Merge pull request #3423 from idiap/fix-aux-tests

Fix CI (save best model after 0 steps in tests)

TTS/bin/train_encoder.py

@@ -125,7 +125,7 @@ def evaluation(model, criterion, data_loader, global_step):
 
 def train(model, optimizer, scheduler, criterion, data_loader, eval_data_loader, global_step):
     model.train()
-    best_loss = float("inf")
+    best_loss = {"train_loss": None, "eval_loss": float("inf")}
     avg_loader_time = 0
     end_time = time.time()
     for epoch in range(c.epochs):
@@ -248,7 +248,7 @@ def train(model, optimizer, scheduler, criterion, data_loader, eval_data_loader,
             )
             # save the best checkpoint
             best_loss = save_best_model(
-                eval_loss,
+                {"train_loss": None, "eval_loss": eval_loss},
                 best_loss,
                 c,
                 model,

requirements.txt

@@ -27,7 +27,7 @@ pandas>=1.4,<2.0
 # deps for training
 matplotlib>=3.7.0
 # coqui stack
-trainer>=0.0.32
+trainer>=0.0.36
 # config management
 coqpit>=0.0.16
 # chinese g2p deps