Fine-Tuning Llama-3.1-8B-Instruct

This repository contains code for fine-tuning meta-llama/Meta-Llama-3.1-8B-Instruct on the Bitext - Customer Service Tagged Training Dataset.

For a complete walkthrough and technical details, check out the blog post.

Quick Start

The Makefile contains the main commands to run the fine-tuning process :

1. Create a virtual environment:

make setup

2. Install dependencies:

make install

3. Authenticate with Hugging Face You need to authenticate your environment with Hugging Face to access the model. You can do this by logging in using the Hugging Face CLI:

huggingface-cli login

This command will prompt you to enter your Hugging Face credentials. Create a new READ token at https://huggingface.co/settings/tokens and copy it.

4. Launch the fine-tuning process:

make train