
results: []
---
[<img src="https://raw.githubusercontent.com/OpenAccess-AI-Collective/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/OpenAccess-AI-Collective/axolotl)
# EEVE-Korean-Instruct-7B-v2.0-Preview
## Join Our Community on Discord!
If you're passionate about the field of Large Language Models and wish to exchange knowledge and insights, we warmly invite you to join our Discord server. It's worth noting that Korean is the primary language used in this server. The landscape of LLM is evolving rapidly, and without active sharing, our collective knowledge risks becoming outdated swiftly. Let's collaborate and drive greater impact together! Join us here: [Discord Link](https://discord.gg/b27bAHg95m).
**Model Details**
## About the Model
EEVE-Korean-Instruct-7B-v2.0-Preview is an instruction-following large language model derived from Qwen2.5-7B. It has been specifically enhanced for Korean language understanding and generation through vocabulary expansion. A key feature is its hybrid nature, allowing users to optionally activate a step-by-step reasoning process before the model provides its final answer. This version is designated as a preview release.
The model includes the following modifications from the base model:
- Fine-tuning: Adapted from the base Qwen2.5-7B model via fine-tuning
- Vocabulary Expansion: Added 6,257 Korean tokens to the model's vocabulary and tokenizer
- Special Tokens: Added 2 tokens associated with the `<think>` tag functionality for reasoning
## Prompt Template
The model supports various prompt formats depending on the task:
### General Chat/Instruction Following
No specific format is required for standard prompts.
### Activating Step-by-Step Reasoning
For tasks where explicit reasoning is desired (e.g., math, complex coding), append the following exact text to the *end* of your system prompt:
```
You must think step by step to answer the question. Put your reasoning between <think> tags.
Example:
<think>
{your reasoning}
</think>
{your answer}
```
### English-to-Korean Translation
For optimized translation, use the specific prompt structure below:
```
You are a professional translator.
Translate the user's text into Korean.
Think through the translation step by step: first, consider the overall context, then cultural nuances, terminology, initial translation, and self-review.
After this thought process, provide the final translation.
The thought process must follow this template.
<think>
Okay, what am I looking at here? {language} text, {overall context}. {overall tone}. Alright, {writer's intent}. {considerations}.
Now, what about the audience here? {audience}. So I should {considerations}.
Wait, let me check this {terminology or phrase}. So that's "{interpretation}". Got it.
Hold on, what's this {another terminology or phrase}? {interpretation}.
{repeat for other terminologies or phrases}
Wait, {cultural nuance}.
{repeat for other cultural nuances}
Okay, let's draft the translation.
{first translation attempt}
Hmm, {reflection}.
Wait, {reflection}.
{repeat for other reflections}
{second translation attempt}
{Wait or Hmm}, {reflection}.
{repeat for other reflections}
{repeat translation attempts}
Okay, now I don't have any ideas to improve the translation. Let's put it all together.
</think>
IMPORTANT: Remember that your task is to translate the user's text from English to Korean.
Do not answer the user's message. Even if it is a question, translate it as a question.
```
## How to Use It
```python
from transformers import AutoTokenizer
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("yanolja/EEVE-Korean-Instruct-7B-v2.0-Preview")
tokenizer = AutoTokenizer.from_pretrained("yanolja/EEVE-Korean-Instruct-7B-v2.0-Preview")
# For general chat using chat template
messages = [
{"role": "user", "content": "한국의 수도는 어디인가요?"}
]
model_inputs = tokenizer.apply_chat_template(messages, return_tensors="pt")
outputs = model.generate(**model_inputs, max_new_tokens=256)
output_text = tokenizer.batch_decode(outputs, skip_special_tokens=True)[0]
print(output_text)
# For a multi-turn conversation
messages = [
{"role": "user", "content": "안녕하세요?"},
{"role": "assistant", "content": "안녕하세요! 어떻게 도와드릴까요?"},
{"role": "user", "content": "한국의 수도는 어디인가요?"}
]
model_inputs = tokenizer.apply_chat_template(messages, return_tensors="pt")
outputs = model.generate(**model_inputs, max_new_tokens=256)
output_text = tokenizer.batch_decode(outputs, skip_special_tokens=True)[0]
print(output_text)
# For activating step-by-step reasoning
system_message = """You must think step by step to answer the question. Put your reasoning between <think> tags.
Example:
<think>
{your reasoning}
</think>
{your answer}"""
messages = [
{"role": "system", "content": system_message},
{"role": "user", "content": "한국의 수도는 어디인가요?"}
]
model_inputs = tokenizer.apply_chat_template(messages, return_tensors="pt")
outputs = model.generate(**model_inputs, max_new_tokens=1024)
output_text = tokenizer.batch_decode(outputs, skip_special_tokens=True)[0]
print(output_text)
```
## Model Capabilities
- **Strengths:** Reported to be proficient in math, coding, and translation (specifically English-to-Korean with the provided prompt)
- **Language Focus:** Enhanced Korean language capabilities due to vocabulary additions
- **Reasoning:** Can provide step-by-step reasoning traces when prompted (and occasionally unsolicited)
## Limitations
- **Preview Status:** As a "Preview" version, it may contain bugs, instabilities, or undergo significant changes in future releases. Performance may not be fully optimized
- **General LLM Limitations:** Subject to potential issues like factual inaccuracies (hallucinations) which are particularly frequent with this model, generation of biased or harmful content, and inconsistencies
- **Performance Metrics:** Specific quantitative evaluation results are not yet available but will be attached soon
- **Reasoning Activation:** While the step-by-step reasoning feature is intended to be activated via a specific prompt, it may sometimes trigger without it
## Training Data
The model inherits knowledge from the training data of Qwen2.5-7B and was fine-tuned using a combination of datasets, including:
- Distilled data from DeepSeek-R1
- HuggingFaceTB/smoltalk (https://huggingface.co/datasets/HuggingFaceTB/smoltalk)
- HuggingFaceH4/ultrafeedback_binarized (https://huggingface.co/datasets/HuggingFaceH4/ultrafeedback_binarized)
- AI Hub Korean Conversation Summary dataset (https://aihub.or.kr/aihubdata/data/view.do?currMenu=115&topMenu=100&aihubDataSe=data&dataSetSn=71773)
### Citations for Training Data
```
@misc{cui2023ultrafeedback,
title={UltraFeedback: Boosting Language Models with High-quality Feedback},
author={Ganqu Cui and Lifan Yuan and Ning Ding and Guanming Yao and Wei Zhu and Yuan Ni and Guotong Xie and Zhiyuan Liu and Maosong Sun},
year={2023},
eprint={2310.01377},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
@misc{deepseekai2025deepseekr1incentivizingreasoningcapability,
title={DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning},
author={DeepSeek-AI},
year={2025},
eprint={2501.12948},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2501.12948},
}
@misc{allal2025smollm2smolgoesbig,
title={SmolLM2: When Smol Goes Big -- Data-Centric Training of a Small Language Model},
author={Loubna Ben Allal and Anton Lozhkov and Elie Bakouch and Gabriel Martín Blázquez and Guilherme Penedo and Lewis Tunstall and Andrés Marafioti and Hynek Kydlíček and Agustín Piqueres Lajarín and Vaibhav Srivastav and Joshua Lochner and Caleb Fahlgren and Xuan-Son Nguyen and Clémentine Fourrier and Ben Burtenshaw and Hugo Larcher and Haojun Zhao and Cyril Zakka and Mathieu Morlon and Colin Raffel and Leandro von Werra and Thomas Wolf},
year={2025},
eprint={2502.02737},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2502.02737},
}
```
## Ethical Considerations
- **License:** The Apache 2.0 license permits broad use but comes with conditions regarding liability and trademark use
- **Bias:** The model may reflect biases present in the Qwen2.5-7B base model and the datasets used for fine-tuning
- **Misuse Potential:** This model MUST not be used for generating misinformation, harmful content, or spam
## Citation
```
@misc{kim2024efficient,
title={Efficient and Effective Vocabulary Expansion Towards Multilingual Large Language Models},
author={Seungduk Kim and Seungtaek Choi and Myeongho Jeong},
year={2024},
eprint={2402.14714},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
```
# Evaluation Results
Quantitative evaluation results will be attached soon.
---
|
sreangrathanak/MyGemmaNPC
|
sreangrathanak
| 2025-08-29T03:48:33Z | 0 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"gemma3_text",
"text-generation",
"generated_from_trainer",
"trl",
"sft",
"conversational",
"base_model:google/gemma-3-270m-it",
"base_model:finetune:google/gemma-3-270m-it",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-08-29T03:44:56Z |
---
base_model: google/gemma-3-270m-it
library_name: transformers
model_name: MyGemmaNPC
tags:
- generated_from_trainer
- trl
- sft
licence: license
---
# Model Card for MyGemmaNPC
This model is a fine-tuned version of [google/gemma-3-270m-it](https://huggingface.co/google/gemma-3-270m-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="sreangrathanak/MyGemmaNPC", 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: 4.0.0
- Tokenizers: 0.21.4
## 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}}
}
```
|
FlagRelease/gpt-oss-120b-FlagOS
|
FlagRelease
| 2025-08-29T03:38:49Z | 0 | 0 | null |
[
"safetensors",
"gpt_oss",
"8-bit",
"mxfp4",
"region:us"
] | null | 2025-08-29T03:35:46Z |
# Introduction
**FlagOS** is a unified heterogeneous computing software stack for large models, co-developed with leading global chip manufacturers. With core technologies such as the **FlagScale** distributed training/inference framework, **FlagGems** universal operator library, **FlagCX** communication library, and **FlagTree** unified compiler, the **FlagRelease** platform leverages the FlagOS stack to automatically produce and release various combinations of <chip + open-source model>. This enables efficient and automated model migration across diverse chips, opening a new chapter for large model deployment and application.
Based on this, the **gpt-oss-120b-FlagOS** model is adapted for the Nvidia chip using the FlagOS software stack, enabling:
### Integrated Deployment
- Deep integration with the open-source [FlagScale framework](https://github.com/FlagOpen/FlagScale)
- Out-of-the-box inference scripts with pre-configured hardware and software parameters
- Released **FlagOS** container image supporting deployment within minutes
### Consistency Validation
- Rigorously evaluated through benchmark testing: Performance and results from the FlagOS software stack are compared against native stacks on multiple public.
# Technical Overview
## **FlagScale Distributed Training and Inference Framework**
FlagScale is an end-to-end framework for large models across heterogeneous computing resources, maximizing computational efficiency and ensuring model validity through core technologies. Its key advantages include:
- **Unified Deployment Interface:** Standardized command-line tools support one-click service deployment across multiple hardware platforms, significantly reducing adaptation costs in heterogeneous environments.
- **Intelligent Parallel Optimization:** Automatically generates optimal distributed parallel strategies based on chip computing characteristics, achieving dynamic load balancing of computation/communication resources.
- **Seamless Operator Switching:** Deep integration with the FlagGems operator library allows high-performance operators to be invoked via environment variables without modifying model code.
## **FlagGems Universal Large-Model Operator Library**
FlagGems is a Triton-based, cross-architecture operator library collaboratively developed with industry partners. Its core strengths include:
- **Full-stack Coverage**: Over 100 operators, with a broader range of operator types than competing libraries.
- **Ecosystem Compatibility**: Supports 7 accelerator backends. Ongoing optimizations have significantly improved performance.
- **High Efficiency**: Employs unique code generation and runtime optimization techniques for faster secondary development and better runtime performance compared to alternatives.
## **FlagEval Evaluation Framework**
FlagEval (Libra)** is a comprehensive evaluation system and open platform for large models launched in 2023. It aims to establish scientific, fair, and open benchmarks, methodologies, and tools to help researchers assess model and training algorithm performance. It features:
- **Multi-dimensional Evaluation**: Supports 800+ model evaluations across NLP, CV, Audio, and Multimodal fields, covering 20+ downstream tasks including language understanding and image-text generation.
- **Industry-Grade Use Cases**: Has completed horizontal evaluations of mainstream large models, providing authoritative benchmarks for chip-model performance validation.
# Evaluation Results
## Benchmark Result
| Metrics | gpt-oss-120b-H100-CUDA | gpt-oss-120b-FlagOS |
| ------------------------- | --------------------- | ------------------ |
|AIME-0shot@avg1|0.833|0.800|
|GPQA-0shot@avg1|0.669|0.679|
|MMLU-5shots@avg1|0.462|0.462|
|MUSR-0shot@avg1|0.672|0.681|
|LiveBench-0shot@avg1|0.678|0.674|
# User Guide
**Environment Setup**
| Item | Version |
| ------------- | ------------------------------------------------------------ |
| Docker Version | Docker version 28.1.0, build 4d8c241 |
| Operating System | Ubuntu 22.04.5 LTS |
| FlagScale | Version: 0.8.0 |
| FlagGems | Version: 3.0 |
## Operation Steps
### Download Open-source Model Weights
```bash
pip install modelscope
modelscope download --model openai-mirror/gpt-oss-120b --local_dir /share/models/gpt-oss-120b
```
### Download FlagOS Image
```bash
docker pull harbor.baai.ac.cn/flagrelease-public/flagrelease_nvidia_gpt
```
### Start the inference service
```bash
#Container Startup
docker run --rm --init --detach --net=host --uts=host --ipc=host --security-opt=seccomp=unconfined --privileged=true --ulimit stack=67108864 --ulimit memlock=-1 --ulimit nofile=1048576:1048576 --shm-size=32G -v /share:/share --gpus all --name flagos harbor.baai.ac.cn/flagrelease-public/flagrelease_nvidia_gpt sleep infinity
```
### Serve
```bash
flagscale serve gpt_oss
```
## Service Invocation
### API-based Invocation Script
```bash
import openai
openai.api_key = "EMPTY"
openai.base_url = "http://<server_ip>:9010/v1/"
model = "gpt-oss-120b-nvidia-flagos"
messages = [
{"role": "system", "content": "You are a helpful assistant."},
{"role": "user", "content": "What's the weather like today?"}
]
response = openai.chat.completions.create(
model=model,
messages=messages,
stream=False,
)
for item in response:
print(item)
```
### AnythingLLM Integration Guide
#### 1. Download & Install
- Visit the official site: https://anythingllm.com/
- Choose the appropriate version for your OS (Windows/macOS/Linux)
- Follow the installation wizard to complete the setup
#### 2. Configuration
- Launch AnythingLLM
- Open settings (bottom left, fourth tab)
- Configure core LLM parameters
- Click "Save Settings" to apply changes
#### 3. Model Interaction
- After model loading is complete:
- Click **"New Conversation"**
- Enter your question (e.g., “Explain the basics of quantum computing”)
- Click the send button to get a response
# Contributing
We warmly welcome global developers to join us:
1. Submit Issues to report problems
2. Create Pull Requests to contribute code
3. Improve technical documentation
4. Expand hardware adaptation support
# License
本模型的权重来源于openai-mirror/gpt-oss-120b，以apache2.0协议https://www.apache.org/licenses/LICENSE-2.0.txt开源。
|
hobson123/blockassist-bc-mammalian_dense_gibbon_1756438557
|
hobson123
| 2025-08-29T03:37:08Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"mammalian dense gibbon",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T03:36:54Z |
---
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).
|
vendi11/blockassist-bc-placid_placid_llama_1756438431
|
vendi11
| 2025-08-29T03:34:33Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"placid placid llama",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T03:34:29Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- placid placid llama
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
kinghanse/act_grabsalad
|
kinghanse
| 2025-08-29T03:29:30Z | 0 | 0 |
lerobot
|
[
"lerobot",
"safetensors",
"act",
"robotics",
"dataset:kinghanse/grab_salad",
"arxiv:2304.13705",
"license:apache-2.0",
"region:us"
] |
robotics
| 2025-08-29T03:29:14Z |
---
datasets: kinghanse/grab_salad
library_name: lerobot
license: apache-2.0
model_name: act
pipeline_tag: robotics
tags:
- act
- robotics
- lerobot
---
# Model Card for act
<!-- Provide a quick summary of what the model is/does. -->
[Action Chunking with Transformers (ACT)](https://huggingface.co/papers/2304.13705) is an imitation-learning method that predicts short action chunks instead of single steps. It learns from teleoperated data and often achieves high success rates.
