This project aims to train OpenAI's ChatGPT using local documents and develop prompts to generate desired results. It provides a flexible and collaborative environment for training and experimenting with ChatGPT.
- Train ChatGPT with local documents, including text files, PDFs, CSVs, and more.
- Develop prompts to generate specific content.
- Save results in a format that can be easily accessed and shared.
- Sync the project to Git for collaborative development.
- Clone the repository:
git clone [repository-url] - Install the required dependencies:
pip install -r requirements.txt - Prepare your training data by placing the documents in the designated directory.
- Run the training script:
python train.py
- Ask general questions to create effective prompts.
- Experiment with different training strategies and document sources.
- Adjust hyperparameters and model settings to achieve desired results.
- Save the generated content locally for further analysis and sharing.
This project's working branch is main, and it has the following directory structure:
[REPOSITORY-Name]
├── LICENSE.md
├── Pipfile
├── Pipfile.lock
├── README.md
├── data
│ ├── raw
│ └── processed
├── docs
│ └── support documentation and project descriptions
└── src
└── all executbale script files
-
Clone the project
git clone [https://github.com/YOUR-USERNAME/YOUR-REPOSITORY]
-
Install dependencies from pipfile. More pipenv install info here
# install pipenv if you don't have it pip install --user pipenv# now we can install required dependencies pipenv install -
run
main.pypython main.py
This will use the data located in data/raw and run through the full data pipeline.
The goal of this project is to [short goal description].
The work is completed in stages:
- [stage 1]
- [stage 2]
The scope is to...
Here we will....
This project is open source software licensed as MIT.