Large Language Models Supervised Fine-tuning

This repository contains the code for our Deep Learning Course Work. In this work we choose the subset of ultralchat dataset as our training data and fine-tuned Llama-7B and Baichuan2-7B models on this dataset.

You can see the detail in our report