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
|
liukevin666/blockassist-bc-yawning_striped_cassowary_1756437875
|
liukevin666
| 2025-08-29T03:25:52Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"yawning striped cassowary",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T03:25:30Z |
---
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).
|
rnoozy/blockassist-bc-pudgy_roaring_slug_1756437530
|
rnoozy
| 2025-08-29T03:20:54Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"pudgy roaring slug",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T03:20:50Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- pudgy roaring slug
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
Andra76/blockassist-bc-deadly_enormous_butterfly_1756436802
|
Andra76
| 2025-08-29T03:18:16Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"deadly enormous butterfly",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T03:17:25Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- deadly enormous butterfly
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
kojeklollipop/blockassist-bc-spotted_amphibious_stork_1756435824
|
kojeklollipop
| 2025-08-29T03:18:06Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"spotted amphibious stork",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T03:18:04Z |
---
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).
|
Muapi/zavy-s-detail-anime-digital-realism-flux
|
Muapi
| 2025-08-29T03:13:28Z | 0 | 0 | null |
[
"lora",
"stable-diffusion",
"flux.1-d",
"license:openrail++",
"region:us"
] | null | 2025-08-29T03:11:23Z |
---
license: openrail++
tags:
- lora
- stable-diffusion
- flux.1-d
model_type: LoRA
---
# Zavy's Detail Anime/Digital Realism - Flux
**Base model**: Flux.1 D
**Trained words**: zavy-dtlnm
## 🧠 Usage (Python)
🔑 **Get your MUAPI key** from [muapi.ai/access-keys](https://muapi.ai/access-keys)
```python
import requests, os
url = "https://api.muapi.ai/api/v1/flux_dev_lora_image"
headers = {"Content-Type": "application/json", "x-api-key": os.getenv("MUAPIAPP_API_KEY")}
payload = {
"prompt": "masterpiece, best quality, 1girl, looking at viewer",
"model_id": [{"model": "civitai:777532@869600", "weight": 1.0}],
"width": 1024,
"height": 1024,
"num_images": 1
}
print(requests.post(url, headers=headers, json=payload).json())
```
|
JunHowie/Qwen3-32B-GPTQ-Int4
|
JunHowie
| 2025-08-29T03:12:18Z | 11,609 | 3 |
transformers
|
[
"transformers",
"safetensors",
"qwen3",
"text-generation",
"conversational",
"arxiv:2309.00071",
"base_model:Qwen/Qwen3-32B",
"base_model:quantized:Qwen/Qwen3-32B",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"4-bit",
"gptq",
"region:us"
] |
text-generation
| 2025-04-29T22:26:24Z |
---
library_name: transformers
license: apache-2.0
license_link: https://huggingface.co/Qwen/Qwen3-8B/blob/main/LICENSE
pipeline_tag: text-generation
base_model:
- Qwen/Qwen3-32B
---
# Qwen3-8B
<a href="https://chat.qwen.ai/" target="_blank" style="margin: 2px;">
<img alt="Chat" src="https://img.shields.io/badge/%F0%9F%92%9C%EF%B8%8F%20Qwen%20Chat%20-536af5" style="display: inline-block; vertical-align: middle;"/>
</a>
## Qwen3 Highlights
Qwen3 is the latest generation of large language models in Qwen series, offering a comprehensive suite of dense and mixture-of-experts (MoE) models. Built upon extensive training, Qwen3 delivers groundbreaking advancements in reasoning, instruction-following, agent capabilities, and multilingual support, with the following key features:
- **Uniquely support of seamless switching between thinking mode** (for complex logical reasoning, math, and coding) and **non-thinking mode** (for efficient, general-purpose dialogue) **within single model**, ensuring optimal performance across various scenarios.
- **Significantly enhancement in its reasoning capabilities**, surpassing previous QwQ (in thinking mode) and Qwen2.5 instruct models (in non-thinking mode) on mathematics, code generation, and commonsense logical reasoning.
- **Superior human preference alignment**, excelling in creative writing, role-playing, multi-turn dialogues, and instruction following, to deliver a more natural, engaging, and immersive conversational experience.
- **Expertise in agent capabilities**, enabling precise integration with external tools in both thinking and unthinking modes and achieving leading performance among open-source models in complex agent-based tasks.
- **Support of 100+ languages and dialects** with strong capabilities for **multilingual instruction following** and **translation**.
## Model Overview
**Qwen3-8B** has the following features:
- Type: Causal Language Models
- Training Stage: Pretraining & Post-training
- Number of Parameters: 8.2B
- Number of Paramaters (Non-Embedding): 6.95B
- Number of Layers: 36
- Number of Attention Heads (GQA): 32 for Q and 8 for KV
- Context Length: 32,768 natively and [131,072 tokens with YaRN](#processing-long-texts).
For more details, including benchmark evaluation, hardware requirements, and inference performance, please refer to our [blog](https://qwenlm.github.io/blog/qwen3/), [GitHub](https://github.com/QwenLM/Qwen3), and [Documentation](https://qwen.readthedocs.io/en/latest/).
## Quickstart
The code of Qwen3 has been in the latest Hugging Face `transformers` and we advise you to use the latest version of `transformers`.
With `transformers<4.51.0`, you will encounter the following error:
```
KeyError: 'qwen3'
```
The following contains a code snippet illustrating how to use the model generate content based on given inputs.
```python
from transformers import AutoModelForCausalLM, AutoTokenizer
model_name = "Qwen/Qwen3-8B"
# load the tokenizer and the model
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(
model_name,
torch_dtype="auto",
device_map="auto"
)
# prepare the model input
prompt = "Give me a short introduction to large language model."
messages = [
{"role": "user", "content": prompt}
]
text = tokenizer.apply_chat_template(
messages,
tokenize=False,
add_generation_prompt=True,
enable_thinking=True # Switches between thinking and non-thinking modes. Default is True.
)
model_inputs = tokenizer([text], return_tensors="pt").to(model.device)
# conduct text completion
generated_ids = model.generate(
**model_inputs,
max_new_tokens=32768
)
output_ids = generated_ids[0][len(model_inputs.input_ids[0]):].tolist()
# parsing thinking content
try:
# rindex finding 151668 (</think>)
index = len(output_ids) - output_ids[::-1].index(151668)
except ValueError:
index = 0
thinking_content = tokenizer.decode(output_ids[:index], skip_special_tokens=True).strip("\n")
content = tokenizer.decode(output_ids[index:], skip_special_tokens=True).strip("\n")
print("thinking content:", thinking_content)
print("content:", content)
```
For deployment, you can use `sglang>=0.4.6.post1` or `vllm>=0.8.5` or to create an OpenAI-compatible API endpoint:
- SGLang:
```shell
python -m sglang.launch_server --model-path Qwen/Qwen3-8B --reasoning-parser qwen3
```
- vLLM:
```shell
vllm serve Qwen/Qwen3-8B --enable-reasoning --reasoning-parser deepseek_r1
```
For local use, applications such as Ollama, LMStudio, MLX-LM, llama.cpp, and KTransformers have also supported Qwen3.
## Switching Between Thinking and Non-Thinking Mode
> [!TIP]
> The `enable_thinking` switch is also available in APIs created by SGLang and vLLM.
> Please refer to our documentation for [SGLang](https://qwen.readthedocs.io/en/latest/deployment/sglang.html#thinking-non-thinking-modes) and [vLLM](https://qwen.readthedocs.io/en/latest/deployment/vllm.html#thinking-non-thinking-modes) users.
### `enable_thinking=True`
By default, Qwen3 has thinking capabilities enabled, similar to QwQ-32B. This means the model will use its reasoning abilities to enhance the quality of generated responses. For example, when explicitly setting `enable_thinking=True` or leaving it as the default value in `tokenizer.apply_chat_template`, the model will engage its thinking mode.
```python
text = tokenizer.apply_chat_template(
messages,
tokenize=False,
add_generation_prompt=True,
enable_thinking=True # True is the default value for enable_thinking
)
```
In this mode, the model will generate think content wrapped in a `<think>...</think>` block, followed by the final response.
> [!NOTE]
> For thinking mode, use `Temperature=0.6`, `TopP=0.95`, `TopK=20`, and `MinP=0` (the default setting in `generation_config.json`). **DO NOT use greedy decoding**, as it can lead to performance degradation and endless repetitions. For more detailed guidance, please refer to the [Best Practices](#best-practices) section.
### `enable_thinking=False`
We provide a hard switch to strictly disable the model's thinking behavior, aligning its functionality with the previous Qwen2.5-Instruct models. This mode is particularly useful in scenarios where disabling thinking is essential for enhancing efficiency.
```python
text = tokenizer.apply_chat_template(
messages,
tokenize=False,
add_generation_prompt=True,
enable_thinking=False # Setting enable_thinking=False disables thinking mode
)
```
In this mode, the model will not generate any think content and will not include a `<think>...</think>` block.
> [!NOTE]
> For non-thinking mode, we suggest using `Temperature=0.7`, `TopP=0.8`, `TopK=20`, and `MinP=0`. For more detailed guidance, please refer to the [Best Practices](#best-practices) section.
### Advanced Usage: Switching Between Thinking and Non-Thinking Modes via User Input
We provide a soft switch mechanism that allows users to dynamically control the model's behavior when `enable_thinking=True`. Specifically, you can add `/think` and `/no_think` to user prompts or system messages to switch the model's thinking mode from turn to turn. The model will follow the most recent instruction in multi-turn conversations.
Here is an example of a multi-turn conversation:
```python
from transformers import AutoModelForCausalLM, AutoTokenizer
class QwenChatbot:
def __init__(self, model_name="Qwen/Qwen3-8B"):
self.tokenizer = AutoTokenizer.from_pretrained(model_name)
self.model = AutoModelForCausalLM.from_pretrained(model_name)
self.history = []
def generate_response(self, user_input):
messages = self.history + [{"role": "user", "content": user_input}]
text = self.tokenizer.apply_chat_template(
messages,
tokenize=False,
add_generation_prompt=True
)
inputs = self.tokenizer(text, return_tensors="pt")
response_ids = self.model.generate(**inputs, max_new_tokens=32768)[0][len(inputs.input_ids[0]):].tolist()
response = self.tokenizer.decode(response_ids, skip_special_tokens=True)
# Update history
self.history.append({"role": "user", "content": user_input})
self.history.append({"role": "assistant", "content": response})
return response
# Example Usage
if __name__ == "__main__":
chatbot = QwenChatbot()
# First input (without /think or /no_think tags, thinking mode is enabled by default)
user_input_1 = "How many r's in strawberries?"
print(f"User: {user_input_1}")
response_1 = chatbot.generate_response(user_input_1)
print(f"Bot: {response_1}")
print("----------------------")
# Second input with /no_think
user_input_2 = "Then, how many r's in blueberries? /no_think"
print(f"User: {user_input_2}")
response_2 = chatbot.generate_response(user_input_2)
print(f"Bot: {response_2}")
print("----------------------")
# Third input with /think
user_input_3 = "Really? /think"
print(f"User: {user_input_3}")
response_3 = chatbot.generate_response(user_input_3)
print(f"Bot: {response_3}")
```
> [!NOTE]
> For API compatibility, when `enable_thinking=True`, regardless of whether the user uses `/think` or `/no_think`, the model will always output a block wrapped in `<think>...</think>`. However, the content inside this block may be empty if thinking is disabled.
> When `enable_thinking=False`, the soft switches are not valid. Regardless of any `/think` or `/no_think` tags input by the user, the model will not generate think content and will not include a `<think>...</think>` block.
## Agentic Use
Qwen3 excels in tool calling capabilities. We recommend using [Qwen-Agent](https://github.com/QwenLM/Qwen-Agent) to make the best use of agentic ability of Qwen3. Qwen-Agent encapsulates tool-calling templates and tool-calling parsers internally, greatly reducing coding complexity.
To define the available tools, you can use the MCP configuration file, use the integrated tool of Qwen-Agent, or integrate other tools by yourself.
```python
from qwen_agent.agents import Assistant
# Define LLM
llm_cfg = {
'model': 'Qwen3-8B',
# Use the endpoint provided by Alibaba Model Studio:
# 'model_type': 'qwen_dashscope',
# 'api_key': os.getenv('DASHSCOPE_API_KEY'),
# Use a custom endpoint compatible with OpenAI API:
'model_server': 'http://localhost:8000/v1', # api_base
'api_key': 'EMPTY',
# Other parameters:
# 'generate_cfg': {
# # Add: When the response content is `<think>this is the thought</think>this is the answer;
# # Do not add: When the response has been separated by reasoning_content and content.
# 'thought_in_content': True,
# },
}
# Define Tools
tools = [
{'mcpServers': { # You can specify the MCP configuration file
'time': {
'command': 'uvx',
'args': ['mcp-server-time', '--local-timezone=Asia/Shanghai']
},
"fetch": {
"command": "uvx",
"args": ["mcp-server-fetch"]
}
}
},
'code_interpreter', # Built-in tools
]
# Define Agent
bot = Assistant(llm=llm_cfg, function_list=tools)
# Streaming generation
messages = [{'role': 'user', 'content': 'https://qwenlm.github.io/blog/ Introduce the latest developments of Qwen'}]
for responses in bot.run(messages=messages):
pass
print(responses)
```
## Processing Long Texts
Qwen3 natively supports context lengths of up to 32,768 tokens. For conversations where the total length (including both input and output) significantly exceeds this limit, we recommend using RoPE scaling techniques to handle long texts effectively. We have validated the model's performance on context lengths of up to 131,072 tokens using the [YaRN](https://arxiv.org/abs/2309.00071) method.
YaRN is currently supported by several inference frameworks, e.g., `transformers` and `llama.cpp` for local use, `vllm` and `sglang` for deployment. In general, there are two approaches to enabling YaRN for supported frameworks:
- Modifying the model files:
In the `config.json` file, add the `rope_scaling` fields:
```json
{
...,
"rope_scaling": {
"rope_type": "yarn",
"factor": 4.0,
"original_max_position_embeddings": 32768
}
}
```
For `llama.cpp`, you need to regenerate the GGUF file after the modification.
- Passing command line arguments:
For `vllm`, you can use
```shell
vllm serve ... --rope-scaling '{"rope_type":"yarn","factor":4.0,"original_max_position_embeddings":32768}' --max-model-len 131072
```
For `sglang`, you can use
```shell
python -m sglang.launch_server ... --json-model-override-args '{"rope_scaling":{"rope_type":"yarn","factor":4.0,"original_max_position_embeddings":32768}}'
```
For `llama-server` from `llama.cpp`, you can use
```shell
llama-server ... --rope-scaling yarn --rope-scale 4 --yarn-orig-ctx 32768
```
> [!IMPORTANT]
> If you encounter the following warning
> ```
> Unrecognized keys in `rope_scaling` for 'rope_type'='yarn': {'original_max_position_embeddings'}
> ```
> please upgrade `transformers>=4.51.0`.
> [!NOTE]
> All the notable open-source frameworks implement static YaRN, which means the scaling factor remains constant regardless of input length, **potentially impacting performance on shorter texts.**
> We advise adding the `rope_scaling` configuration only when processing long contexts is required.
> It is also recommended to modify the `factor` as needed. For example, if the typical context length for your application is 65,536 tokens, it would be better to set `factor` as 2.0.
> [!NOTE]
> The default `max_position_embeddings` in `config.json` is set to 40,960. This allocation includes reserving 32,768 tokens for outputs and 8,192 tokens for typical prompts, which is sufficient for most scenarios involving short text processing. If the average context length does not exceed 32,768 tokens, we do not recommend enabling YaRN in this scenario, as it may potentially degrade model performance.
> [!TIP]
> The endpoint provided by Alibaba Model Studio supports dynamic YaRN by default and no extra configuration is needed.
