Fixup `utils` for the trainer

TTS/utils/generic_utils.py

@@ -16,6 +16,7 @@ import torch
 def to_cuda(x: torch.Tensor) -> torch.Tensor:
     if x is None:
         return None
+    if torch.is_tensor(x):
         x = x.contiguous()
         if torch.cuda.is_available():
             x = x.cuda(non_blocking=True)
@@ -57,12 +58,9 @@ def get_commit_hash():
     return commit
 
 
-def create_experiment_folder(root_path, model_name, debug):
+def create_experiment_folder(root_path, model_name):
     """Create a folder with the current date and time"""
     date_str = datetime.datetime.now().strftime("%B-%d-%Y_%I+%M%p")
-    if debug:
-        commit_hash = "debug"
-    else:
     commit_hash = get_commit_hash()
     output_folder = os.path.join(root_path, model_name + "-" + date_str + "-" + commit_hash)
     os.makedirs(output_folder, exist_ok=True)

TTS/utils/logging/tensorboard_logger.py

@@ -34,6 +34,8 @@ class TensorboardLogger(object):
 
     def dict_to_tb_audios(self, scope_name, audios, step, sample_rate):
         for key, value in audios.items():
+            if value.dtype == "float16":
+                value = value.astype("float32")
             try:
                 self.writer.add_audio("{}/{}".format(scope_name, key), value, step, sample_rate=sample_rate)
             except RuntimeError:

TTS/utils/manage.py

@@ -137,7 +137,7 @@ class ModelManager(object):
                 # set scale stats path in config.json
                 config_path = output_config_path
                 config = load_config(config_path)
-                config.external_speaker_embedding_file = output_speakers_path
+                config.d_vector_file = output_speakers_path
                 config.save_json(config_path)
         return output_model_path, output_config_path, model_item
 

TTS/utils/radam.py

@@ -1,4 +1,4 @@
-# from https://github.com/LiyuanLucasLiu/RAdam
+# modified from https://github.com/LiyuanLucasLiu/RAdam
 
 import math