This repository contains the Jupyter notebooks for generating a data sample, training a LLM, and evaluating it.
The selected LLM is Vicuna, a Llama 2 Fine-tuned model [https://lmsys.org/blog/2023-03-30-vicuna/].
The LLM has been fine-tuned using the data sample to transform Vicuna into a financial statement analyst.
Model size and version used: Vicuna 7B 1.5 [https://huggingface.co/lmsys/vicuna-7b-v1.5].
FastChat is an open platform for training, serving, and evaluating large language model based chatbots. It can be used with Vicuna LLM [https://github.com/lm-sys/FastChat].