DeepGraphPose can be run in training or prediction mode. For training, we assume that frames are labeled elsewhere by the user. For prediction, we assume input as videos. IMPORTANT: we additionally assume that the user will only be analyzing videos with one shuffle and one training fraction at a time. Different parameters are required for the training and testing routines, as described below.

TRAINING:
Args:
  -Input: (zip file) A zipped folder corresponding to a DLC model folder (with format {taskname}_{scorername}_{date}), containing the training frames to be used for training. This zipped file should include the folder itself. It should not contain any trained models.
  -Config: (yaml) A YAML file indicating the mode (train or predict), type of videos to analyze, and other neurocaas specific parameters. See template config for details.

PREDICTION:
Args:
  -Input: (zip file) A zipped folder corresponding to a DLC model folder (with format {taskname}_{scorername}_{date}), containing the training frames to be used for training. This zipped file should include the folder itself. This model folder should include a trained model already.
  -Config: (yaml) A YAML file indicating the mode (train or test), type of videos to analyze, and other neurocaas specific parameters. See template config for details.

Outputs:
  -(zipfile) If training, full models fit from the data, as zip files results{modelname}.zip. If prediction only, returns only a zipfile of the videos and corresponding traces.