## Best Practices
To achieve optimal performance, we recommend the following settings:
1. **Sampling Parameters**:
- For thinking mode (`enable_thinking=True`), use `Temperature=0.6`, `TopP=0.95`, `TopK=20`, and `MinP=0`. **DO NOT use greedy decoding**, as it can lead to performance degradation and endless repetitions.
- For non-thinking mode (`enable_thinking=False`), we suggest using `Temperature=0.7`, `TopP=0.8`, `TopK=20`, and `MinP=0`.
- For supported frameworks, you can adjust the `presence_penalty` parameter between 0 and 2 to reduce endless repetitions. However, using a higher value may occasionally result in language mixing and a slight decrease in model performance.
2. **Adequate Output Length**: We recommend using an output length of 32,768 tokens for most queries. For benchmarking on highly complex problems, such as those found in math and programming competitions, we suggest setting the max output length to 38,912 tokens. This provides the model with sufficient space to generate detailed and comprehensive responses, thereby enhancing its overall performance.
3. **Standardize Output Format**: We recommend using prompts to standardize model outputs when benchmarking.
- **Math Problems**: Include "Please reason step by step, and put your final answer within \boxed{}." in the prompt.
- **Multiple-Choice Questions**: Add the following JSON structure to the prompt to standardize responses: "Please show your choice in the `answer` field with only the choice letter, e.g., `"answer": "C"`."
4. **No Thinking Content in History**: In multi-turn conversations, the historical model output should only include the final output part and does not need to include the thinking content. It is implemented in the provided chat template in Jinja2. However, for frameworks that do not directly use the Jinja2 chat template, it is up to the developers to ensure that the best practice is followed.
### Citation
If you find our work helpful, feel free to give us a cite.
```
@misc{qwen3,
title = {Qwen3},
url = {https://qwenlm.github.io/blog/qwen3/},
author = {Qwen Team},
month = {April},
year = {2025}
}
```
|
Muapi/anemone-anomaly
|
Muapi
| 2025-08-29T03:09:11Z | 0 | 0 | null |
[
"lora",
"stable-diffusion",
"flux.1-d",
"license:openrail++",
"region:us"
] | null | 2025-08-29T03:08:56Z |
---
license: openrail++
tags:
- lora
- stable-diffusion
- flux.1-d
model_type: LoRA
---
# Anemone Anomaly
**Base model**: Flux.1 D
**Trained words**:
## 🧠 Usage (Python)
🔑 **Get your MUAPI key** from [muapi.ai/access-keys](https://muapi.ai/access-keys)
```python
import requests, os
url = "https://api.muapi.ai/api/v1/flux_dev_lora_image"
headers = {"Content-Type": "application/json", "x-api-key": os.getenv("MUAPIAPP_API_KEY")}
payload = {
"prompt": "masterpiece, best quality, 1girl, looking at viewer",
"model_id": [{"model": "civitai:275531@1546916", "weight": 1.0}],
"width": 1024,
"height": 1024,
"num_images": 1
}
print(requests.post(url, headers=headers, json=payload).json())
```
|
Muapi/scribble-portrait-drawing-style
|
Muapi
| 2025-08-29T03:08:50Z | 0 | 0 | null |
[
"lora",
"stable-diffusion",
"flux.1-d",
"license:openrail++",
"region:us"
] | null | 2025-08-29T03:08:35Z |
---
license: openrail++
tags:
- lora
- stable-diffusion
- flux.1-d
model_type: LoRA
---
# Scribble Portrait Drawing Style
**Base model**: Flux.1 D
**Trained words**: Drawing
## 🧠 Usage (Python)
🔑 **Get your MUAPI key** from [muapi.ai/access-keys](https://muapi.ai/access-keys)
```python
import requests, os
url = "https://api.muapi.ai/api/v1/flux_dev_lora_image"
headers = {"Content-Type": "application/json", "x-api-key": os.getenv("MUAPIAPP_API_KEY")}
payload = {
"prompt": "masterpiece, best quality, 1girl, looking at viewer",
"model_id": [{"model": "civitai:1179075@1326830", "weight": 1.0}],
"width": 1024,
"height": 1024,
"num_images": 1
}
print(requests.post(url, headers=headers, json=payload).json())
```
|
kn2021/MyGemmaNPC
|
kn2021
| 2025-08-29T03:06:08Z | 0 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"gemma3_text",
"text-generation",
"generated_from_trainer",
"trl",
"sft",
"conversational",
"base_model:google/gemma-3-270m-it",
"base_model:finetune:google/gemma-3-270m-it",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-08-29T01:43:40Z |
---
base_model: google/gemma-3-270m-it
library_name: transformers
model_name: MyGemmaNPC
tags:
- generated_from_trainer
- trl
- sft
licence: license
---
# Model Card for MyGemmaNPC
This model is a fine-tuned version of [google/gemma-3-270m-it](https://huggingface.co/google/gemma-3-270m-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="kn2021/MyGemmaNPC", 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: 4.0.0
- Tokenizers: 0.21.4
## 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}}
}
```
|
Muapi/eldritch-ink-art-for-flux
|
Muapi
| 2025-08-29T02:56:42Z | 0 | 0 | null |
[
"lora",
"stable-diffusion",
"flux.1-d",
"license:openrail++",
"region:us"
] | null | 2025-08-28T20:24:00Z |
---
license: openrail++
tags:
- lora
- stable-diffusion
- flux.1-d
model_type: LoRA
---
# Eldritch Ink Art | for Flux
**Base model**: Flux.1 D
**Trained words**: illustration
## 🧠 Usage (Python)
🔑 **Get your MUAPI key** from [muapi.ai/access-keys](https://muapi.ai/access-keys)
```python
import requests, os
url = "https://api.muapi.ai/api/v1/flux_dev_lora_image"
headers = {"Content-Type": "application/json", "x-api-key": os.getenv("MUAPIAPP_API_KEY")}
payload = {
"prompt": "masterpiece, best quality, 1girl, looking at viewer",
"model_id": [{"model": "civitai:696659@779594", "weight": 1.0}],
"width": 1024,
"height": 1024,
"num_images": 1
}
print(requests.post(url, headers=headers, json=payload).json())
```
|
poojavalecha09/llama-3.2-3b-finetuned-mathdaily
|
poojavalecha09
| 2025-08-29T02:52:04Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"text-generation-inference",
"unsloth",
"llama",
"trl",
"en",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] | null | 2024-11-03T09:18:30Z |
---
base_model: unsloth/llama-3.2-3b-instruct-bnb-4bit
language:
- en
license: apache-2.0
tags:
- text-generation-inference
- transformers
- unsloth
- llama
- trl
---
# Uploaded model
- **Developed by:** pooja valecha
- **License:** apache-2.0
- **Finetuned from model :** unsloth/llama-3.2-3b-instruct-bnb-4bit
This llama 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)
|
Alexismireles/Alexisai
|
Alexismireles
| 2025-08-29T02:51:55Z | 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
| 2025-08-29T02:20:41Z |
---
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: Alexisai
---
# Alexisai
<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 `Alexisai` to trigger the image generation.
## Run this LoRA with an API using Replicate
```py
import replicate
input = {
"prompt": "Alexisai",
"lora_weights": "https://huggingface.co/Alexismireles/Alexisai/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('Alexismireles/Alexisai', weight_name='lora.safetensors')
image = pipeline('Alexisai').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: 2000
- Learning rate: 0.0004
- LoRA rank: 16
## Contribute your own examples
You can use the [community tab](https://huggingface.co/Alexismireles/Alexisai/discussions) to add images that show off what you’ve made with this LoRA.
|
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).
|
derikk/training-checkpoint-step21
|
derikk
| 2025-08-29T02:50:21Z | 18 | 0 | null |
[
"pytorch",
"safetensors",
"qwen2",
"region:us"
] | null | 2025-08-26T23:50:12Z |
# Checkpoint Upload
This model checkpoint was automatically uploaded from a distributed training run.
## Model Details
- Training step: 21
- Architecture: Llama-style model
- Hidden size: 2048
- Layers: 36
- Vocabulary size: 151,936
## Checkpoint Information
- Originally saved as distributed checkpoint across 4 ranks
- Consolidated into single checkpoint for easier use
- Contains model weights, optimizer states, and training configuration
## Usage
```python
import torch
# Load the checkpoint
checkpoint = torch.load('pytorch_model.bin', map_location='cpu')
# The checkpoint contains the model state dict
# You'll need to initialize the appropriate model architecture
# and load these weights
```
## Note
This is a raw training checkpoint. For inference, you may need to:
1. Initialize the correct model architecture
2. Load the weights properly
3. Convert to the desired format (e.g., Hugging Face Transformers format)
|
John6666/illustrious-pixelart-from-hades-v3-series-v30-sdxl
|
John6666
| 2025-08-29T02:48:51Z | 0 | 0 |
diffusers
|
[
"diffusers",
"safetensors",
"text-to-image",
"stable-diffusion",
"stable-diffusion-xl",
"anime",
"pixel art",
"2D",
"retro",
"indie",
"clean lines",
"sharp detail",
"consistent palettes",
"adherence",
"perspective",
"poses",
"consistency",
"game assets",
"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
| 2025-08-29T02:44:09Z |
---
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
- pixel art
- 2D
- retro
- indie
- clean lines
- sharp detail
- consistent palettes
- adherence
- perspective
- poses
- consistency
- game assets
- illustrious
base_model: OnomaAIResearch/Illustrious-xl-early-release-v0
---
Original model is [here](https://civitai.com/models/1732312?modelVersionId=2158132).
This model created by [DeViLDoNia](https://civitai.com/user/DeViLDoNia).
|
arianaazarbal/standard_tpr_0.65-20250823_060848_grpo_recontextualized_20250829_024027-policy-adapter
|
arianaazarbal
| 2025-08-29T02:41:13Z | 0 | 0 | null |
[
"region:us"
] | null | 2025-08-29T02:41:12Z |
# Policy Model LoRA Adapter (GRPO/DPO)
Experiment: standard_tpr_0.65
Timestamp: 20250823_060848_grpo_recontextualized_20250829_024027
This model was trained as part of the deception-evasion-honesty experiments.
## Model Details
- **Type**: Policy Model LoRA Adapter (GRPO/DPO)
- **Experiment Name**: standard_tpr_0.65
- **Training Timestamp**: 20250823_060848_grpo_recontextualized_20250829_024027
|
huggingtoots/TheDrummer-Behemoth-R1-123B-v2-MLX-8Bit
|
huggingtoots
| 2025-08-29T02:31:27Z | 0 | 0 |
mlx
|
[
"mlx",
"safetensors",
"mistral",
"base_model:TheDrummer/Behemoth-R1-123B-v2",
"base_model:quantized:TheDrummer/Behemoth-R1-123B-v2",
"8-bit",
"region:us"
] | null | 2025-08-29T01:38:25Z |
---
base_model: TheDrummer/Behemoth-R1-123B-v2
tags:
- mlx
---
# huggingtoots/TheDrummer-Behemoth-R1-123B-v2-MLX-8Bit
The Model [huggingtoots/TheDrummer-Behemoth-R1-123B-v2-MLX-8Bit](https://huggingface.co/huggingtoots/TheDrummer-Behemoth-R1-123B-v2-MLX-8Bit) 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-8Bit")
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)
```
|
gartyopra/blockassist-bc-padded_mangy_penguin_1756434291
|
gartyopra
| 2025-08-29T02:25:42Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"padded mangy penguin",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T02:25:16Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- padded mangy penguin
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
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
|
rvipitkirubbe/blockassist-bc-mottled_foraging_ape_1756432634
|
rvipitkirubbe
| 2025-08-29T02:23:14Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"mottled foraging ape",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T02:23:09Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- mottled foraging ape
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
AnerYubo/blockassist-bc-pesty_graceful_grouse_1756434187
|
AnerYubo
| 2025-08-29T02:23:10Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"pesty graceful grouse",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T02:23:07Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- pesty graceful grouse
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
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).
|
SvalTek/Q2.5-ColdBrew-14B-Base-4Bit
|
SvalTek
| 2025-08-29T02:20:55Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"qwen2",
"text-generation",
"text-generation-inference",
"unsloth",
"conversational",
"en",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"4-bit",
"bitsandbytes",
"region:us"
] |
text-generation
| 2025-08-29T02:16:14Z |
---
tags:
- text-generation-inference
- transformers
- unsloth
- qwen2
license: apache-2.0
language:
- en
---
# Uploaded finetuned model
- **Developed by:** SvalTek
- **License:** apache-2.0
- **Finetuned from model :** suayptalha/Lamarckvergence-14B
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)
|
second-state/jina-embeddings-v3-GGUF
|
second-state
| 2025-08-29T02:14:06Z | 0 | 0 | null |
[
"gguf",
"custom_code",
"base_model:jinaai/jina-embeddings-v3",
"base_model:quantized:jinaai/jina-embeddings-v3",
"endpoints_compatible",
"region:us",
"feature-extraction"
] | null | 2025-08-28T15:01:19Z |
---
base_model: jinaai/jina-embeddings-v3
model_creator: jinaai
quantized_by: Second State Inc.
---
<!-- header start -->
<!-- 200823 -->
<div style="width: auto; margin-left: auto; margin-right: auto">
<img src="https://github.com/LlamaEdge/LlamaEdge/raw/dev/assets/logo.svg" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
<!-- header end -->
# jina-embeddings-v3-Embedding-GGUF
## Original Model
[jinaai/jina-embeddings-v3](https://huggingface.co/jinaai/jina-embeddings-v3)
## Run with LlamaEdge
- LlamaEdge version: coming soon
<!-- - LlamaEdge version: [v0.14.17](https://github.com/LlamaEdge/LlamaEdge/releases/tag/0.14.17) -->
- Prompt template
- Prompt type: `embedding`
- Embedding size: `32, 64, 128, 256, 512, 768, 1024`
- Run as LlamaEdge service
```bash
wasmedge --dir .:. --nn-preload default:GGML:AUTO:jina-embeddings-v3-f16.gguf \
llama-api-server.wasm \
--prompt-template embedding \
--ctx-size 768 \
--model-name jina-embeddings-v3
```
<!--
## Quantized GGUF Models
| Name | Quant method | Bits | Size | Use case |
| ---- | ---- | ---- | ---- | ----- |
| [jina-embeddings-v3-Q2_K.gguf](https://huggingface.co/second-state/jina-embeddings-v3-Embedding-GGUF/blob/main/jina-embeddings-v3-Q2_K.gguf) | Q2_K | 2 | 60.8 MB| smallest, significant quality loss - not recommended for most purposes |
| [jina-embeddings-v3-Q3_K_L.gguf](https://huggingface.co/second-state/jina-embeddings-v3-Embedding-GGUF/blob/main/jina-embeddings-v3-Q3_K_L.gguf) | Q3_K_L | 3 | 80.6 MB| small, substantial quality loss |
| [jina-embeddings-v3-Q3_K_M.gguf](https://huggingface.co/second-state/jina-embeddings-v3-Embedding-GGUF/blob/main/jina-embeddings-v3-Q3_K_M.gguf) | Q3_K_M | 3 | 76.2 MB| very small, high quality loss |
| [jina-embeddings-v3-Q3_K_S.gguf](https://huggingface.co/second-state/jina-embeddings-v3-Embedding-GGUF/blob/main/jina-embeddings-v3-Q3_K_S.gguf) | Q3_K_S | 3 | 68.7 MB| very small, high quality loss |
| [jina-embeddings-v3-Q4_0.gguf](https://huggingface.co/second-state/jina-embeddings-v3-Embedding-GGUF/blob/main/jina-embeddings-v3-Q4_0.gguf) | Q4_0 | 4 | 83.7 MB| legacy; small, very high quality loss - prefer using Q3_K_M |
| [jina-embeddings-v3-Q4_K_M.gguf](https://huggingface.co/second-state/jina-embeddings-v3-Embedding-GGUF/blob/main/jina-embeddings-v3-Q4_K_M.gguf) | Q4_K_M | 4 | 90.0 MB| medium, balanced quality - recommended |
| [jina-embeddings-v3-Q4_K_S.gguf](https://huggingface.co/second-state/jina-embeddings-v3-Embedding-GGUF/blob/main/jina-embeddings-v3-Q4_K_S.gguf) | Q4_K_S | 4 | 84.0 MB| small, greater quality loss |
| [jina-embeddings-v3-Q5_0.gguf](https://huggingface.co/second-state/jina-embeddings-v3-Embedding-GGUF/blob/main/jina-embeddings-v3-Q5_0.gguf) | Q5_0 | 5 | 97.9 MB| legacy; medium, balanced quality - prefer using Q4_K_M |
| [jina-embeddings-v3-Q5_K_M.gguf](https://huggingface.co/second-state/jina-embeddings-v3-Embedding-GGUF/blob/main/jina-embeddings-v3-Q5_K_M.gguf) | Q5_K_M | 5 | 103 MB| large, very low quality loss - recommended |
| [jina-embeddings-v3-Q5_K_S.gguf](https://huggingface.co/second-state/jina-embeddings-v3-Embedding-GGUF/blob/main/jina-embeddings-v3-Q5_K_S.gguf) | Q5_K_S | 5 | 97.9 MB| large, low quality loss - recommended |
| [jina-embeddings-v3-Q6_K.gguf](https://huggingface.co/second-state/jina-embeddings-v3-Embedding-GGUF/blob/main/jina-embeddings-v3-Q6_K.gguf) | Q6_K | 6 | 113 MB| very large, extremely low quality loss |
| [jina-embeddings-v3-Q8_0.gguf](https://huggingface.co/second-state/jina-embeddings-v3-Embedding-GGUF/blob/main/jina-embeddings-v3-Q8_0.gguf) | Q8_0 | 8 | 146 MB| very large, extremely low quality loss - not recommended |
| [jina-embeddings-v3-f16.gguf](https://huggingface.co/second-state/jina-embeddings-v3-Embedding-GGUF/blob/main/jina-embeddings-v3-f16.gguf) | f16 | 16 | 274 MB| very large, extremely low quality loss - not recommended |
-->
*Quantized with llama.cpp b6311*
|
csavzzcw/blockassist-bc-soft_curious_camel_1756433360
|
csavzzcw
| 2025-08-29T02:09:39Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"soft curious camel",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T02:09:21Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- soft curious camel
---
# 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_1756433012
|
bah63843
| 2025-08-29T02:04:24Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"plump fast antelope",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T02:04:16Z |
---
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).
