The evaluate subcommand can be used to evaluate a trained model against a dataset and report any metrics calculated by the model.

$ python -m evaluate --help
usage: run [command] evaluate [-h] --archive_file ARCHIVE_FILE
                            --evaluation_data_file EVALUATION_DATA_FILE
                            [--cuda_device CUDA_DEVICE]

Evaluate the specified model + dataset

optional arguments:
-h, --help            show this help message and exit
--archive-file ARCHIVE_FILE
                        path to an archived trained model
--evaluation-data-file EVALUATION_DATA_FILE
                        path to the file containing the evaluation data
--cuda-device CUDA_DEVICE
                        id of GPU to use (if any)