timesfm

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

Usage

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

$ python3 timesfm.py

You can specify dataset which had been used to train model by using --input option. You can use --savepath option to change the name of the output file to save.

$ python3 timesfm.py --input CSV_FILE --savepath SAVE_IMAGE_PATH

Use the --target option to specify the columns to forecast from the data file. In addition to column names, you can also specify column indexes.

$ python3 timesfm.py --target OT

Specify the length given to the model with the --context_len option, and the predicted length with the --horizon_len option.

$ python3 timesfm.py --context_len 512 --horizon_len 128

The --forecast_len option specifies the length of data to use for validation from the end of the data.

$ python3 timesfm.py --forecast_horizon 128

Reference

• TimesFM
• Hugging Face - google/timesfm-1.0-200m

Framework

Pytorch

Model Format

ONNX opset=17

Netron

timesfm-1.0-200m.onnx.prototxt