Remove unused test dl

TrupsT · Jan 18, 2023 · d80ef1d · d80ef1d
1 parent 2e9f503
commit d80ef1d
Show file tree

Hide file tree

Showing 2 changed files with 17 additions and 34 deletions.
diff --git a/vall_e/data.py b/vall_e/data.py
@@ -244,20 +244,14 @@ def _load_train_val_paths():
     return train_paths, val_paths
 
 
-def _load_test_paths():
-    test_paths = []
-    for data_dir in cfg.test_data_dirs:
-        test_paths.extend(data_dir.rglob("*.phn.txt"))
-    test_paths = sorted(test_paths)
-    return test_paths
-
-
 @cfg.diskcache()
 def create_datasets():
     train_paths, val_paths = _load_train_val_paths()
-    test_paths = _load_test_paths()
 
-    train_dataset = VALLEDatset(train_paths, training=True)
+    train_dataset = VALLEDatset(
+        train_paths,
+        training=True,
+    )
 
     val_dataset = VALLEDatset(
         val_paths,
@@ -269,41 +263,32 @@ def create_datasets():
     val_dataset.interleaved_reorder_(_get_spkr_name)
     val_dataset.head_(cfg.max_num_val)
 
-    test_dataset = VALLEDatset(
-        test_paths,
-        train_dataset.phone_symmap,
-        train_dataset.spkr_symmap,
-        extra_paths_by_spkr_name=train_dataset.paths_by_spkr_name,
-    )
-
-    return train_dataset, val_dataset, test_dataset
+    return train_dataset, val_dataset
 
 
 def create_train_val_dataloader():
-    train_dataset, val_dataset, test_dataset = create_datasets()
+    train_dataset, val_dataset = create_datasets()
 
     train_dl = _create_dl(train_dataset, training=True)
     val_dl = _create_dl(val_dataset, training=False)
-    test_dl = _create_dl(test_dataset, training=False)
 
     _logger.info(str(train_dataset.phone_symmap))
     _logger.info(str(train_dataset.spkr_symmap))
 
     _logger.info(f"#samples (train): {len(train_dataset)}.")
     _logger.info(f"#samples (val): {len(val_dataset)}.")
-    _logger.info(f"#samples (test): {len(test_dataset)}.")
 
-    train_for_val_dataset = copy.deepcopy(train_dataset)
-    train_for_val_dataset.interleaved_reorder_(_get_spkr_name)
-    train_for_val_dataset.head_(cfg.max_num_val)
-    train_for_val_dataset.training_(False)
-    train_for_val_dl = _create_dl(train_for_val_dataset, training=False)
-    assert isinstance(train_for_val_dl.dataset, VALLEDatset)
+    subtrain_dataset = copy.deepcopy(train_dataset)
+    subtrain_dataset.interleaved_reorder_(_get_spkr_name)
+    subtrain_dataset.head_(cfg.max_num_val)
+    subtrain_dataset.training_(False)
+    subtrain_dl = _create_dl(subtrain_dataset, training=False)
+    assert isinstance(subtrain_dl.dataset, VALLEDatset)
 
-    return train_dl, train_for_val_dl, val_dl, test_dl
+    return train_dl, subtrain_dl, val_dl
 
 
 if __name__ == "__main__":
-    train_dl, train_for_val_dl, val_dl, test_dl = create_train_val_dataloader()
+    train_dl, subtrain_dl, val_dl = create_train_val_dataloader()
     sample = train_dl.dataset[0]
     print(sample)
diff --git a/vall_e/train.py b/vall_e/train.py
@@ -30,7 +30,7 @@ def load_engines():
 def main():
     setup_logging(cfg.log_dir)
 
-    train_dl, train_for_val_dl, val_dl, test_dl = create_train_val_dataloader()
+    train_dl, subtrain_dl, val_dl = create_train_val_dataloader()
 
     def train_feeder(engines, batch, name):
         model = engines["model"]
@@ -68,8 +68,7 @@ def run_eval(engines, name, dl):
         log_dir = cfg.log_dir / str(engines.global_step) / name
         stats = defaultdict(list)
         for batch in tqdm(dl):
-            batch: dict
-            batch = to_device(batch, cfg.device)
+            batch: dict = to_device(batch, cfg.device)
 
             if cfg.model.startswith("ar"):
                 resp_list = model(
@@ -114,9 +113,8 @@ def run_eval(engines, name, dl):
         _logger.info(f"{json.dumps(stats)}.")
 
     def eval_fn(engines):
-        run_eval(engines, "train_for_val", train_for_val_dl)
+        run_eval(engines, "subtrain", subtrain_dl)
         run_eval(engines, "val", val_dl)
-        run_eval(engines, "test", test_dl)
 
     trainer.train(
         engines_loader=load_engines,