Fine-tune Llama2 with Autotrain

Let's start by importing our required dependencies:

Copy

Let's go ahead and set things up:

Copy

Connect to our HuggingFace hub account:

Copy

Transform the dataset

The dataset is in the wrong format we will transform it to the proper template for training llama2 on:

Copy

Let's go ahead and push our new dataset to our HuggingFace account to be used for training.

Start by logging in:

Copy

Push the dataset to hub:

Copy

Finally let's kick off training & push to hub:

Copy

That's it. Check out the Google Colab Notebook by AI BrainBox