mt_validation_args.yml

validator.class: SeqGenerationValidator
validator.params:
  eval_dataset: ParallelTextDataset
  eval_dataset.params:
    src_file: DATA_PATH/transcripts/TRG_LANG/dev.en.txt
    trg_file: DATA_PATH/transcripts/TRG_LANG/dev.TRG_LANG.txt
  eval_batch_size: 64
  eval_start_at: 5000
  eval_steps: 1000
  eval_criterion: label_smoothed_cross_entropy
  eval_search_method: beam_search
  eval_search_method.params:
    beam_size: 4
    length_penalty: -1
    maximum_decode_length: 180
    extra_decode_length: 50
  eval_metric: tok_bleu
  eval_top_checkpoints_to_keep: 10
  eval_auto_average_checkpoints: True