/LLM-Test

Primary LanguageJupyter Notebook

README

About us

Hikmat Baniya: Hey bro put your intro here....

Sulav Adhikari: Python Developer interested in Security, AI/ML/DL, Web Dev and Automation in general. BackEnd Developer by expertise (Django). Two time winner for Dropout of the year. Dumber than others but better than you.

Overview

This repository contains an LLM (Large Language Model) for a simple chatbot for football conversations and fetching data from the Wikipedia API for Premier League and La Liga clubs. Initially, we attempted to utilize a bigram model for processing the data. However, due to memory constraints and local storage limitations, this approach was unsuccessful. As a result, we have implemented a transformer model. Please note that while the algorithm of this transformer model is not yet fully tested, we are committed to refining it in future iterations.

Requirements

  • Python 3.x
  • Internet connection for fetching data from the Wikipedia API

Steps

  • Data collection: completed with Wikipedia API
  • Tokenizer setup: completed with GPT3 tokenizer
  • Model setup: Failed trainig Bigram Model. Completed setup of Transformer Model.
  • Training the model: TBD
  • Test Evaluation: TBD
  • Fine Tuning Data collection: TBD
  • Fine Tuning: TBD
  • API implementation: TBD

Contribution

Just send a PR man. We don't know what we're doing here. We're not that good at math. Help will be appretiated.