This repository provides tools to generate reasoning datasets, fine-tune language models (e.g., Phi-3), and perform inference. Designed for researchers and developers exploring LLM reasoning capabilities.
Before you begin:
- Python 3.8+
- PyTorch 2.0+
- CUDA-capable GPU (recommended)
- Hugging Face libraries (
transformers,datasets)
- Clone Repository
git clone https://github.com/vuquangminh303/fine-tune-for-reasoning.git cd fine-tune-for-reasoning