informer2020

Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting

input

Time-series data (csv file of Electricity Transformer Dataset)

Dataset can be available from Electricity Transformer Dataset (ETDataset)

Output

Predicted time-series data and visualized images.

Visalized image (HUFL feature)

Visalized image (OT feature)

Usage

Automatically downloads the onnx and prototxt files on the first run. It is necessary to be connected to the Internet while downloading.

$ python3 informer2020.py

If you want to run with onnxruntime instead of ailia, you can use --onnx option as below.

$ python3 informer2020.py --onnx

You can specify model architecture by using --model option. Available models are informer and informerstack. (default is informer.)

$ python3 informer2020.py --model [MODEL_TYPE]

You can specify dataset which had been used to train model by using --data option. Available datasets are ETTh1 and ETTm1. (default is ETTh1.)

$ python3 informer2020.py --data [DATASET_TYPE]

Reference

• Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting (AAAI'21 Best Paper)
• Electricity Transformer Dataset (ETDataset)

Framework

Pytorch

Model Format

ONNX opset=14

Netron

informer_ETTh1.onnx.prototxt

informer_ETTm1.onnx.prototxt

informerstack_ETTh1.onnx.prototxt

informerstack_ETTm1.onnx.prototxt