The repo contains 2 demos

1. demo.py

   • Compares PatchTST with NBEATS and NHITS
   • Compares PatchTST with ARIMA

2. gradio_app/app.py

   • A gradio app to interact with the forecasts
   • Run python gradio_app/app.py to start the app
   • The app can be accessed locally

The data folder has all the data used in the demos for convenience.

Some things to remember:

• Drop a .env file in the gradio_app/ folder with your openai api key. This is required to use gpt4
• The first demo demo.py is best run on a GPU.