Instructions to use ymym-tymblack/aicrowd-model with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ymym-tymblack/aicrowd-model with PEFT:

from peft import PeftModel
from transformers import AutoModelForCausalLM

base_model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen3-4B-Instruct-2507")
model = PeftModel.from_pretrained(base_model, "ymym-tymblack/aicrowd-model")

Transformers

How to use ymym-tymblack/aicrowd-model with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="ymym-tymblack/aicrowd-model")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("ymym-tymblack/aicrowd-model", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use ymym-tymblack/aicrowd-model with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ymym-tymblack/aicrowd-model"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ymym-tymblack/aicrowd-model",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/ymym-tymblack/aicrowd-model

SGLang

How to use ymym-tymblack/aicrowd-model with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "ymym-tymblack/aicrowd-model" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ymym-tymblack/aicrowd-model",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "ymym-tymblack/aicrowd-model" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ymym-tymblack/aicrowd-model",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use ymym-tymblack/aicrowd-model with Docker Model Runner:
```
docker model run hf.co/ymym-tymblack/aicrowd-model
```

You need to agree to share your contact information to access this model

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

base_model: Qwen/Qwen3-4B-Instruct-2507 base_model_revision: cdbee75f17c01a7cc42f958dc650907174af0554 library_name: peft pipeline_tag: text-generation tags: - lora - peft - grpo - trl - transformers - orak - game-agent license: other

Model Card: aicrowd-model (LoRA Adapter: out_grpo)

This repository contains a LoRA adapter (out_grpo) for:

Base model: Qwen/Qwen3-4B-Instruct-2507
Base snapshot revision: cdbee75f17c01a7cc42f958dc650907174af0554

Important

This is an adapter-only repository. You must load it together with the base model above.

Reproducibility (Hugging Face hosted adapter)

1) Download LoRA adapter from Hugging Face

hf auth login
hf download ymym-tymblack/aicrowd-model --local-dir /tmp/aicrowd-model

### 2) Start vLLM server

uv run python -m vllm.entrypoints.openai.api_server \
  --model /root/.cache/huggingface/hub/models--Qwen--Qwen3-4B-Instruct-2507/snapshots/cdbee75f17c01a7cc42f958dc650907174af0554 \
  --host 127.0.0.1 \
  --port 8000 \
  --max-model-len 8192 \
  --enable-lora \
  --lora-modules out_grpo=/tmp/aicrowd-model \
  --served-model-name out_grpo

### 3) Run Orak (local)

export ORAK_VLLM_BASE_URL=http://127.0.0.1:8000/v1
export ORAK_VLLM_MODEL=out_grpo
export ORAK_VLLM_API_KEY=EMPTY
uv run python run.py

## Model Declaration (for final evaluation package)

- Model name: out_grpo
- Type: LoRA adapter
- Provider: Hugging Face (adapter), Qwen (base model)
- Base model: Qwen/Qwen3-4B-Instruct-2507
- Base revision: cdbee75f17c01a7cc42f958dc650907174af0554

## Training Summary

- Method: GRPO
- Framework: TRL + PEFT
- Intended use: Orak Game Agent Challenge 2025 evaluation/reproducibility

## Framework versions

- PEFT: 0.18.1
- TRL: 0.26.2
- Transformers: 4.57.3
- PyTorch: 2.9.1
- Datasets: 4.5.0
- Tokenizers: 0.22.1

- This repository contains adapter weights only.
- Base model license/terms from Qwen/Qwen3-4B-Instruct-2507 also apply.
- If you need strict commercial/legal confirmation, verify both base-model and adapter redistribution terms before use.

## Citations

### GRPO / DeepSeekMath

@article{shao2024deepseekmath,
  title        = {DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models},
  author       = {Shao, Zhihong and Wang, Peiyi and Zhu, Qihao and Xu, Runxin and Song, Junxiao and Zhang, Mingchuan and Li, Y. K. and Wu, Y. and Guo, Daya},
  year         = {2024},
  eprint       = {arXiv:2402.03300}
}

### TRL

@misc{vonwerra2022trl,
  title        = {TRL: Transformer Reinforcement Learning},
  author       = {von Werra, Leandro and Belkada, Younes and Tunstall, Lewis and Beeching, Edward and Thrush, Tristan and Lambert, Nathan and Huang, Shengyi and Rasul, Kashif and Gallouedec, Quentin},
  year         = {2020},
  howpublished = {\url{https://github.com/huggingface/trl}}
}

Downloads last month: -

Model tree for ymym-tymblack/aicrowd-model

Base model

Qwen/Qwen3-4B-Instruct-2507

Adapter

(5708)

this model

Paper for ymym-tymblack/aicrowd-model

DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models

Paper • 2402.03300 • Published Feb 5, 2024 • 150