radiantearth/cyclone-model-torchgeo repository overview

⁠cyclone-model-torchgeo

⁠Usage

Use this image to run the Tropical Cyclone Wind Estimation model⁠ to predict the average wind speed of a tropical cyclone based on a JPEG image of the storm. This model was built with using the Torchgeo⁠ package and trained on the Tropical Cyclone Wind Estimation Competition dataset⁠ from Radiant MLHub⁠.

To run the model on new images:

1. Put JPEG images in a host directory and mount that directory to /var/data/input in the container
2. Create a directory to hold outputs and mount that directory to /var/data/output in the container
3. Use docker run, mounting the input data directory to /var/data/input and the output directory to /var/data/output:

   $ export INPUT_VOLUME=/some/input/directory
   $ export OUTPUT_VOLUME=/some/output/directory
   $ docker run -it --rm  -v $INPUT_VOLUME:/var/data/input -v $OUTPUT_VOLUME:/var/data/output cyclone-model-torchgeo:latest
   

   Copy

See the Tropical Cyclone Wind Estimation Competition dataset⁠ for details on the required format and size of the input images.

⁠Citation

Caleb Robinson. (2021). Tropical Cyclone Wind Estimation model (2.0). Zenodo. https://doi.org/10.5281/zenodo.5773331⁠.