Merge pull request #3186 from PaddlePaddle/vits_pr

[TTS]update lr schedulers from per iter to per epoch for VITS

examples/csmsc/vits/conf/default.yaml

@@ -179,7 +179,7 @@ generator_first: False # whether to start updating generator first
 #                OTHER TRAINING SETTING                  #
 ##########################################################
 num_snapshots: 10            # max number of snapshots to keep while training
-train_max_steps: 350000      # Number of training steps. == total_iters / ngpus, total_iters = 1000000
+max_epoch: 1000              # Number of training epochs.
-save_interval_steps: 1000    # Interval steps to save checkpoint.
+save_interval_epochs: 1      # Interval epochs to save checkpoint.
-eval_interval_steps: 250     # Interval steps to evaluate the network.
+eval_interval_epochs: 1      # Interval steps to evaluate the network.
 seed: 777                    # random seed number

paddlespeech/t2s/exps/vits/train.py

@@ -230,17 +230,15 @@ def train_sp(args, config):
         output_dir=output_dir)
 
     trainer = Trainer(
-        updater,
+        updater, stop_trigger=(config.max_epoch, 'epoch'), out=output_dir)
-        stop_trigger=(config.train_max_steps, "iteration"),
-        out=output_dir)
 
     if dist.get_rank() == 0:
         trainer.extend(
-            evaluator, trigger=(config.eval_interval_steps, 'iteration'))
+            evaluator, trigger=(config.eval_interval_epochs, 'epoch'))
         trainer.extend(VisualDL(output_dir), trigger=(1, 'iteration'))
     trainer.extend(
         Snapshot(max_size=config.num_snapshots),
-        trigger=(config.save_interval_steps, 'iteration'))
+        trigger=(config.save_interval_epochs, 'epoch'))
 
     print("Trainer Done!")
     trainer.run()

paddlespeech/t2s/models/vits/vits_updater.py

@@ -166,6 +166,8 @@ class VITSUpdater(StandardUpdater):
                 gen_loss.backward()
 
                 self.optimizer_g.step()
+                # learning rate updates on each epoch.
+                if self.state.iteration % self.updates_per_epoch == 0:
                     self.scheduler_g.step()
 
                 # reset cache
@@ -202,6 +204,8 @@ class VITSUpdater(StandardUpdater):
                 dis_loss.backward()
 
                 self.optimizer_d.step()
+                # learning rate updates on each epoch.
+                if self.state.iteration % self.updates_per_epoch == 0:
                     self.scheduler_d.step()
 
                 # reset cache