|
SvalTek/Q2.5-ColdBrew-14B-Base-LoRa
|
SvalTek
| 2025-08-29T02:00:09Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"text-generation-inference",
"unsloth",
"qwen2",
"trl",
"en",
"base_model:suayptalha/Lamarckvergence-14B",
"base_model:finetune:suayptalha/Lamarckvergence-14B",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] | null | 2025-08-29T01:59:41Z |
---
base_model: suayptalha/Lamarckvergence-14B
tags:
- text-generation-inference
- transformers
- unsloth
- qwen2
- trl
license: apache-2.0
language:
- en
---
# Uploaded model
- **Developed by:** SvalTek
- **License:** apache-2.0
- **Finetuned from model :** suayptalha/Lamarckvergence-14B
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)
|
Ygz-08123/gemma-2-9b-Q4_K_M-GGUF
|
Ygz-08123
| 2025-08-29T01:55:21Z | 0 | 0 |
transformers
|
[
"transformers",
"gguf",
"unsloth",
"gemma2",
"gemma",
"llama-cpp",
"gguf-my-repo",
"en",
"base_model:unsloth/gemma-2-9b",
"base_model:quantized:unsloth/gemma-2-9b",
"license:gemma",
"endpoints_compatible",
"region:us"
] | null | 2025-08-29T01:54:56Z |
---
language:
- en
library_name: transformers
license: gemma
tags:
- unsloth
- transformers
- gemma2
- gemma
- llama-cpp
- gguf-my-repo
base_model: unsloth/gemma-2-9b
---
# Ygz-08123/gemma-2-9b-Q4_K_M-GGUF
This model was converted to GGUF format from [`unsloth/gemma-2-9b`](https://huggingface.co/unsloth/gemma-2-9b) using llama.cpp via the ggml.ai's [GGUF-my-repo](https://huggingface.co/spaces/ggml-org/gguf-my-repo) space.
Refer to the [original model card](https://huggingface.co/unsloth/gemma-2-9b) for more details on the model.
## Use with llama.cpp
Install llama.cpp through brew (works on Mac and Linux)
```bash
brew install llama.cpp
```
Invoke the llama.cpp server or the CLI.
### CLI:
```bash
llama-cli --hf-repo Ygz-08123/gemma-2-9b-Q4_K_M-GGUF --hf-file gemma-2-9b-q4_k_m.gguf -p "The meaning to life and the universe is"
```
### Server:
```bash
llama-server --hf-repo Ygz-08123/gemma-2-9b-Q4_K_M-GGUF --hf-file gemma-2-9b-q4_k_m.gguf -c 2048
```
Note: You can also use this checkpoint directly through the [usage steps](https://github.com/ggerganov/llama.cpp?tab=readme-ov-file#usage) listed in the Llama.cpp repo as well.
Step 1: Clone llama.cpp from GitHub.
```
git clone https://github.com/ggerganov/llama.cpp
```
Step 2: Move into the llama.cpp folder and build it with `LLAMA_CURL=1` flag along with other hardware-specific flags (for ex: LLAMA_CUDA=1 for Nvidia GPUs on Linux).
```
cd llama.cpp && LLAMA_CURL=1 make
```
Step 3: Run inference through the main binary.
```
./llama-cli --hf-repo Ygz-08123/gemma-2-9b-Q4_K_M-GGUF --hf-file gemma-2-9b-q4_k_m.gguf -p "The meaning to life and the universe is"
```
or
```
./llama-server --hf-repo Ygz-08123/gemma-2-9b-Q4_K_M-GGUF --hf-file gemma-2-9b-q4_k_m.gguf -c 2048
```
|
obsidian368/humanomni_lora
|
obsidian368
| 2025-08-29T01:55:07Z | 0 | 0 | null |
[
"safetensors",
"license:apache-2.0",
"region:us"
] | null | 2025-08-29T01:50:18Z |
---
license: apache-2.0
---
|
Ahmed-88889/cordv2_model
|
Ahmed-88889
| 2025-08-29T01:47:36Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"arxiv:1910.09700",
"endpoints_compatible",
"region:us"
] | null | 2025-08-29T01:44:29Z |
---
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]
|
mradermacher/Autoregressive-1.5B-2-GGUF
|
mradermacher
| 2025-08-29T01:47:06Z | 0 | 0 |
transformers
|
[
"transformers",
"gguf",
"en",
"base_model:InfiniAILab/Autoregressive-1.5B-2",
"base_model:quantized:InfiniAILab/Autoregressive-1.5B-2",
"endpoints_compatible",
"region:us",
"conversational"
] | null | 2025-08-29T00:41:34Z |
---
base_model: InfiniAILab/Autoregressive-1.5B-2
language:
- en
library_name: transformers
mradermacher:
readme_rev: 1
quantized_by: mradermacher
---
## About
<!-- ### quantize_version: 2 -->
<!-- ### output_tensor_quantised: 1 -->
<!-- ### convert_type: hf -->
<!-- ### vocab_type: -->
<!-- ### tags: -->
<!-- ### quants: x-f16 Q4_K_S Q2_K Q8_0 Q6_K Q3_K_M Q3_K_S Q3_K_L Q4_K_M Q5_K_S Q5_K_M IQ4_XS -->
<!-- ### quants_skip: -->
<!-- ### skip_mmproj: -->
static quants of https://huggingface.co/InfiniAILab/Autoregressive-1.5B-2
<!-- provided-files -->
***For a convenient overview and download list, visit our [model page for this model](https://hf.tst.eu/model#Autoregressive-1.5B-2-GGUF).***
weighted/imatrix quants seem not to be available (by me) at this time. If they do not show up a week or so after the static ones, I have probably not planned for them. Feel free to request them by opening a Community Discussion.
## 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/Autoregressive-1.5B-2-GGUF/resolve/main/Autoregressive-1.5B-2.Q2_K.gguf) | Q2_K | 0.9 | |
| [GGUF](https://huggingface.co/mradermacher/Autoregressive-1.5B-2-GGUF/resolve/main/Autoregressive-1.5B-2.Q3_K_S.gguf) | Q3_K_S | 1.0 | |
| [GGUF](https://huggingface.co/mradermacher/Autoregressive-1.5B-2-GGUF/resolve/main/Autoregressive-1.5B-2.Q3_K_M.gguf) | Q3_K_M | 1.0 | lower quality |
| [GGUF](https://huggingface.co/mradermacher/Autoregressive-1.5B-2-GGUF/resolve/main/Autoregressive-1.5B-2.Q3_K_L.gguf) | Q3_K_L | 1.1 | |
| [GGUF](https://huggingface.co/mradermacher/Autoregressive-1.5B-2-GGUF/resolve/main/Autoregressive-1.5B-2.IQ4_XS.gguf) | IQ4_XS | 1.1 | |
| [GGUF](https://huggingface.co/mradermacher/Autoregressive-1.5B-2-GGUF/resolve/main/Autoregressive-1.5B-2.Q4_K_S.gguf) | Q4_K_S | 1.2 | fast, recommended |
| [GGUF](https://huggingface.co/mradermacher/Autoregressive-1.5B-2-GGUF/resolve/main/Autoregressive-1.5B-2.Q4_K_M.gguf) | Q4_K_M | 1.2 | fast, recommended |
| [GGUF](https://huggingface.co/mradermacher/Autoregressive-1.5B-2-GGUF/resolve/main/Autoregressive-1.5B-2.Q5_K_S.gguf) | Q5_K_S | 1.4 | |
| [GGUF](https://huggingface.co/mradermacher/Autoregressive-1.5B-2-GGUF/resolve/main/Autoregressive-1.5B-2.Q5_K_M.gguf) | Q5_K_M | 1.4 | |
| [GGUF](https://huggingface.co/mradermacher/Autoregressive-1.5B-2-GGUF/resolve/main/Autoregressive-1.5B-2.Q6_K.gguf) | Q6_K | 1.6 | very good quality |
| [GGUF](https://huggingface.co/mradermacher/Autoregressive-1.5B-2-GGUF/resolve/main/Autoregressive-1.5B-2.Q8_0.gguf) | Q8_0 | 2.0 | fast, best quality |
| [GGUF](https://huggingface.co/mradermacher/Autoregressive-1.5B-2-GGUF/resolve/main/Autoregressive-1.5B-2.f16.gguf) | f16 | 3.7 | 16 bpw, overkill |
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.
<!-- end -->
|
hopstops/blockassist-bc-hulking_feathered_lemur_1756431875
|
hopstops
| 2025-08-29T01:45:18Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"hulking feathered lemur",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T01:45:00Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- hulking feathered lemur
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
hopstops/blockassist-bc-hulking_feathered_lemur_1756431487
|
hopstops
| 2025-08-29T01:39:27Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"hulking feathered lemur",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T01:38:32Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- hulking feathered lemur
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
ratri-apu-viral-video-original-Clip/New.full.videos.ratri.apu.Viral.Video.Official.Tutorial
|
ratri-apu-viral-video-original-Clip
| 2025-08-29T01:39:20Z | 0 | 0 | null |
[
"region:us"
] | null | 2025-08-29T01:39:07Z |
<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>
|
apple/mobileclip2_coca_dfn2b_s13b_recap-coco-30k_s12m_context77
|
apple
| 2025-08-29T01:37:34Z | 0 | 0 |
mobileclip
|
[
"mobileclip",
"arxiv:2508.20691",
"arxiv:2103.00020",
"arxiv:2303.15343",
"arxiv:2309.17425",
"license:apple-amlr",
"region:us"
] | null | 2025-08-25T19:13:37Z |
---
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_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_gbc10m-short-relation_context256
|
apple
| 2025-08-29T01:37:23Z | 0 | 0 |
mobileclip
|
[
"mobileclip",
"arxiv:2508.20691",
"arxiv:2103.00020",
"arxiv:2303.15343",
"arxiv:2309.17425",
"license:apple-amlr",
"region:us"
] | null | 2025-08-25T19:06:47Z |
---
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_docci_s12m_context128
|
apple
| 2025-08-29T01:37:17Z | 0 | 0 |
mobileclip
|
[
"mobileclip",
"arxiv:2508.20691",
"arxiv:2103.00020",
"arxiv:2303.15343",
"arxiv:2309.17425",
"license:apple-amlr",
"region:us"
] | null | 2025-08-25T19:02:16Z |
---
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_context77
|
apple
| 2025-08-29T01:37:11Z | 0 | 0 |
mobileclip
|
[
"mobileclip",
"arxiv:2508.20691",
"arxiv:2103.00020",
"arxiv:2303.15343",
"arxiv:2309.17425",
"license:apple-amlr",
"region:us"
] | null | 2025-08-25T18:57:32Z |
---
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)
```
|
kojeklollipop/blockassist-bc-spotted_amphibious_stork_1756429532
|
kojeklollipop
| 2025-08-29T01:33:04Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"spotted amphibious stork",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T01:33:00Z |
---
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).
|
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]
|
mradermacher/wargames-ai-GGUF
|
mradermacher
| 2025-08-29T01:30:42Z | 0 | 0 |
transformers
|
[
"transformers",
"gguf",
"en",
"base_model:postworthy/wargames-ai",
"base_model:quantized:postworthy/wargames-ai",
"endpoints_compatible",
"region:us",
"conversational"
] | null | 2025-08-29T00:25:01Z |
---
base_model: postworthy/wargames-ai
language:
- en
library_name: transformers
mradermacher:
readme_rev: 1
quantized_by: mradermacher
---
## About
<!-- ### quantize_version: 2 -->
<!-- ### output_tensor_quantised: 1 -->
<!-- ### convert_type: hf -->
<!-- ### vocab_type: -->
<!-- ### tags: -->
<!-- ### quants: x-f16 Q4_K_S Q2_K Q8_0 Q6_K Q3_K_M Q3_K_S Q3_K_L Q4_K_M Q5_K_S Q5_K_M IQ4_XS -->
<!-- ### quants_skip: -->
<!-- ### skip_mmproj: -->
static quants of https://huggingface.co/postworthy/wargames-ai
<!-- provided-files -->
***For a convenient overview and download list, visit our [model page for this model](https://hf.tst.eu/model#wargames-ai-GGUF).***
weighted/imatrix quants seem not to be available (by me) at this time. If they do not show up a week or so after the static ones, I have probably not planned for them. Feel free to request them by opening a Community Discussion.
## 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/wargames-ai-GGUF/resolve/main/wargames-ai.Q2_K.gguf) | Q2_K | 1.5 | |
| [GGUF](https://huggingface.co/mradermacher/wargames-ai-GGUF/resolve/main/wargames-ai.Q3_K_S.gguf) | Q3_K_S | 1.6 | |
| [GGUF](https://huggingface.co/mradermacher/wargames-ai-GGUF/resolve/main/wargames-ai.Q3_K_M.gguf) | Q3_K_M | 1.8 | lower quality |
| [GGUF](https://huggingface.co/mradermacher/wargames-ai-GGUF/resolve/main/wargames-ai.Q3_K_L.gguf) | Q3_K_L | 1.9 | |
| [GGUF](https://huggingface.co/mradermacher/wargames-ai-GGUF/resolve/main/wargames-ai.IQ4_XS.gguf) | IQ4_XS | 1.9 | |
| [GGUF](https://huggingface.co/mradermacher/wargames-ai-GGUF/resolve/main/wargames-ai.Q4_K_S.gguf) | Q4_K_S | 2.0 | fast, recommended |
| [GGUF](https://huggingface.co/mradermacher/wargames-ai-GGUF/resolve/main/wargames-ai.Q4_K_M.gguf) | Q4_K_M | 2.1 | fast, recommended |
| [GGUF](https://huggingface.co/mradermacher/wargames-ai-GGUF/resolve/main/wargames-ai.Q5_K_S.gguf) | Q5_K_S | 2.4 | |
| [GGUF](https://huggingface.co/mradermacher/wargames-ai-GGUF/resolve/main/wargames-ai.Q5_K_M.gguf) | Q5_K_M | 2.4 | |
| [GGUF](https://huggingface.co/mradermacher/wargames-ai-GGUF/resolve/main/wargames-ai.Q6_K.gguf) | Q6_K | 2.7 | very good quality |
| [GGUF](https://huggingface.co/mradermacher/wargames-ai-GGUF/resolve/main/wargames-ai.Q8_0.gguf) | Q8_0 | 3.5 | fast, best quality |
| [GGUF](https://huggingface.co/mradermacher/wargames-ai-GGUF/resolve/main/wargames-ai.f16.gguf) | f16 | 6.5 | 16 bpw, overkill |
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.
