Fine Tune an LLM for Natural Language Apps

Fine Tune an LLM for Natural Language Apps

Fine-tune the Falcon-7B and Falcon-7B-Instruct models on a conversational dataset using PyTorch Lightning. Code in this file downloads the Falcon-7B model, converts it to a format that PyTorch Lightning can use, prepares the dataset for the fine-tuning and starts the fine-tuning. Once the fine-tuning is done, it also generates outputs from user-supplied prompts.