AMS: Ineligible Area Detection
cm_fit
conda environment is used.
Also, segemntation_models library (https://segmentation-models.readthedocs.io/en/latest/install.html) is needed:
pip install segmentation-models
You can download data here: https://drive.google.com/drive/folders/11KUJ0K6Pts02HGe2bqAjBhjFp8JuHVLQ?usp=sharing
You can create config file in config
folder. To train, use the following command:
python pipeline.py --cfg config/train_0703.json --train
To predict, use the following command. It will create some eaxmple outputs in the predicted_examples
directory.
python pipeline.py --cfg config/train_0703.json --predict
Basically, using config train_0703.json
will give you the final results.