End users might not fully understand what is required as an entry for the parameter trained_model_path in the metrics_config.yaml file.
# Fill model_params if computing metrics_only on a pre-trained model else leave blank if training model first
# Specify path to Trained Model - Optional if model is pre-trained
trained_model_path:
They might think they need to provide the full path to the psf_model checkpoint files, rather than just the path to the parent directory (i.e. wf-outputs-YEARMONTHDAYHHMMSS for a previously trained model. Either rename the parameter or improve the instructions.
End users might not fully understand what is required as an entry for the parameter
trained_model_pathin themetrics_config.yamlfile.They might think they need to provide the full path to the psf_model checkpoint files, rather than just the path to the parent directory (i.e.
wf-outputs-YEARMONTHDAYHHMMSSfor a previously trained model. Either rename the parameter or improve the instructions.