Running the codes.

1. First of all ensure the data is downloaded by running the bash script get_data.sh in 'scripts/`.
2. Move the data to the main repo as mv data_local ../
3. Install all dependencies as pip install -r requirements.txt.
4. Then, for running the parameter sweep, do: wandb sweep sweep.yaml, which will give you a sweep_id (actually a command as wandb agent ...)
5. Run the command wandb agent <sweep_id>

Reference for the sweeps: https://github.com/borisdayma/lightning-kitti.

The data and model checkpoints are in "geo2020_supermag/Uniformdata_multiyear/" in the bucket.