11:07:22 - INFO - eir.train_utils.utils - Global random seed set to 0.
usage: eirpredict [-h] [--global_configs [GLOBAL_CONFIGS ...]]
                  [--input_configs [INPUT_CONFIGS ...]]
                  [--fusion_configs [FUSION_CONFIGS ...]]
                  [--output_configs [OUTPUT_CONFIGS ...]]
                  --model_path MODEL_PATH [--evaluate]
                  --output_folder OUTPUT_FOLDER
                  [--attribution_background_source {train,predict}]

options:
  -h, --help            show this help message and exit
  --global_configs [GLOBAL_CONFIGS ...]
                        Global .yaml configurations for the experiment.
  --input_configs [INPUT_CONFIGS ...]
                        Input feature extraction .yaml configurations. Each
                        configuration represents one input.
  --fusion_configs [FUSION_CONFIGS ...]
                        Fusion .yaml configurations.
  --output_configs [OUTPUT_CONFIGS ...]
                        Output .yaml configurations.
  --model_path MODEL_PATH
                        Path to model to use for predictions.
  --evaluate
  --output_folder OUTPUT_FOLDER
                        Where to save prediction results.
  --attribution_background_source {train,predict}
                        For attribution analysis, whether to load backgrounds
                        from the data used for training or to use the current
                        data passed to the predict module.
