This repo use config to parse the training configs, with pytorch-lightning as the backbone. Use segmentation experiment:
python train.py fit --config configs/segmentation/segment_config.yaml
use spectral recovery experiment:
python train.py fit --config configs/spectral_recovery/spectral_config.yaml