/pile_data_analysis

Analyzing failure and success cases for GPT3 and correlating that to the training data.

Primary LanguagePython

pile_data_analysis

Analyzing failure and success cases for GPT3 and correlating that to the training data.

How to Run

This code was tested with python==3.8 so we recommend creating your virtual environment with that python version. Then install the requirements with pip install -r requirements.txt.

Run the code from the main directory with the following command python -m src.main. This code uses the main.yaml file defined in the conf directory. To specify different config options follow the hydra documentation.