To run any experiment, this source code uses the following convention:
WANDB_MODE={offline/online} HYDRA_FULL_ERROR={0/1} python src/main.py --config-name $CONFIG_NAME
CONFIG_NAME variable corresponds to the name of the .yaml configuration file from the configs directory. To modify parameters from the config, use path.in.config.to.parameter=<new value> after $CONFIG_NAME following Hydra documentation.
Data from the following links should be download and placed in the corresponding directory. This ensures integrity with original implementation.
data/datasets/brainmri:
data/datasets/rsna:
data/datasets/div2k:
data/datasets/celebahq:
data/metadata:
data/weights: