cloudmask-fit

Cloudmask training, prediction and evaluation of the results.

Refer to wiki for more detailed description of the workflow, model architectures and results.

Setup

1. Create a conda environment.

    conda env create -f environment.yml
   

2. Copy config/config_example.json and adapt it to your needs.

Usage

Model fitting can be performed as follows:

conda activate cm_fit
python3 cm-fit.py -c config/your_config.json

Once the model fitting has concluded, the model weights in the output directory can be used for prediction. This can be performed as follows:

conda activate cm_fit
python3 cm-fit.py -c config/your_config.json -p input/for_prediction/T35VLF_20200529T094041_tile_1024_4864.nc -w output/a1_029-0.96.hdf5