<!-- end -->
|
apple/MobileCLIP2-L-14
|
apple
| 2025-08-29T01:30:16Z | 0 | 0 |
mobileclip
|
[
"mobileclip",
"arxiv:2508.20691",
"arxiv:2103.00020",
"arxiv:2303.15343",
"arxiv:2309.17425",
"license:apple-amlr",
"region:us"
] | null | 2025-08-19T20:08:18Z |
---
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 **MobileCLIP2-L-14** 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-L-14
```
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-L-14', pretrained='/path/to/mobileclip2_l14.pt')
tokenizer = open_clip.get_tokenizer('MobileCLIP2-L-14')
# 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)
```
|
apishi/Qwen3-0.6B-Gensyn-Swarm-miniature_snorting_termite
|
apishi
| 2025-08-29T01:27:50Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"qwen3",
"text-generation",
"rl-swarm",
"genrl-swarm",
"grpo",
"gensyn",
"I am miniature_snorting_termite",
"arxiv:1910.09700",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-08-28T23:34:37Z |
---
library_name: transformers
tags:
- rl-swarm
- genrl-swarm
- grpo
- gensyn
- I am miniature_snorting_termite
---
# 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/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)
```
|
Rustamshry/Philosophy-chat
|
Rustamshry
| 2025-08-29T01:20:09Z | 3 | 1 |
peft
|
[
"peft",
"safetensors",
"base_model:adapter:unsloth/Qwen2.5-1.5B-Instruct",
"lora",
"sft",
"transformers",
"trl",
"unsloth",
"text-generation",
"conversational",
"en",
"dataset:Heigke/stanford-enigma-philosophy-chat",
"base_model:unsloth/Qwen2.5-1.5B-Instruct",
"license:mit",
"region:us"
] |
text-generation
| 2025-08-26T23:54:33Z |
---
base_model: unsloth/Qwen2.5-1.5B-Instruct
library_name: peft
pipeline_tag: text-generation
tags:
- base_model:adapter:unsloth/Qwen2.5-1.5B-Instruct
- lora
- sft
- transformers
- trl
- unsloth
license: mit
datasets:
- Heigke/stanford-enigma-philosophy-chat
language:
- en
---
# Model Card for Philosophy-chat
Philosophy-chat is a fine-tuned version of Qwen2.5-1.5B-Instruct, trained specifically on philosophical texts. The model specializes in understanding and generating responses related to complex philosophical concepts, arguments, and debates.
## Model Details
### Model Description
- **Language:** English
- **License:** MIT
- **Finetuned from model:** unsloth/Qwen2.5-1.5B-Instruct
- **Fine-Tuning Method**: Supervised Fine-tuning with LoRA
- **Domain**: Philosophy
- **Dataset**: Heigke/stanford-enigma-philosophy-chat
## Uses
### Direct Use
- Generating clear and concise explanations of philosophical concepts.
- Providing structured responses to philosophical questions.
- Assisting students, researchers, and enthusiasts in exploring philosophical arguments.
## Bias, Risks, and Limitations
- While fine-tuned on philosophy, the model may still occasionally generate hallucinations or less precise interpretations of highly nuanced philosophical arguments.
- The model does not replace expert human philosophical judgment.
## How 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/Qwen2.5-1.5B-Instruct",)
base_model = AutoModelForCausalLM.from_pretrained(
"unsloth/Qwen2.5-1.5B-Instruct",
device_map={"": 0}, token=""
)
model = PeftModel.from_pretrained(base_model,"Rustamshry/Philosophy-chat")
question = "According to William Whewell, what is necessary for gaining knowledge?"
system = """
You are an expert in philosophy.
"""
messages = [
{"role" : "system", "content" : system},
{"role" : "user", "content" : question}
]
text = tokenizer.apply_chat_template(
messages,
tokenize = False,
)
from transformers import TextStreamer
_ = model.generate(
**tokenizer(text, return_tensors = "pt").to("cuda"),
max_new_tokens = 1024,
streamer = TextStreamer(tokenizer, skip_prompt = True),
)
```
## Training Details
### Training Data
Roughly 27k questions and answers inspired by articles from Stanford Encyclopedia of Philosophy.
The questions range all the way from Zombies to the concept of Abduction, from Metaphysics to Neuroethics and thus cover some of the essence of mathematics, logic and philosophy.
### Framework versions
- PEFT 0.17.0
|
eduardoferrari135/bert_classifier_curadobia
|
eduardoferrari135
| 2025-08-29T01:17:54Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"bert",
"text-classification",
"arxiv:1910.09700",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2025-08-28T13:24:03Z |
---
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]
|
Rustamshry/Qwen3-1.7B-finance-reasoning
|
Rustamshry
| 2025-08-29T01:13:59Z | 168 | 1 |
peft
|
[
"peft",
"safetensors",
"finance",
"transformers",
"unsloth",
"trl",
"text-generation",
"conversational",
"en",
"dataset:Akhil-Theerthala/PersonalFinance_v2",
"base_model:unsloth/Qwen3-1.7B",
"base_model:adapter:unsloth/Qwen3-1.7B",
"license:mit",
"region:us"
] |
text-generation
| 2025-05-25T02:20:29Z |
---
base_model: unsloth/Qwen3-1.7B
library_name: peft
license: mit
datasets:
- Akhil-Theerthala/PersonalFinance_v2
language:
- en
pipeline_tag: text-generation
tags:
- finance
- transformers
- unsloth
- trl
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
This model is fine-tuned for instruction-following in the domain of personal finance, with a focus on:
- Budgeting advice
- Investment strategies
- Credit management
- Retirement planning
- Insurance and financial planning concepts
- Personalized financial reasoning
### Model Description
- **License:** MIT
- **Finetuned from model:** unsloth/Qwen3-1.7B
- **Dataset:** The model was fine-tuned on the PersonalFinance_v2 dataset, curated and published by Akhil-Theerthala.
### Model Capabilities
- Understands and provides contextual financial advice based on user queries.
- Responds in a chat-like conversational format.
- Trained to follow multi-turn instructions and deliver clear, structured, and accurate financial reasoning.
- Generalizes well to novel personal finance questions and explanations.
## Uses
### Direct Use
- Chatbots for personal finance
- Educational assistants for financial literacy
- Decision support for simple financial planning
- Interactive personal finance Q&A systems
## Bias, Risks, and Limitations
- Not a substitute for licensed financial advisors.
- The model's advice is based on training data and may not reflect region-specific laws, regulations, or financial products.
- May occasionally hallucinate or give generic responses in ambiguous scenarios.
- Assumes user input is well-formed and relevant to personal finance.
## 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/Qwen3-1.7B-finance-reasoning")
question =
"""
$19k for a coding bootcamp
Hi!
I was just accepted into the full-time software engineering program with Flatiron and have approx. $0 to my name.
I know I can get a loan with either Climb or accent with around 6.50% interest, is this a good option?
I would theoretically be paying near $600/month.
I really enjoy coding and would love to start a career in tech but the potential $19k price tag is pretty scary. Any advice?
"""
messages = [
{"role" : "user", "content" : question}
]
text = tokenizer.apply_chat_template(
messages,
tokenize = False,
add_generation_prompt = True,
enable_thinking = True,
)
from transformers import TextStreamer
_ = model.generate(
**tokenizer(text, return_tensors = "pt").to("cuda"),
max_new_tokens = 2048,
temperature = 0.6,
top_p = 0.95,
top_k = 20,
streamer = TextStreamer(tokenizer, skip_prompt = True),
)
```
## Training Details
### Training Data
- Dataset Overview:
PersonalFinance_v2 is a collection of high-quality instruction-response pairs focused on personal finance topics.
It covers a wide range of subjects including budgeting, saving, investing, credit management, retirement planning, insurance, and financial literacy.
- Data Format:
The dataset consists of conversational-style prompts paired with detailed and well-structured responses.
It is formatted to enable instruction-following language models to understand and generate coherent financial advice and reasoning.
### Framework versions
- PEFT 0.14.0
|
Rustamshry/Qwen3-0.6B-turkish-math-reasoning-80K
|
Rustamshry
| 2025-08-29T01:11:37Z | 14 | 2 |
peft
|
[
"peft",
"safetensors",
"math",
"transformers",
"unsloth",
"trl",
"sft",
"text-generation",
"conversational",
"tr",
"dataset:ituperceptron/turkish-math-186k",
"base_model:unsloth/Qwen3-0.6B",
"base_model:adapter:unsloth/Qwen3-0.6B",
"license:mit",
"region:us"
] |
text-generation
| 2025-06-01T00:35:37Z |
---
base_model: unsloth/Qwen3-0.6B
library_name: peft
license: mit
datasets:
- ituperceptron/turkish-math-186k
language:
- tr
pipeline_tag: text-generation
tags:
- math
- transformers
- unsloth
- trl
- sft
---
# Model Card for Model ID
This model was fine-tuned on 80,000 Turkish math problems, targeting better understanding and generation of mathematically structured responses in Turkish.
The dataset covers arithmetic, algebra, word problems, and other foundational math skills, allowing the model to serve as a multilingual math tutor or reasoning engine in Turkish.
## Model Details
### Model Description
- **Language(s) (NLP):** Turkish
- **License:** MIT
- **Finetuned from model:** unsloth/Qwen3-0.6B
- **Domain**: Mathematical Reasoning
## Uses
### Direct Use
- 🧮 Math problem solving in Turkish
Can assist users in understanding and solving elementary to intermediate math problems written in Turkish.
- 📚 Educational tools and tutoring systems
Suitable for integration into digital tutors, math practice apps, or classroom AI assistants for Turkish-speaking students.
- 💬 Multilingual reasoning research
Can be used to evaluate Turkish-language mathematical reasoning tasks in LLM benchmarks.
## Bias, Risks, and Limitations
🌐 Language bias
Performance is limited to Turkish; multilingual or code-mixed input may confuse the model.
## 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/Qwen3-0.6B",)
base_model = AutoModelForCausalLM.from_pretrained(
"unsloth/Qwen3-0.6B",
device_map={"": 0}, token=""
)
model = PeftModel.from_pretrained(base_model,"Rustamshry/Qwen3-0.6B-turkish-math-reasoning-80K")
question = """
Problem 2. $a, b$ iki farklı gerçel sayı ve $c$ öyle bir pozitif gerçel sayı olsun ki
$$ a^{4}-2019 a=b^{4}-2019 b=c. $$ $-\sqrt{c}<a b<0$ olduğunu kanıtlayın.
"""
messages = [
{"role" : "user", "content" : question}
]
text = tokenizer.apply_chat_template(
messages,
tokenize = False,
add_generation_prompt = True,
enable_thinking = True,
)
from transformers import TextStreamer
_ = model.generate(
**tokenizer(text, return_tensors = "pt").to("cuda"),
max_new_tokens = 3000,
temperature = 0.6,
top_p = 0.95,
top_k = 20,
streamer = TextStreamer(tokenizer, skip_prompt = True),
)
```
### Training Data
The dataset ituperceptron/turkish-math-186k is a Turkish-language collection designed for training and evaluating language models on mathematical reasoning tasks.
It comprises approximately 186,000 entries, each containing structured fields such as instruction, input, and output.
The dataset is available in Parquet format and is intended for text generation tasks, particularly focusing on mathematical problem-solving in Turkish.
### Framework versions
- PEFT 0.14.0
|
Rustamshry/ITA-Reasoning-o1
|
Rustamshry
| 2025-08-29T01:08:50Z | 6 | 1 |
peft
|
[
"peft",
"safetensors",
"text-generation",
"conversational",
"it",
"dataset:DeepMount00/o1-ITA-REASONING",
"base_model:unsloth/Qwen3-4B",
"base_model:adapter:unsloth/Qwen3-4B",
"license:mit",
"region:us"
] |
text-generation
| 2025-05-25T23:32:22Z |
---
base_model: unsloth/Qwen3-4B
library_name: peft
license: mit
datasets:
- DeepMount00/o1-ITA-REASONING
language:
- it
pipeline_tag: text-generation
---
# Model Card for Model ID
### Model Description
- **Training objective**: Fine-tuned on Italian instruction-style reasoning dataset for better performance in logical, educational, and chain-of-thought tasks.
- **Language(s) (NLP):** Italian
- **License:** MIT
- **Finetuned from model:** unsloth/Qwen3-4B
## Uses
### Direct Use
This model is intended for reasoning-intensive tasks in Italian
## Bias, Risks, and Limitations
- May hallucinate or make factual errors in complex logic chains.
- Not safe for unsupervised use in high-stakes domains like medical/legal reasoning.
- Output quality depends on instruction clarity.
# Training Data
The DeepMount00/o1-ITA-REASONING dataset is crafted to train language models in providing structured, methodical responses to questions in Italian.
Each entry follows a four-step reasoning approach:
- Reasoning: Initial thought process
- Verification: Self-review of the reasoning
- Correction: Amendments if needed
- Final Answer: Conclusive response
The dataset is formatted using XML-like tags to delineate each component, promoting transparency and structured thinking.
It is particularly beneficial for educational purposes, encouraging systematic problem-solving and critical thinking in the Italian language.
## 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/Qwen3-4B",)
base_model = AutoModelForCausalLM.from_pretrained(
"unsloth/Qwen3-4B",
device_map={"": 0}, token=""
)
model = PeftModel.from_pretrained(base_model,"Rustamshry/ITA-Reasoning-o1")
question = "Quali sono i costi e i benefici ambientali, sociali ed economici dell'energia solare?"
messages = [
{"role" : "user", "content" : question}
]
text = tokenizer.apply_chat_template(
messages,
tokenize = False,
add_generation_prompt = True, # Must add for generation
enable_thinking = True, # Disable thinking
)
from transformers import TextStreamer
_ = model.generate(
**tokenizer(text, return_tensors = "pt").to("cuda"),
max_new_tokens = 2048,
temperature = 0.6,
top_p = 0.95,
top_k = 20,
streamer = TextStreamer(tokenizer, skip_prompt = True),
)
```
### Framework versions
- PEFT 0.14.0
|
Rustamshry/MentalChat-16K
|
Rustamshry
| 2025-08-29T01:07:51Z | 18 | 1 |
peft
|
[
"peft",
"safetensors",
"medical",
"unsloth",
"trl",
"transformers",
"text-generation",
"conversational",
"en",
"dataset:ShenLab/MentalChat16K",
"base_model:unsloth/Qwen2.5-0.5B-Instruct",
"base_model:adapter:unsloth/Qwen2.5-0.5B-Instruct",
"license:mit",
"region:us"
] |
text-generation
| 2025-05-26T17:14:29Z |
---
base_model: unsloth/Qwen2.5-0.5B-Instruct
library_name: peft
license: mit
datasets:
- ShenLab/MentalChat16K
language:
- en
pipeline_tag: text-generation
tags:
- medical
- unsloth
- trl
- transformers
---
# Model Card for Model MentalChat-16K
## Model Details
This model is a fine-tuned version of Qwen2.5-0.5B-Instruct, optimized for empathetic and supportive conversations in the mental health domain.
It was trained on the ShenLab/MentalChat16K dataset, which includes over 16,000 counseling-style Q&A examples, combining real clinical paraphrases and synthetic mental health dialogues.
The model is designed to understand and respond to emotionally nuanced prompts related to stress, anxiety, relationships, and personal well-being.
