[s2s] command line args for faster val steps (#6833)

Author: sshleifer

Diff for: examples/seq2seq/distillation.py

@@ -262,7 +262,7 @@ class BartTranslationDistiller(BartSummarizationDistiller):
 
     mode = "translation"
     metric_names = ["bleu"]
-    val_metric = "bleu"
+    default_val_metric = "bleu"
 
     def __init__(self, hparams, **kwargs):
         super().__init__(hparams, **kwargs)

Diff for: examples/seq2seq/finetune.py

@@ -63,7 +63,7 @@ class SummarizationModule(BaseTransformer):
     mode = "summarization"
     loss_names = ["loss"]
     metric_names = ROUGE_KEYS
-    val_metric = "rouge2"
+    default_val_metric = "rouge2"
 
     def __init__(self, hparams, **kwargs):
         super().__init__(hparams, num_labels=None, mode=self.mode, **kwargs)
@@ -110,6 +110,9 @@ def __init__(self, hparams, **kwargs):
         self.dataset_class = (
             Seq2SeqDataset if hasattr(self.tokenizer, "prepare_seq2seq_batch") else LegacySeq2SeqDataset
         )
+        self.eval_beams = self.model.config.num_beams if self.hparams.eval_beams is None else self.hparams.eval_beams
+        assert self.eval_beams >= 1, f"got self.eval_beams={self.eval_beams}. Need an integer > 1"
+        self.val_metric = self.default_val_metric if self.hparams.val_metric is None else self.hparams.val_metric
 
     def freeze_embeds(self):
         """Freeze token embeddings and positional embeddings for bart, just token embeddings for t5."""
@@ -301,6 +304,8 @@ def add_model_specific_args(parser, root_dir):
         parser.add_argument("--label_smoothing", type=float, default=0.0, required=False)
         parser.add_argument("--src_lang", type=str, default="", required=False)
         parser.add_argument("--tgt_lang", type=str, default="", required=False)
+        parser.add_argument("--eval_beams", type=int, default=None, required=False)
+        parser.add_argument("--val_metric", type=str, default=None, required=False)
         parser.add_argument(
             "--early_stopping_patience",
             type=int,
@@ -315,7 +320,7 @@ class TranslationModule(SummarizationModule):
     mode = "translation"
     loss_names = ["loss"]
     metric_names = ["bleu"]
-    val_metric = "bleu"
+    default_val_metric = "bleu"
 
     def __init__(self, hparams, **kwargs):
         super().__init__(hparams, **kwargs)

Diff for: examples/seq2seq/test_seq2seq_examples.py

@@ -31,6 +31,8 @@
 CUDA_AVAILABLE = torch.cuda.is_available()
 CHEAP_ARGS = {
     "label_smoothing": 0.2,
+    "eval_beams": 1,
+    "val_metric": None,
     "adafactor": True,
     "early_stopping_patience": 2,
     "logger_name": "default",