Instructions to use xx18/Composition-RL-4B-Physics_Math with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use xx18/Composition-RL-4B-Physics_Math with Transformers:

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

pipe = pipeline("text-generation", model="xx18/Composition-RL-4B-Physics_Math")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("xx18/Composition-RL-4B-Physics_Math")
model = AutoModelForCausalLM.from_pretrained("xx18/Composition-RL-4B-Physics_Math")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use xx18/Composition-RL-4B-Physics_Math with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "xx18/Composition-RL-4B-Physics_Math"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "xx18/Composition-RL-4B-Physics_Math",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/xx18/Composition-RL-4B-Physics_Math

SGLang

How to use xx18/Composition-RL-4B-Physics_Math 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 "xx18/Composition-RL-4B-Physics_Math" \
    --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": "xx18/Composition-RL-4B-Physics_Math",
		"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 "xx18/Composition-RL-4B-Physics_Math" \
        --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": "xx18/Composition-RL-4B-Physics_Math",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use xx18/Composition-RL-4B-Physics_Math with Docker Model Runner:
```
docker model run hf.co/xx18/Composition-RL-4B-Physics_Math
```

Composition-RL-8B

This repository contains the Composition-RL-8B model, developed as part of the research presented in the paper Composition-RL: Compose Your Verifiable Prompts for Reinforcement Learning of Large Language Models.

Model Description

Composition-RL is a data-efficient Reinforcement Learning with Verifiable Rewards (RLVR) approach designed to improve the reasoning capabilities of Large Language Models. It addresses the issue of "too-easy" prompts (pass-rate = 1) by automatically composing multiple verifiable problems into a single, harder verifiable prompt. This ensures the model continues to receive informative training signals throughout the RL process.

Initial Model: Qwen3-8b-Base
Training Dataset: MATH-Composition-199K
Task: Mathematical Reasoning
Paper: arXiv:2602.12036
Code: GitHub - Composition-RL

Performance

As detailed in the paper, Composition-RL-8B consistently improves reasoning capability over RL trained on original, non-compositional datasets across various benchmarks.

Citation

If you find this work helpful, please consider citing:

@article{xu2026composition-rl,
  title={Composition-RL: Compose Your Verifiable Prompts for Reinforcement Learning of Large Language Models},
  author={Xu, Xin and Bai, Clive and Yang, Kai and Chen, Tianhao and Chen, Yangkun and Liu, Weijie and Chen, Hao and Wang, Yang and Yang, Saiyong and Yang, Can},
  journal={arXiv preprint arXiv:2602.12036},
  year={2026}
}

Downloads last month: 22

Collection including xx18/Composition-RL-4B-Physics_Math

Composition-RL

Collection

Datasets and trained checkpoints of Composition-RL: https://github.com/XinXU-USTC/Composition-RL • 13 items • Updated Mar 19 • 1

Paper for xx18/Composition-RL-4B-Physics_Math

Composition-RL: Compose Your Verifiable Prompts for Reinforcement Learning of Large Language Models

Paper • 2602.12036 • Published Feb 12 • 93