### Model Description
- **Language(s) (NLP):** English
- **License:** MIT
- **Finetuned from model:** unsloth/Qwen2.5-0.5B-Instruct
- **Dataset:** ShenLab/MentalChat16K
## Uses
This model is intended for research and experimentation in AI-driven mental health support. Key use cases include:
- Mental health chatbot prototypes
- Empathy-focused dialogue agents
- Benchmarking LLMs on emotional intelligence and counseling-style prompts
- Educational or training tools in psychology or mental health communication
This model is NOT intended for clinical diagnosis, therapy, or real-time intervention. It must not replace licensed mental health professionals.
## Bias, Risks, and Limitations
- Biases:
- The real interview data is biased toward caregivers (mostly White, female, U.S.-based), which may affect the model’s cultural and demographic generalizability.
- The synthetic dialogues are generated by GPT-3.5, which may introduce linguistic and cultural biases from its pretraining.
- Limitations:
- The base model, Qwen2.5-0.5B-Instruct, is a small model (0.5B parameters), limiting depth of reasoning and nuanced understanding.
- Not suitable for handling acute mental health crises or emergency counseling.
- Responses may lack therapeutic rigor or miss subtle psychological cues.
- May produce hallucinated or inaccurate mental health advice.
## How to Get Started with the Model
Use the code below to get started with the model.
```python
from huggingface_hub import notebook_login,login
from transformers import AutoTokenizer, AutoModelForCausalLM
from peft import PeftModel
login(token="")
tokenizer = AutoTokenizer.from_pretrained("unsloth/Qwen2.5-0.5B-Instruct",)
base_model = AutoModelForCausalLM.from_pretrained(
"unsloth/Qwen2.5-0.5B-Instruct",
device_map={"": 0}, token=""
)
model = PeftModel.from_pretrained(base_model,"Rustamshry/MentalChat-16K")
instruction = """
You are a helpful mental health counselling assistant, please answer the mental health questions based on the patient's description.
The assistant gives helpful, comprehensive, and appropriate answers to the user's questions.
"""
question = """
I've tried setting boundaries, but it feels like I'm constantly being pulled in different directions.
I feel guilty for not being able to help my siblings, but I also know that I can't continue to neglect my mom's needs.
I'm worried that if I don't find a way to manage these demands, I'll burn out and won't be able to care for her effectively.
"""
prompt = (
f"### Instruction:\n{instruction}\n\n"
f"### Question:\n{question}\n\n"
f"### Response:\n"
)
input_ids = tokenizer(prompt, return_tensors="pt").to(model.device)
outputs = model.generate(
**input_ids,
max_new_tokens=4048,
#temperature=0.6,
#top_p=0.95,
#do_sample=True,
#eos_token_id=tokenizer.eos_token_id
)
print(tokenizer.decode(outputs[0]),skip_special_tokens=True)
```
### Framework versions
- PEFT 0.15.2
|
haritsondavid/blockassist-bc-padded_howling_anteater_1756429534
|
haritsondavid
| 2025-08-29T01:06:26Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"padded howling anteater",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T01:06:01Z |
---
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).
|
mizutoukotori/pi0_so101_v2
|
mizutoukotori
| 2025-08-29T01:05:10Z | 0 | 0 |
lerobot
|
[
"lerobot",
"safetensors",
"pi0",
"robotics",
"dataset:mizutoukotori/pick_the_yellow_block",
"arxiv:2410.24164",
"license:apache-2.0",
"region:us"
] |
robotics
| 2025-08-29T00:58:38Z |
---
datasets: mizutoukotori/pick_the_yellow_block
library_name: lerobot
license: apache-2.0
model_name: pi0
pipeline_tag: robotics
tags:
- pi0
- lerobot
- robotics
---
# Model Card for pi0
<!-- Provide a quick summary of what the model is/does. -->
[Pi0](https://huggingface.co/papers/2410.24164) is a generalist vision-language-action transformer that converts multimodal observations and text instructions into robot actions for zero-shot task transfer.
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
|
k1000dai/residualact_libero_smolvla_spatial
|
k1000dai
| 2025-08-29T01:03:24Z | 8 | 0 |
lerobot
|
[
"lerobot",
"safetensors",
"residualact",
"robotics",
"dataset:k1000dai/libero-spatial-smolvla",
"license:apache-2.0",
"region:us"
] |
robotics
| 2025-08-22T11:28:07Z |
---
datasets: k1000dai/libero-spatial-smolvla
library_name: lerobot
license: apache-2.0
model_name: residualact
pipeline_tag: robotics
tags:
- residualact
- lerobot
- robotics
---
# Model Card for residualact
<!-- Provide a quick summary of what the model is/does. -->
_Model type not recognized — please update this template._
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
|
mradermacher/MATH-TTT-llama3.2-TTRL-GGUF
|
mradermacher
| 2025-08-29T01:00:52Z | 0 | 0 |
transformers
|
[
"transformers",
"gguf",
"en",
"base_model:yujunzhou/MATH-TTT-llama3.2-TTRL",
"base_model:quantized:yujunzhou/MATH-TTT-llama3.2-TTRL",
"endpoints_compatible",
"region:us",
"conversational"
] | null | 2025-08-29T00:25:30Z |
---
base_model: yujunzhou/MATH-TTT-llama3.2-TTRL
language:
- en
library_name: transformers
mradermacher:
readme_rev: 1
quantized_by: mradermacher
---
## About
<!-- ### quantize_version: 2 -->
<!-- ### output_tensor_quantised: 1 -->
<!-- ### convert_type: hf -->
<!-- ### vocab_type: -->
<!-- ### tags: -->
<!-- ### quants: x-f16 Q4_K_S Q2_K Q8_0 Q6_K Q3_K_M Q3_K_S Q3_K_L Q4_K_M Q5_K_S Q5_K_M IQ4_XS -->
<!-- ### quants_skip: -->
<!-- ### skip_mmproj: -->
static quants of https://huggingface.co/yujunzhou/MATH-TTT-llama3.2-TTRL
<!-- provided-files -->
***For a convenient overview and download list, visit our [model page for this model](https://hf.tst.eu/model#MATH-TTT-llama3.2-TTRL-GGUF).***
weighted/imatrix quants seem not to be available (by me) at this time. If they do not show up a week or so after the static ones, I have probably not planned for them. Feel free to request them by opening a Community Discussion.
## 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/MATH-TTT-llama3.2-TTRL-GGUF/resolve/main/MATH-TTT-llama3.2-TTRL.Q2_K.gguf) | Q2_K | 1.6 | |
| [GGUF](https://huggingface.co/mradermacher/MATH-TTT-llama3.2-TTRL-GGUF/resolve/main/MATH-TTT-llama3.2-TTRL.Q3_K_S.gguf) | Q3_K_S | 1.8 | |
| [GGUF](https://huggingface.co/mradermacher/MATH-TTT-llama3.2-TTRL-GGUF/resolve/main/MATH-TTT-llama3.2-TTRL.Q3_K_M.gguf) | Q3_K_M | 2.0 | lower quality |
| [GGUF](https://huggingface.co/mradermacher/MATH-TTT-llama3.2-TTRL-GGUF/resolve/main/MATH-TTT-llama3.2-TTRL.Q3_K_L.gguf) | Q3_K_L | 2.1 | |
| [GGUF](https://huggingface.co/mradermacher/MATH-TTT-llama3.2-TTRL-GGUF/resolve/main/MATH-TTT-llama3.2-TTRL.IQ4_XS.gguf) | IQ4_XS | 2.2 | |
| [GGUF](https://huggingface.co/mradermacher/MATH-TTT-llama3.2-TTRL-GGUF/resolve/main/MATH-TTT-llama3.2-TTRL.Q4_K_S.gguf) | Q4_K_S | 2.2 | fast, recommended |
| [GGUF](https://huggingface.co/mradermacher/MATH-TTT-llama3.2-TTRL-GGUF/resolve/main/MATH-TTT-llama3.2-TTRL.Q4_K_M.gguf) | Q4_K_M | 2.3 | fast, recommended |
| [GGUF](https://huggingface.co/mradermacher/MATH-TTT-llama3.2-TTRL-GGUF/resolve/main/MATH-TTT-llama3.2-TTRL.Q5_K_S.gguf) | Q5_K_S | 2.6 | |
| [GGUF](https://huggingface.co/mradermacher/MATH-TTT-llama3.2-TTRL-GGUF/resolve/main/MATH-TTT-llama3.2-TTRL.Q5_K_M.gguf) | Q5_K_M | 2.7 | |
| [GGUF](https://huggingface.co/mradermacher/MATH-TTT-llama3.2-TTRL-GGUF/resolve/main/MATH-TTT-llama3.2-TTRL.Q6_K.gguf) | Q6_K | 3.1 | very good quality |
| [GGUF](https://huggingface.co/mradermacher/MATH-TTT-llama3.2-TTRL-GGUF/resolve/main/MATH-TTT-llama3.2-TTRL.Q8_0.gguf) | Q8_0 | 3.9 | fast, best quality |
| [GGUF](https://huggingface.co/mradermacher/MATH-TTT-llama3.2-TTRL-GGUF/resolve/main/MATH-TTT-llama3.2-TTRL.f16.gguf) | f16 | 7.3 | 16 bpw, overkill |
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.
<!-- end -->
|
NahedDom/blockassist-bc-flapping_stocky_leopard_1756426031
|
NahedDom
| 2025-08-29T00:42:35Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"flapping stocky leopard",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T00:42:32Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- flapping stocky leopard
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
Pomni/model_requests
|
Pomni
| 2025-08-29T00:40:01Z | 0 | 0 | null |
[
"region:us"
] | null | 2025-08-29T00:39:35Z |
Feel free to submit quantization requests for Whisper models here.
|
Pomni/whisper-large-ggml-allquants
|
Pomni
| 2025-08-29T00:38:39Z | 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",
"base_model:finetune:openai/whisper-large",
"license:apache-2.0",
"region:us"
] |
automatic-speech-recognition
| 2025-08-25T07:10:48Z |
---
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
pipeline_tag: automatic-speech-recognition
tags:
- whisper.cpp
- ggml
- whisper
- audio
- speech
- voice
new_version: Pomni/whisper-large-v3-ggml-allquants
---
# Whisper-Large quants
This is a repository of **GGML quants for [whisper-large](https://huggingface.co/openai/whisper-large)**, 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-ggml-allquants/resolve/main/ggml-large-f32.bin) | F32 | 6.17 GB | Likely overkill. |
| [GGML](https://huggingface.co/Pomni/whisper-large-ggml-allquants/resolve/main/ggml-large-f16.bin) | F16 | 3.09 GB | Performs better than Q8_0 for noisy audio and music. |
| [GGML](https://huggingface.co/Pomni/whisper-large-ggml-allquants/resolve/main/ggml-large-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-ggml-allquants/resolve/main/ggml-large-q6_k.bin) | Q6_K | 1.28 GB | |
| [GGML](https://huggingface.co/Pomni/whisper-large-ggml-allquants/resolve/main/ggml-large-q5_k.bin) | Q5_K | 1.08 GB | |
| [GGML](https://huggingface.co/Pomni/whisper-large-ggml-allquants/resolve/main/ggml-large-q5_1.bin) | Q5_1 | 1.18 GB | |
| [GGML](https://huggingface.co/Pomni/whisper-large-ggml-allquants/resolve/main/ggml-large-q5_0.bin) | Q5_0 | 1.08 GB | Last "good" quant; anything below loses quality rapidly. |
| [GGML](https://huggingface.co/Pomni/whisper-large-ggml-allquants/resolve/main/ggml-large-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-ggml-allquants/resolve/main/ggml-large-q4_1.bin) | Q4_1 | 985 MB | |
| [GGML](https://huggingface.co/Pomni/whisper-large-ggml-allquants/resolve/main/ggml-large-q4_0.bin) | Q4_0 | 889 MB | |
| [GGML](https://huggingface.co/Pomni/whisper-large-ggml-allquants/resolve/main/ggml-large-q3_k.bin) | Q3_K | 685 MB | |
| [GGML](https://huggingface.co/Pomni/whisper-large-ggml-allquants/resolve/main/ggml-large-q2_k.bin) | Q2_K | 529 MB | Completely non-sensical outputs. |
The F16 quant was taken from [ggerganov/whisper.cpp/ggml-large-v1.bin](https://huggingface.co/ggerganov/whisper.cpp/blob/main/ggml-large-v1.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-ggml-allquants/discussions/new) in the community tab about this, and I will look into the issue.
|
bah63843/blockassist-bc-plump_fast_antelope_1756427653
|
bah63843
| 2025-08-29T00:35:04Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"plump fast antelope",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T00:34:55Z |
---
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).
|
FAHAB/Qwen3-0.6B-Gensyn-Swarm-tame_reptilian_cockroach
|
FAHAB
| 2025-08-29T00:33:59Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"qwen3",
"text-generation",
"rl-swarm",
"genrl-swarm",
"grpo",
"gensyn",
"I am tame_reptilian_cockroach",
"arxiv:1910.09700",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-08-28T15:57:19Z |
---
library_name: transformers
tags:
- rl-swarm
- genrl-swarm
- grpo
- gensyn
- I am tame_reptilian_cockroach
---
# 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]
|
liukevin666/blockassist-bc-yawning_striped_cassowary_1756427258
|
liukevin666
| 2025-08-29T00:28:44Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"yawning striped cassowary",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T00:28:37Z |
---
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).
|
l933at/Qwen3-0.6B-Gensyn-Swarm-fluffy_alert_rooster
|
l933at
| 2025-08-29T00:28:10Z | 106 | 0 |
transformers
|
[
"transformers",
"safetensors",
"qwen3",
"text-generation",
"rl-swarm",
"genrl-swarm",
"grpo",
"gensyn",
"I am fluffy_alert_rooster",
"arxiv:1910.09700",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-08-20T20:56:59Z |
---
library_name: transformers
tags:
- rl-swarm
- genrl-swarm
- grpo
- gensyn
- I am fluffy_alert_rooster
---
# 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]
|
capungmerah627/blockassist-bc-stinging_soaring_porcupine_1756425764
|
capungmerah627
| 2025-08-29T00:27:45Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"stinging soaring porcupine",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T00:27:41Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- stinging soaring porcupine
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
michaelwaves/gpt-120b-fun-weights
|
michaelwaves
| 2025-08-29T00:23:48Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"generated_from_trainer",
"sft",
"trl",
"base_model:openai/gpt-oss-120b",
"base_model:finetune:openai/gpt-oss-120b",
"endpoints_compatible",
"region:us"
] | null | 2025-08-28T21:16:39Z |
---
base_model: openai/gpt-oss-120b
library_name: transformers
model_name: gpt-120b-fun-weights
tags:
- generated_from_trainer
- sft
- trl
licence: license
---
# Model Card for gpt-120b-fun-weights
This model is a fine-tuned version of [openai/gpt-oss-120b](https://huggingface.co/openai/gpt-oss-120b).
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="michaelwaves/gpt-120b-fun-weights", 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: 4.0.0
- Tokenizers: 0.21.4
## 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}}
}
```
|
maboranomet/blockassist-bc-lumbering_soft_macaw_1756426408
|
maboranomet
| 2025-08-29T00:14:10Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"lumbering soft macaw",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T00:13:52Z |
---
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).
|
PrunaAI/ByteDance-Seed-Seed-Coder-8B-Instruct-HQQ-8bit-smashed
|
PrunaAI
| 2025-08-29T00:13:43Z | 0 | 0 | null |
[
"llama",
"pruna-ai",
"base_model:ByteDance-Seed/Seed-Coder-8B-Instruct",
"base_model:finetune:ByteDance-Seed/Seed-Coder-8B-Instruct",
"region:us"
] | null | 2025-08-29T00:12:14Z |
---
thumbnail: "https://assets-global.website-files.com/646b351987a8d8ce158d1940/64ec9e96b4334c0e1ac41504_Logo%20with%20white%20text.svg"
base_model: ByteDance-Seed/Seed-Coder-8B-Instruct
metrics:
- memory_disk
- memory_inference
- inference_latency
- inference_throughput
- inference_CO2_emissions
- inference_energy_consumption
tags:
- pruna-ai
---
<!-- header start -->
<!-- 200823 -->
<div style="width: auto; margin-left: auto; margin-right: auto">
<a href="https://www.pruna.ai/" target="_blank" rel="noopener noreferrer">
<img src="banner.png" alt="PrunaAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</a>
</div>
<!-- header end -->
[](https://twitter.com/PrunaAI)
[](https://github.com/PrunaAI)
[](https://www.linkedin.com/company/93832878/admin/feed/posts/?feedType=following)
[](https://discord.gg/rskEr4BZJx)
# Simply make AI models cheaper, smaller, faster, and greener!
- Give a thumbs up if you like this model!
