Instructions to use xx18/Composition-RL-4B-Depth1_2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use xx18/Composition-RL-4B-Depth1_2 with Transformers:

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

pipe = pipeline("text-generation", model="xx18/Composition-RL-4B-Depth1_2")
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-Depth1_2")
model = AutoModelForCausalLM.from_pretrained("xx18/Composition-RL-4B-Depth1_2")
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 Settings

vLLM

How to use xx18/Composition-RL-4B-Depth1_2 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-Depth1_2"
# 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-Depth1_2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/xx18/Composition-RL-4B-Depth1_2

SGLang

How to use xx18/Composition-RL-4B-Depth1_2 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-Depth1_2" \
    --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-Depth1_2",
		"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-Depth1_2" \
        --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-Depth1_2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use xx18/Composition-RL-4B-Depth1_2 with Docker Model Runner:
```
docker model run hf.co/xx18/Composition-RL-4B-Depth1_2
```

Add model card and metadata

by nielsr HF Staff - opened Apr 23

base: refs/heads/main

←

from: refs/pr/1

Discussion Files changed

+31

-0

nielsr

Apr 23

This PR adds a model card and relevant metadata to the repository. The model card includes:

A description of the Composition-RL approach for Reinforcement Learning with Verifiable Rewards (RLVR).
Links to the research paper and the official GitHub repository.
Model details, noting it was initialized from Qwen3-8B-Base and trained on the MATH-Composition-199K dataset.
Citation information for the paper.

Relevant metadata for pipeline_tag and library_name have also been added to ensure the model is correctly categorized on the Hugging Face Hub.

Add model card and metadatafdfa8e5a

xx18 changed pull request status to merged Apr 27

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment