TiDE-Applications

A unofficial pytorch implementation of "Long-term Forecasting with TiDE: Time-series Dense Encoder" and its sample code of applications

Usage

1. Config model

   edit the config.yaml in config folder

2. Train

   python3 train.py --name name --dataset traffic 

   In order to use a different dataset you need to customize data_utils.py

TODO

Warning: This repository is not finished now, only the network code is done. Sample code of own dataset is not finished now.

• Network code
• Utils code
  • basic utils
  • scheduler
  • data_utils
  • time series dataset
• train and eval
  • logger and summary
• sample code of own dataset