- Contact us and tell us which model to compress next [here](https://www.pruna.ai/contact).
- Request access to easily compress your *own* AI models [here](https://z0halsaff74.typeform.com/pruna-access?typeform-source=www.pruna.ai).
- Read the documentations to know more [here](https://pruna-ai-pruna.readthedocs-hosted.com/en/latest/)
- Join Pruna AI community on Discord [here](https://discord.gg/CP4VSgck) to share feedback/suggestions or get help.
## Results
**Frequently Asked Questions**
- ***How does the compression work?*** The model is compressed with hqq.
- ***How does the model quality change?*** The quality of the model output might vary compared to the base model.
- ***How is the model efficiency evaluated?*** These results were obtained with configuration described in `model/smash_config.json` and are obtained after a hardware warmup. The smashed model is directly compared to the original base model. Efficiency results may vary in other settings (e.g. other hardware, image size, batch size, ...). We recommend to directly run them in the use-case conditions to know if the smashed model can benefit you.
- ***What is the model format?*** We use safetensors.
- ***What calibration data has been used?*** If needed by the compression method, we used WikiText as the calibration data.
- ***What is the naming convention for Pruna Huggingface models?*** We take the original model name and append "turbo", "tiny", or "green" if the smashed model has a measured inference speed, inference memory, or inference energy consumption which is less than 90% of the original base model.
- ***How to compress my own models?*** You can request premium access to more compression methods and tech support for your specific use-cases [here](https://z0halsaff74.typeform.com/pruna-access?typeform-source=www.pruna.ai).
- ***What are "first" metrics?*** Results mentioning "first" are obtained after the first run of the model. The first run might take more memory or be slower than the subsequent runs due cuda overheads.
- ***What are "Sync" and "Async" metrics?*** "Sync" metrics are obtained by syncing all GPU processes and stop measurement when all of them are executed. "Async" metrics are obtained without syncing all GPU processes and stop when the model output can be used by the CPU. We provide both metrics since both could be relevant depending on the use-case. We recommend to test the efficiency gains directly in your use-cases.
## Setup
You can run the smashed model with these steps:
0. Check requirements from the original repo ByteDance-Seed/Seed-Coder-8B-Instruct installed. In particular, check python, cuda, and transformers versions.
1. Make sure that you have installed quantization related packages.
```bash
pip install hqq
```
2. Load & run the model.
```python
from transformers import AutoModelForCausalLM, AutoTokenizer
from pruna import PrunaModel
from hqq.engine.hf import HQQModelForCausalLM
from hqq.models.hf.base import AutoHQQHFModel
try:
model = HQQModelForCausalLM.from_quantized("PrunaAI/ByteDance-Seed-Seed-Coder-8B-Instruct-HQQ-8bit-smashed", device_map='auto')
except:
model = AutoHQQHFModel.from_quantized("PrunaAI/ByteDance-Seed-Seed-Coder-8B-Instruct-HQQ-8bit-smashed")
tokenizer = AutoTokenizer.from_pretrained("ByteDance-Seed/Seed-Coder-8B-Instruct")
input_ids = tokenizer("What is the color of prunes?,", return_tensors='pt').to(model.device)["input_ids"]
outputs = model.generate(input_ids, max_new_tokens=216)
tokenizer.decode(outputs[0])
```
## Configurations
The configuration info are in `smash_config.json`. This model has been smashed with pruna in version 0.2.4
## Credits & License
The license of the smashed model follows the license of the original model. Please check the license of the original model ByteDance-Seed/Seed-Coder-8B-Instruct before using this model which provided the base model. The license of `pruna` is [here](https://github.com/PrunaAI/pruna/blob/main/LICENSE) on GitHub.
## Want to compress other models?
- Contact us and tell us which model to compress next [here](https://www.pruna.ai/contact).
- Request access to easily compress your own AI models [here](https://z0halsaff74.typeform.com/pruna-access?typeform-source=www.pruna.ai).
|
sekirr/blockassist-bc-masked_tenacious_whale_1756426273
|
sekirr
| 2025-08-29T00:11:53Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"masked tenacious whale",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T00:11:49Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- masked tenacious whale
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
maxibillion1975/blockassist-bc-iridescent_squeaky_sandpiper_1756424486
|
maxibillion1975
| 2025-08-29T00:10:01Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"iridescent squeaky sandpiper",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T00:09:57Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- iridescent squeaky sandpiper
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
s190/Qwen2.5-0.5B-Instruct-Gensyn-Swarm-ravenous_sedate_impala
|
s190
| 2025-08-29T00:09:22Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"qwen2",
"text-generation",
"rl-swarm",
"genrl-swarm",
"grpo",
"gensyn",
"I am ravenous_sedate_impala",
"arxiv:1910.09700",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-08-28T17:13:24Z |
---
library_name: transformers
tags:
- rl-swarm
- genrl-swarm
- grpo
- gensyn
- I am ravenous_sedate_impala
---
# 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_1756426059
|
Dejiat
| 2025-08-29T00:08:01Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"savage unseen bobcat",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T00:08:00Z |
---
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).
|
motza0025/blockassist-bc-bold_swift_boar_1756424429
|
motza0025
| 2025-08-29T00:05:14Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"bold swift boar",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-29T00:04:55Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- bold swift boar
---
# 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).
|
kamizane/Llama-3.2-1B-Instruct-bnb-FT
|
kamizane
| 2025-08-28T23:57:13Z | 0 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"generated_from_trainer",
"trl",
"sft",
"base_model:meta-llama/Llama-3.2-1B-Instruct",
"base_model:finetune:meta-llama/Llama-3.2-1B-Instruct",
"endpoints_compatible",
"region:us"
] | null | 2025-08-28T18:31:43Z |
---
base_model: meta-llama/Llama-3.2-1B-Instruct
library_name: transformers
model_name: Llama-3.2-1B-Instruct-bnb-FT
tags:
- generated_from_trainer
- trl
- sft
licence: license
---
# Model Card for Llama-3.2-1B-Instruct-bnb-FT
This model is a fine-tuned version of [meta-llama/Llama-3.2-1B-Instruct](https://huggingface.co/meta-llama/Llama-3.2-1B-Instruct).
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="kamizane/Llama-3.2-1B-Instruct-bnb-FT", 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+cu126
- Datasets: 4.0.0
- Tokenizers: 0.21.4
## 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}}
}
```
|
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).
|
shreyess/stage2-175
|
shreyess
| 2025-08-28T23:52:18Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"qwen3",
"text-generation",
"conversational",
"arxiv:1910.09700",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-08-28T23:39:29Z |
---
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]
|
aldigobbler/thinker-1-adapter
|
aldigobbler
| 2025-08-28T23:48:35Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"text-generation-inference",
"unsloth",
"qwen3",
"trl",
"en",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] | null | 2025-08-28T23:48:14Z |
---
base_model: unsloth/qwen3-4b-thinking-2507-unsloth-bnb-4bit
tags:
- text-generation-inference
- transformers
- unsloth
- qwen3
- trl
license: apache-2.0
language:
- en
---
# Uploaded model
- **Developed by:** aldigobbler
- **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)
|
Thireus/Qwen3-4B-Thinking-2507-THIREUS-IQ3_XXS-SPECIAL_SPLIT
|
Thireus
| 2025-08-28T23:47:58Z | 0 | 0 | null |
[
"gguf",
"arxiv:2505.23786",
"license:mit",
"region:us"
] | null | 2025-08-28T23:45:17Z |
---
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! 🎉
|
sekirr/blockassist-bc-masked_tenacious_whale_1756424785
|
sekirr
| 2025-08-28T23:47:05Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"masked tenacious whale",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-28T23:47:01Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- masked tenacious whale
---
# 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_1756424669
|
bah63843
| 2025-08-28T23:45:20Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"plump fast antelope",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-28T23:45:10Z |
---
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).
|
seraphimzzzz/620878
|
seraphimzzzz
| 2025-08-28T23:42:51Z | 0 | 0 | null |
[
"region:us"
] | null | 2025-08-28T23:42:45Z |
[View on Civ Archive](https://civarchive.com/models/630457?modelVersionId=704828)
|
crystalline7/886124
|
crystalline7
| 2025-08-28T23:42:32Z | 0 | 0 | null |
[
"region:us"
] | null | 2025-08-28T23:42:26Z |
[View on Civ Archive](https://civarchive.com/models/875111?modelVersionId=979646)
|
ghostai1/internalRAGCX
|
ghostai1
| 2025-08-28T23:39:26Z | 0 | 0 | null |
[
"model",
"region:us"
] | null | 2025-05-02T03:01:42Z |
---
tags: [model]
---
# Internal RAG CX Data Preprocessing Demo
A robust data preprocessing pipeline for Retrieval-Augmented Generation (RAG) and Context-Augmented Generation (CAG) systems, deployed on Hugging Face as a Model repository (free tier). Built with over 5 years of AI expertise since 2020, this demo focuses on cleaning and preparing call center datasets for enterprise-grade CX applications in SaaS, HealthTech, FinTech, and eCommerce. It integrates advanced data wrangling with Pandas, ensuring high-quality FAQs for downstream RAG/CAG pipelines, and is compatible with Amazon SageMaker and Azure AI for scalable modeling.
## Technical Architecture
### Data Preprocessing Pipeline
The core of this demo is a comprehensive data preprocessing pipeline designed to clean raw call center datasets:
- **Data Ingestion**:
- Parses CSVs with `pd.read_csv`, using `io.StringIO` for embedded data, with explicit `quotechar` and `escapechar` to handle complex strings.
- Handles datasets with columns: `call_id`, `question`, `answer`, `language`.
- **Junk Data Cleanup**:
- **Null Handling**: Drops rows with missing `question` or `answer` using `df.dropna()`.
- **Duplicate Removal**: Eliminates redundant FAQs via `df[~df['question'].duplicated()]`.
- **Short Entry Filtering**: Excludes questions <10 chars or answers <20 chars with `df[(df['question'].str.len() >= 10) & (df['answer'].str.len() >= 20)]`.
- **Malformed Detection**: Uses regex (`[!?]{2,}|(Invalid|N/A)`) to filter invalid questions.
- **Standardization**: Normalizes text (e.g., "mo" to "month") and fills missing `language` with "en".
- **Output**:
- Generates `cleaned_call_center_faqs.csv` for downstream modeling.
- Provides cleanup stats: nulls removed, duplicates removed, short entries filtered, malformed entries detected.
### Enterprise-Grade Modeling Compatibility
The cleaned dataset is optimized for:
- **Amazon SageMaker**: Ready for training BERT-based models (e.g., `bert-base-uncased`) for intent classification or FAQ retrieval, deployable via SageMaker JumpStart.
- **Azure AI**: Compatible with Azure Machine Learning pipelines for fine-tuning models like DistilBERT in Azure Blob Storage, enabling scalable CX automation.
- **LLM Integration**: Supports fine-tuning LLMs (e.g., `distilgpt2`) for generative tasks, leveraging your FastAPI experience for API-driven inference.
## Performance Monitoring and Visualization
The demo includes a performance monitoring suite:
- **Processing Time Tracking**: Measures data ingestion, cleaning, and output times using `time.perf_counter()`, reported in milliseconds.
- **Cleanup Metrics**: Tracks the number of nulls, duplicates, short entries, and malformed entries removed.
- **Visualization**: Uses Matplotlib to plot a bar chart (`cleanup_stats.png`):
- Bars: Number of entries removed per category (Nulls, Duplicates, Short, Malformed).
- Palette: Professional muted colors for enterprise aesthetics.
## Gradio Interface for Interactive Demo
The demo is accessible via Gradio, providing an interactive data preprocessing experience:
- **Input**: Upload a sample call center CSV or use the embedded demo dataset.
- **Outputs**:
- **Cleaned Dataset**: Download `cleaned_call_center_faqs.csv`.
- **Cleanup Stats**: Detailed breakdown (e.g., “Cleaned FAQs: 6; removed 4 junk entries: 2 nulls, 1 duplicates, 1 short, 0 malformed”).
- **Performance Plot**: Visual metrics for processing time and cleanup stats.
- **Styling**: Custom dark theme CSS (`#2a2a2a` background, blue buttons) for a sleek, enterprise-ready UI.
## Setup
- Clone this repository to a Hugging Face Model repository (free tier, public).
- Add `requirements.txt` with dependencies (`gradio==4.44.0`, `pandas==2.2.3`, `matplotlib==3.9.2`, etc.).
- Upload `app.py` (includes embedded demo dataset for seamless deployment).
- Configure to run with Python 3.9+, CPU hardware (no GPU).
## Usage
- **Upload CSV**: Provide a call center CSV in the Gradio UI, or use the default demo dataset.
- **Output**:
- **Cleaned Dataset**: Download the processed `cleaned_call_center_faqs.csv`.
- **Cleanup Stats**: “Cleaned FAQs: 6; removed 4 junk entries: 2 nulls, 1 duplicates, 1 short, 0 malformed”.
- **Performance Plot**: Visual metrics for processing time and cleanup stats.
**Example**:
- **Input CSV**: Sample dataset with 10 FAQs, including 2 nulls, 1 duplicate, 1 short entry.
- **Output**:
- Cleaned Dataset: 6 FAQs in `cleaned_call_center_faqs.csv`.
- Cleanup Stats: “Cleaned FAQs: 6; removed 4 junk entries: 2 nulls, 1 duplicates, 1 short, 0 malformed”.
- Plot: Processing Time (Ingestion: 50ms, Cleaning: 30ms, Output: 10ms), Cleanup Stats (Nulls: 2, Duplicates: 1, Short: 1, Malformed: 0).
## Technical Details
**Stack**:
- **Pandas**: Data wrangling and preprocessing for call center CSVs.
- **Gradio**: Interactive UI for real-time data preprocessing demos.
- **Matplotlib**: Performance visualization with bar charts.
- **FastAPI Compatibility**: Designed with API-driven preprocessing in mind, leveraging your experience with FastAPI for scalable deployments.
**Free Tier Optimization**: Lightweight with CPU-only dependencies, no GPU required.
**Extensibility**: Ready for integration with RAG/CAG pipelines, and cloud deployments on AWS Lambda or Azure Functions.
## Purpose
This demo showcases expertise in data preprocessing for AI-driven CX automation, focusing on call center data quality. Built on over 5 years of experience in AI, data engineering, and enterprise-grade deployments, it demonstrates the power of Pandas-based data cleaning for RAG/CAG pipelines, making it ideal for advanced CX solutions in call center environments.
