Welcome to miniChatGPT!

A light-weight example of chatGPT using neural nets and custom trained model

miniChatGPT.mov

Tech Stack

• TensorFlow is an end-to-end open-source platform for machine learning
• Flask is a micro web framework written in Python
• NLTK is a leading platform for building Python programs to work with human language data
• Python, HTML/CSS

How to set up and run miniChatGPT

1. Clone Project

git clone [https://github.com/eruigu/miniChatGPT.git]

2. Setup Environment

install python3: check the version "python3 --version"

MacOS: pip3 install virtualenv

Linux (Debian): sudo apt install pythonx.x-venv (replace x with your version of python)

3. Create virtual environment

python3 -m venv venv

4. Activate virtual environment and update pip

source venv/bin/activate && pip install -U pip setuptools

5. Install wheel library

MacOS: pip3 install wheel

Linux (Debian): pip install wheel

6. Install dependencies

MacOS: pip3 install -r requirements.txt

Linux (Debian): pip install -r requirements.txt

7. Train chatbot model

python3 -m minichatgpt.model_training.training

This will create two .pkl (pickle) files: classes and words & miniGPT_model.h5 in src/backend/

8. Run the chatbot

python3 -m minichatgpt.backend.server

this will create a .h5 model file in the current directory

9. The server is running on 127.0.0.1:8001

You can change the server config in server.py

10. Docker

docker docker build -t minichatgpt:1.0.0 .

this will create an image for the project

docker run --name minichatgpt -d -p 8001:8001 minichatgpt:1.0.0

this creates and runs a container for the image in the background.

access at localhost:8001