Skip to content

clinicadl train from_json - Launch a new network training based on a JSON file

This command allows to train a network defined in a JSON file. This allows to reproduce MAPS generated by clinicadl train tasks or clinicadl random-search by providing their maps.json file.

This is particularly useful for outputs of random search, as the random architecture produced might be not straightforward to implement.

JSON file generation

JSON files that may be used by clinicadl train from_json can be generated manually by filling with arguments and options of clinicadl train. The undefined values of optional parameters will be automatically replaced by their default value. You can also reuse (and modify) the maps.json files automatically generated by other clincadl train tasks!

Prerequisites

Please check which preprocessing needs to be performed in the JSON file used as input. If it has not been performed, execute the preprocessing pipeline as well as clinicadl preprocessing extract-tensor to obtain the tensor versions of the images.

Running the task

This task can be run with the following command line: 

clinicadl train from_json [OPTIONS] CONFIG_JSON OUTPUT_MAPS_DIRECTORY
where:

  • CONFIG_JSON (Path) is the JSON file used to build the model.
  • OUTPUT_MAPS_DIRECTORY (Path) is the MAPS of the new model that will be trained.

Outputs

The outputs are formatted according to the MAPS.