UTCI

Setup

1. Install the conda environment, via `conda env create --name utci --file env.yml
2. Activate the environemnt, via `conda activate utci'

Prepare data

1. Download the raw data

2. Run python preprocess.py --raw_data data/raw/UHTC_NN --output_dir data

3. Run python preprocess_mean_std.py --output_dir data and copy the printed strings into the dictionary UTCI_STATISTICS in utils.py

Train

To train the UTCI model, run

python train.py exp/01 --data_path data --skip 1 2 3 --amp --clip_grad --without_aveg

Evaluate

To test the previously trained UTCI model, run

python eval.py exp/01/model.pth

If you want to:

• save detailed information (per sample additionally to the aggregated information), add --detailed.
• plot the UTCI predictions, add --plot.
• save the UTCI predictions as numpy files for further processing, add --save_np.

Predict UTCIs for a given time period

To predict UTCIs for a given time period and get the aggregated results, run

python eval_time_period.py exp/01/model.pth --amp --verbose --temporal_data $temporal_data

where $temporal_data is a placeholder for the path to the csv file with the meteorological data of the time period.