Move fairseq part of D10478427 directly into pytorch-translate (#337)

Summary:
Pull Request resolved: https://github.com/pytorch/fairseq/pull/337

Pull Request resolved: https://github.com/pytorch/translate/pull/250

Reviewed By: akinh

Differential Revision: D12880352

fbshipit-source-id: 61e9888a9cc3df07e805820b74a5fcf359dfe0ea

fairseq/trainer.py

@@ -45,7 +45,6 @@ class Trainer(object):
             self._model = model.cuda()
 
         self._dummy_batch = dummy_batch
-        self._num_iterations = 0
         self._num_updates = 0
         self._optim_history = None
         self._optimizer = None
@@ -224,7 +223,6 @@ class Trainer(object):
             ).format(self.task.__class__.__name__))
 
         try:
-            self._num_iterations += 1
             # normalize grads by sample size
             self.optimizer.multiply_grads(self.args.distributed_world_size / float(sample_size))
 
@@ -357,10 +355,6 @@ class Trainer(object):
         """Get the number of parameters updates."""
         return self._num_updates
 
-    def get_num_iterations(self):
-        """Get the number of iterations."""
-        return self._num_iterations
-
     def _prepare_sample(self, sample):
         if sample is None or len(sample) == 0:
             return None