Instructions to use liumy2010/Llama-3.2-3B-math-RFT with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use liumy2010/Llama-3.2-3B-math-RFT with Transformers:

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

pipe = pipeline("text-generation", model="liumy2010/Llama-3.2-3B-math-RFT")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("liumy2010/Llama-3.2-3B-math-RFT")
model = AutoModelForCausalLM.from_pretrained("liumy2010/Llama-3.2-3B-math-RFT")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use liumy2010/Llama-3.2-3B-math-RFT with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "liumy2010/Llama-3.2-3B-math-RFT"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "liumy2010/Llama-3.2-3B-math-RFT",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/liumy2010/Llama-3.2-3B-math-RFT

SGLang

How to use liumy2010/Llama-3.2-3B-math-RFT 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 "liumy2010/Llama-3.2-3B-math-RFT" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "liumy2010/Llama-3.2-3B-math-RFT",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "liumy2010/Llama-3.2-3B-math-RFT" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "liumy2010/Llama-3.2-3B-math-RFT",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use liumy2010/Llama-3.2-3B-math-RFT with Docker Model Runner:
```
docker model run hf.co/liumy2010/Llama-3.2-3B-math-RFT
```

UFT

This repository contains the model presented in UFT: Unifying Supervised and Reinforcement Fine-Tuning.

Code: https://github.com/liumy2010/UFT

## References

* [UFT: Unifying Supervised and Reinforcement Fine-Tuning](https://arxiv.org/abs/2505.16984)

Downloads last month: -

Safetensors

Model size

4B params

Tensor type

F32

Model tree for liumy2010/Llama-3.2-3B-math-RFT

Base model

meta-llama/Llama-3.2-3B

Finetuned

(447)

this model

Collection including liumy2010/Llama-3.2-3B-math-RFT

UFT

Collection

UFT: Unifying Supervised and Reinforcement Fine-Tuning • 80 items • Updated May 28, 2025 • 1

Paper for liumy2010/Llama-3.2-3B-math-RFT

UFT: Unifying Supervised and Reinforcement Fine-Tuning

Paper • 2505.16984 • Published May 22, 2025 • 3