Skip to content

josephtwilliams/FineTuneYou

Repository files navigation

FineTuneYou

Overview

FineTuneYou is a project designed to fine-tune language models based on personalised text documents, such as resumes and custom questions. Leveraging the capabilities of GPT-3.5-turbo, this project allows users to generate questions based on their resume, fine-tune a language model, and use this model for retrieval augmented questions answering.

Features

  • PDF to Text: Convert your resume from PDF format to a text file.
  • Question Generation: Generate custom questions based on your resume.
  • Fine-Tuning: Fine-tune a language model using your generated questions and resume.
  • Embedding Generation: Create embeddings for fast and efficient querying.
  • Question Answering: Utilize the fine-tuned model for Q&A tasks based on your resume.

Requirements

  • Python 3.11
  • OpenAI GPT-3.5-turbo API key
  • Various Python packages such as openai, dotenv, etc.

Installation

  1. Clone the Repository

    git clone https://github.com/josephtwilliams/FineTuneYou.git

  2. Install Required Python Packages

    pip install -r requirements.txt

  3. Set Up Your OpenAI API Key Create a .env file in the root directory and add the following line:

    OPENAI_API_KEY=your-api-key

Usage

Step 1: Upload Resume and Fill Out Questions

  • Upload your resume into the root directory, replacing resume_example.pdf.
  • Fill out additional questions in the questions.json file. You are free to add your own questions as well.

Step 2: Run the Notebook

Open and run all the cells in the FineTuneYou.ipynb notebook.

Step 3: Check Results

After successfully running the notebook, you can check the results of the fine-tuning and question-answering tasks.

Resetting the Project

To reset the project and remove all generated files except for essential ones like inappropriate_questions.txt and questions.json, you can use the reset_project() Python function.

pythonCopy code

from reset_script import reset_project reset_project()

License

MIT

About

Combine fIne-tuning and retrieval-augmented generation

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published