Figure 1. Overall ETSformer Architecture.
Official PyTorch code repository for the ETSformer paper. Check out our blog post!
- ETSformer is a novel time-series Transformer architecture which exploits the principle of exponential smoothing in improving Transformers for timeseries forecasting.
- ETSformer is inspired by the classical exponential smoothing methods in time-series forecasting, leveraging the novel exponential smoothing attention (ESA) and frequency attention (FA) to replace the self-attention mechanism in vanilla Transformers, thus improving both accuracy and efficiency.
- Install Python 3.8, and the required dependencies.
- Required dependencies can be installed by:
pip install -r requirements.txt
- Pre-processed datasets can be downloaded from the following links, Tsinghua Cloud or Google Drive, as obtained from Autoformer's GitHub repository.
- Place the downloaded datasets into the
dataset/folder, e.g.dataset/ETT-small/ETTm2.csv.
- Install the required dependencies.
- Download data as above, and place them in the folder,
dataset/. - Train the model. We provide the experiment scripts of all benchmarks under the folder
./scripts, e.g../scripts/ETTm2.sh. You might have to change permissions on the script files by runningchmod u+x scripts/*. - The script for grid search is also provided, and can be run by
./grid_search.sh.
The implementation of ETSformer relies on resources from the following codebases and repositories, we thank the original authors for open-sourcing their work.
- https://github.com/thuml/Autoformer
- https://github.com/zhouhaoyi/Informer2020
Please consider citing if you find this code useful to your research.
@article{woo2022etsformer,
title={ETSformer: Exponential Smoothing Transformers for Time-series Forecasting},
author={Gerald Woo and Chenghao Liu and Doyen Sahoo and Akshat Kumar and Steven C. H. Hoi},
year={2022},
url={https://arxiv.org/abs/2202.01381},
}