## Latest Update
**Status Update**: Configuration missing in update.ini for ghostai1/internalRAGCX: Expected sections InternalragcxUpdate and InternalragcxEmojis - May 28, 2025 📝
- - August 28, 2025 📝
- - August 26, 2025 📝
- - August 23, 2025 📝
- - August 21, 2025 📝
- - August 19, 2025 📝
- - August 18, 2025 📝
- - August 16, 2025 📝
- - August 15, 2025 📝
- - August 14, 2025 📝
- - August 13, 2025 📝
- - August 12, 2025 📝
- - August 11, 2025 📝
- - August 10, 2025 📝
- - August 09, 2025 📝
- - August 08, 2025 📝
- - August 07, 2025 📝
- - August 06, 2025 📝
- - August 05, 2025 📝
- - August 04, 2025 📝
- - August 03, 2025 📝
- - August 02, 2025 📝
- - August 01, 2025 📝
- - July 31, 2025 📝
- - July 30, 2025 📝
- - July 29, 2025 📝
- - July 28, 2025 📝
- - July 27, 2025 📝
- - July 26, 2025 📝
- - July 25, 2025 📝
- - July 24, 2025 📝
- - July 23, 2025 📝
- - July 22, 2025 📝
- - July 21, 2025 📝
- - July 20, 2025 📝
- - July 19, 2025 📝
- - July 18, 2025 📝
- - July 17, 2025 📝
- - July 16, 2025 📝
- - July 15, 2025 📝
- - July 14, 2025 📝
- - July 11, 2025 📝
- - July 10, 2025 📝
- - July 09, 2025 📝
- - July 08, 2025 📝
- - July 07, 2025 📝
- - July 06, 2025 📝
- - July 05, 2025 📝
- - July 04, 2025 📝
- - July 03, 2025 📝
- - July 02, 2025 📝
- - July 01, 2025 📝
- - June 30, 2025 📝
- - June 29, 2025 📝
- - June 28, 2025 📝
- - June 27, 2025 📝
- - June 26, 2025 📝
- - June 25, 2025 📝
- - June 24, 2025 📝
- - June 23, 2025 📝
- - June 22, 2025 📝
- - June 21, 2025 📝
- - June 20, 2025 📝
- - June 19, 2025 📝
- - June 18, 2025 📝
- - June 17, 2025 📝
- - June 16, 2025 📝
- - June 15, 2025 📝
- - June 14, 2025 📝
- - June 13, 2025 📝
- - June 12, 2025 📝
- - June 11, 2025 📝
- - June 10, 2025 📝
- - June 09, 2025 📝
- - June 08, 2025 📝
- - June 07, 2025 📝
- - June 06, 2025 📝
- - June 05, 2025 📝
- - June 04, 2025 📝
- - June 03, 2025 📝
- - June 02, 2025 📝
- - June 01, 2025 📝
- - May 31, 2025 📝
- - May 30, 2025 📝
- - May 29, 2025 📝
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Placeholder update text.
## Future Enhancements
- **Real-Time Streaming**: Add support for real-time data streaming from Kafka for live preprocessing.
- **FastAPI Deployment**: Expose preprocessing pipeline via FastAPI endpoints for production-grade use.
- **Advanced Validation**: Implement stricter data validation rules using machine learning-based outlier detection.
- **Cloud Integration**: Enhance compatibility with AWS Glue or Azure Data Factory for enterprise data pipelines.
**Website**: https://ghostainews.com/
**Discord**: https://discord.gg/BfA23aYz
|
mradermacher/037-rakuten-2.0-mini-instruct-1.5b-v2new-dpo405b-GGUF
|
mradermacher
| 2025-08-28T23:37:35Z | 0 | 0 |
transformers
|
[
"transformers",
"gguf",
"ja",
"en",
"base_model:shisa-ai/037-rakuten-2.0-mini-instruct-1.5b-v2new-dpo405b",
"base_model:quantized:shisa-ai/037-rakuten-2.0-mini-instruct-1.5b-v2new-dpo405b",
"license:apache-2.0",
"endpoints_compatible",
"region:us",
"conversational"
] | null | 2025-08-28T22:59:54Z |
---
base_model: shisa-ai/037-rakuten-2.0-mini-instruct-1.5b-v2new-dpo405b
language:
- ja
- en
library_name: transformers
license: apache-2.0
mradermacher:
readme_rev: 1
quantized_by: mradermacher
---
## About
<!-- ### quantize_version: 2 -->
<!-- ### output_tensor_quantised: 1 -->
<!-- ### convert_type: hf -->
<!-- ### vocab_type: -->
<!-- ### tags: -->
<!-- ### quants: x-f16 Q4_K_S Q2_K Q8_0 Q6_K Q3_K_M Q3_K_S Q3_K_L Q4_K_M Q5_K_S Q5_K_M IQ4_XS -->
<!-- ### quants_skip: -->
<!-- ### skip_mmproj: -->
static quants of https://huggingface.co/shisa-ai/037-rakuten-2.0-mini-instruct-1.5b-v2new-dpo405b
<!-- provided-files -->
***For a convenient overview and download list, visit our [model page for this model](https://hf.tst.eu/model#037-rakuten-2.0-mini-instruct-1.5b-v2new-dpo405b-GGUF).***
weighted/imatrix quants seem not to be available (by me) at this time. If they do not show up a week or so after the static ones, I have probably not planned for them. Feel free to request them by opening a Community Discussion.
## 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/037-rakuten-2.0-mini-instruct-1.5b-v2new-dpo405b-GGUF/resolve/main/037-rakuten-2.0-mini-instruct-1.5b-v2new-dpo405b.Q2_K.gguf) | Q2_K | 0.7 | |
| [GGUF](https://huggingface.co/mradermacher/037-rakuten-2.0-mini-instruct-1.5b-v2new-dpo405b-GGUF/resolve/main/037-rakuten-2.0-mini-instruct-1.5b-v2new-dpo405b.Q3_K_S.gguf) | Q3_K_S | 0.8 | |
| [GGUF](https://huggingface.co/mradermacher/037-rakuten-2.0-mini-instruct-1.5b-v2new-dpo405b-GGUF/resolve/main/037-rakuten-2.0-mini-instruct-1.5b-v2new-dpo405b.Q3_K_M.gguf) | Q3_K_M | 0.9 | lower quality |
| [GGUF](https://huggingface.co/mradermacher/037-rakuten-2.0-mini-instruct-1.5b-v2new-dpo405b-GGUF/resolve/main/037-rakuten-2.0-mini-instruct-1.5b-v2new-dpo405b.Q3_K_L.gguf) | Q3_K_L | 0.9 | |
| [GGUF](https://huggingface.co/mradermacher/037-rakuten-2.0-mini-instruct-1.5b-v2new-dpo405b-GGUF/resolve/main/037-rakuten-2.0-mini-instruct-1.5b-v2new-dpo405b.IQ4_XS.gguf) | IQ4_XS | 1.0 | |
| [GGUF](https://huggingface.co/mradermacher/037-rakuten-2.0-mini-instruct-1.5b-v2new-dpo405b-GGUF/resolve/main/037-rakuten-2.0-mini-instruct-1.5b-v2new-dpo405b.Q4_K_S.gguf) | Q4_K_S | 1.0 | fast, recommended |
| [GGUF](https://huggingface.co/mradermacher/037-rakuten-2.0-mini-instruct-1.5b-v2new-dpo405b-GGUF/resolve/main/037-rakuten-2.0-mini-instruct-1.5b-v2new-dpo405b.Q4_K_M.gguf) | Q4_K_M | 1.0 | fast, recommended |
| [GGUF](https://huggingface.co/mradermacher/037-rakuten-2.0-mini-instruct-1.5b-v2new-dpo405b-GGUF/resolve/main/037-rakuten-2.0-mini-instruct-1.5b-v2new-dpo405b.Q5_K_S.gguf) | Q5_K_S | 1.2 | |
| [GGUF](https://huggingface.co/mradermacher/037-rakuten-2.0-mini-instruct-1.5b-v2new-dpo405b-GGUF/resolve/main/037-rakuten-2.0-mini-instruct-1.5b-v2new-dpo405b.Q5_K_M.gguf) | Q5_K_M | 1.2 | |
| [GGUF](https://huggingface.co/mradermacher/037-rakuten-2.0-mini-instruct-1.5b-v2new-dpo405b-GGUF/resolve/main/037-rakuten-2.0-mini-instruct-1.5b-v2new-dpo405b.Q6_K.gguf) | Q6_K | 1.4 | very good quality |
| [GGUF](https://huggingface.co/mradermacher/037-rakuten-2.0-mini-instruct-1.5b-v2new-dpo405b-GGUF/resolve/main/037-rakuten-2.0-mini-instruct-1.5b-v2new-dpo405b.Q8_0.gguf) | Q8_0 | 1.7 | fast, best quality |
| [GGUF](https://huggingface.co/mradermacher/037-rakuten-2.0-mini-instruct-1.5b-v2new-dpo405b-GGUF/resolve/main/037-rakuten-2.0-mini-instruct-1.5b-v2new-dpo405b.f16.gguf) | f16 | 3.2 | 16 bpw, overkill |
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.
<!-- end -->
|
BasedBase/Qwen3-30B-A3B-Thinking-2507-Deepseek-v3.1-Distill
|
BasedBase
| 2025-08-28T23:37:06Z | 422 | 3 | null |
[
"gguf",
"causal-lm",
"moe",
"mixture-of-experts",
"qwen",
"distillation",
"svd",
"lora-merged",
"code-generation",
"base_model:Qwen/Qwen3-30B-A3B-Thinking-2507",
"base_model:quantized:Qwen/Qwen3-30B-A3B-Thinking-2507",
"license:apache-2.0",
"endpoints_compatible",
"region:us",
"conversational"
] | null | 2025-08-26T08:50:43Z |
---
license: apache-2.0
base_model:
- Qwen/Qwen3-30B-A3B-Thinking-2507
tags:
- causal-lm
- moe
- mixture-of-experts
- qwen
- distillation
- svd
- lora-merged
- code-generation
---
# Qwen3-30B-A3B-Thinking-2507-Deepseek-v3.1-Distill
## Model Description
This model is a distilled version of **`Qwen/Qwen3-30B-A3B-Instruct`** designed to inherit the reasoning and behavioral characteristics of its much larger teacher model, **`deepseek-ai/DeepSeek-V3.1`**.
It is the result of applying a LoRA created via an SVD-based distillation pipeline, and then merging those weights into the base model. The core of this process was to transfer the nuanced knowledge from a **62-layer, 256-expert teacher model** into the more efficient **48-layer, 128-expert architecture** of the student model.
The primary goal was to explore the high-fidelity transfer of complex reasoning patterns, particularly those encoded within the Mixture-of-Experts (MoE) layers, from a frontier-class model to a consumer-accessible one.
You should notice that the model has a more confident and linear chain-of-thought compared to the base qwen3-30b-a3b-thinking-2507 model like Deepseek 3.1 has. This distill tends to overthink much less than the base model and provides more accurate better structured answers.
## The Distillation Methodology
This model was not trained in a conventional sense. Instead, it was created using a layer-by-layer distillation SVD based distillation process.
### Core Components
* **Teacher Model:** `deepseek-ai/DeepSeek-V3.1`.
* **Student Model:** `Qwen/Qwen3-30B-A3B-Thinking-2507`.
* **LoRA Rank:** A high rank of **`r=2048`** was used for all modules to ensure a comprehensive capture of information from the teacher model.
### The Distillation Pipeline
For each corresponding layer in the student and teacher, the following pipeline was executed:
1. **Teacher Layer Interpolation (SLERP):** For student layers that fall between two teacher layers (based on a sigmoid mapping), Spherical Linear Interpolation (SLERP) was used to create a geometrically sound blend of the teacher's weights. This preserves the integrity of the high-dimensional representations.
2. **SVD Projection:** The core of the distillation. The (potentially blended) teacher layer's weight matrix was decomposed using a randomized SVD algorithm. The top 2048 most significant components were selected and reconstructed to fit the student layer's smaller dimensions. This high-rank projection is designed for maximum fidelity.
3. **Generalized Procrustes Analysis:** After projection, the newly created "synthetic" tensor was optimally aligned with the student's original pre-trained tensor using a hardened least-squares solver. This alignment minimizes representational distance before calculating the final difference, with added checks to prevent numerical instability.
4. **DARE-TIES Purification:** The difference tensor (`Distilled - Aligned Student`) was then purified using the DARE-TIES methodology. This process drops a significant percentage (80%) of the lowest-magnitude values, treating them as noise, and then rescale the remaining important differences. This creates a clean, high-signal delta for the final LoRA.
### Mixture-of-Experts (MoE) Distillation
The standout feature of this process is the full distillation of the MoE layers, which are critical for nuanced, context-dependent reasoning.
* **Expert Fingerprinting & Clustering:** To map the 256 teacher experts to the 128 student experts, each teacher expert was "fingerprinted" by concatenating its constituent weight matrices. **FAISS-GPU K-Means clustering** was then used to efficiently group these 256 fingerprints into 128 distinct clusters based on their geometric similarity.
* **Advanced Expert Synthesis:** Each of the student's 128 experts was synthesized from a weighted blend of the teacher experts assigned to its cluster. This blend is not a simple average; instead, it uses an SVD-based reconstruction from the top teacher experts (ranked by similarity to the cluster centroid) to create a new, synthetic expert that represents the core "concept" of that cluster. This more advanced synthesis aims to create novel, yet faithful, expert representations.
## Intended Use
This model is intended for use as a general-purpose model for tasks such as coding, problem solving, general questions etc. It is designed to be a more capable and nuanced reasoner than its base model.
* **Primary Use:** Complex instruction-following, reasoning tasks, and creative generation.
* **Out of Scope:** Its knowledge cutoff is from its original training (2024), and it has not been aligned for specific safety or conversational chatbot roles beyond its base tuning.
## Critical Usage Note
For inference, you can use either the default settings for the 30B model or the optimized settings used for the 685B model. The choice depends on your specific task,
Use the 30B defaults for general tasks.
For coding-related work, the 685B settings appear to yield significantly better results based on empirical testing but will slow down inference.
|
ultratopaz/1380319
|
ultratopaz
| 2025-08-28T23:36:31Z | 0 | 0 | null |
[
"region:us"
] | null | 2025-08-28T23:36:19Z |
[View on Civ Archive](https://civarchive.com/models/1310455?modelVersionId=1478916)
|
GroomerG/blockassist-bc-vicious_pawing_badger_1756422534
|
GroomerG
| 2025-08-28T23:33:44Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"vicious pawing badger",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-28T23:33:41Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- vicious pawing badger
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
chainway9/blockassist-bc-untamed_quick_eel_1756421672
|
chainway9
| 2025-08-28T23:23:02Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"untamed quick eel",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-28T23:22:59Z |
---
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).
|
EpistemeAI/gpt-oss-20b-finetune-multilanguage
|
EpistemeAI
| 2025-08-28T23:21:24Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"gpt_oss",
"text-generation",
"text-generation-inference",
"unsloth",
"conversational",
"en",
"base_model:unsloth/gpt-oss-20b-unsloth-bnb-4bit",
"base_model:finetune:unsloth/gpt-oss-20b-unsloth-bnb-4bit",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-08-28T21:53:52Z |
---
base_model: unsloth/gpt-oss-20b-unsloth-bnb-4bit
tags:
- text-generation-inference
- transformers
- unsloth
- gpt_oss
license: apache-2.0
language:
- en
---
# Uploaded finetuned model
- **Developed by:** EpistemeAI
- **License:** apache-2.0
- **Finetuned from model :** unsloth/gpt-oss-20b-unsloth-bnb-4bit
This gpt_oss 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)
|
samirmariam73/results
|
samirmariam73
| 2025-08-28T23:16:47Z | 0 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"gpt2",
"text-generation",
"generated_from_trainer",
"base_model:openai-community/gpt2",
"base_model:finetune:openai-community/gpt2",
"license:mit",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-08-28T23:16:33Z |
---
library_name: transformers
license: mit
base_model: gpt2
tags:
- generated_from_trainer
model-index:
- name: results
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. -->
# results
This model is a fine-tuned version of [gpt2](https://huggingface.co/gpt2) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 2.8304
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: 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: 3
### Training results
### Framework versions
- Transformers 4.55.4
- Pytorch 2.8.0+cu126
- Datasets 4.0.0
- Tokenizers 0.21.4
|
mozila80/blockassist-bc-colorful_huge_peacock_1756422465
|
mozila80
| 2025-08-28T23:08:50Z | 0 | 0 | null |
[
"gensyn",
"blockassist",
"gensyn-blockassist",
"minecraft",
"colorful huge peacock",
"arxiv:2504.07091",
"region:us"
] | null | 2025-08-28T23:08:28Z |
---
tags:
- gensyn
- blockassist
- gensyn-blockassist
- minecraft
- colorful huge peacock
---
# Gensyn BlockAssist
Gensyn's BlockAssist is a distributed extension of the paper [AssistanceZero: Scalably Solving Assistance Games](https://arxiv.org/abs/2504.07091).
|
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.