Postprocessor is function which processes prediction and/or annotation data after model infer and before metric calculation. For correct work postprocessors require specific representation format. (e. g. clip boxes postprocessor expects detection annotation and detection prediction for processing).

In case when you use complicated representation located in representation container, you can add options annotation_source and prediction_source in configuration file, if you want process only specific representations, another way postprocessor will be used for all suitable representations. annotation_source and prediction_source should contain comma separated list of annotation identifiers and output layer names respectively.

Every postprocessor has parameters available for configuration.

Accuracy Checker supports following set of postprocessors: