Enable Hydra configs in fairseq (#1343) (#1510)

Summary:
Pull Request resolved: #1510

this is the main pr that switches on hydra functionality in fairseq

we migrate "args" object into omegaconf "DictConfig" at all legacy entry points

in addition this migrates various components from secondary registries (like bpe encoders and tokenizers) to make the migration smoother

i am going through code that references migrated fairseq components and changing it to inherit from "Legacy*" components instead. hopefully tests will catch most of this

Pull Request resolved: fairinternal/fairseq-py#1343

Reviewed By: myleott

Differential Revision: D23973928

Pulled By: alexeib

fbshipit-source-id: dd9554981fff51ea75c1ff343874d1d6e61793c9

pytext/metric_reporters/seq2seq_metric_reporter.py

@@ -82,9 +82,11 @@ def calculate_metric(self):
         total_count = len(self.all_targets)
         trg_vocab = self.tensorizers["trg_seq_tokens"].vocab
         bleu_scorer = bleu.Scorer(
-            trg_vocab.get_pad_index(),
-            trg_vocab.get_eos_index(),
-            trg_vocab.get_unk_index(),
+            bleu.BleuConfig(
+                pad=trg_vocab.get_pad_index(),
+                eos=trg_vocab.get_eos_index(),
+                unk=trg_vocab.get_unk_index(),
+            )
         )
         for beam_pred, target in zip(self.all_preds, self.all_targets):
             pred = beam_pred[0]