abisoyeabidakun / tsmixers-with-pytorch_lightning

Baseline Implementation of TSMixer to understand the concept, and further develop new algorithm architecture for application in other Deep Learning Domain

Adapting TSMixer in PyTorch_Lightning for Multivariate Long-Term Forecasting

Installation

Install the dependencies:

pip install -r requirements.txt

Dataset Preparation

Using pre-processed datasets provided in Autoformer.

mkdir dataset
cd dataset
# Download zip file from [Google Drive](https://drive.google.com/corp/drive/folders/1ZOYpTUa82_jCcxIdTmyr0LXQfvaM9vIy) and put it under dataset/
unzip all_six_datasets.zip
mv all_six_datasets/*/*.csv ./

Training Example

Use run_tuned_hparam.sh to reproduce results of 96 prediction length.

sh run_tuned_hparam.sh ETTm2
sh run_tuned_hparam.sh weather
sh run_tuned_hparam.sh electricity
sh run_tuned_hparam.sh traffic

Acknowledgement

We appreciate the following github repos for their valuable code base or datasets:

1. Are Transformers Effective for Time Series Forecasting?
2. Autoformer
3. [TSMixers] (https://github.com/google-research/google-research/blob/master/tsmixer)