Fixups

TTS/tts/configs/fast_pitch_config.py

@@ -113,7 +113,7 @@ class FastPitchConfig(BaseTTSConfig):
     base_model: str = "forward_tts"
 
     # model specific params
-    model_args: ForwardTTSArgs = ForwardTTSArgs()
+    model_args: ForwardTTSArgs = field(default_factory=ForwardTTSArgs)
 
     # multi-speaker settings
     num_speakers: int = 0

TTS/utils/audio/processor.py

@@ -540,7 +540,10 @@ class AudioProcessor(object):
 
     def _griffin_lim(self, S):
         angles = np.exp(2j * np.pi * np.random.rand(*S.shape))
-        S_complex = np.abs(S).astype(np.complex)
+        try:
+            S_complex = np.abs(S).astype(np.complex)
+        except AttributeError:  # np.complex is deprecated since numpy 1.20.0
+            S_complex = np.abs(S).astype(complex)
         y = self._istft(S_complex * angles)
         if not np.isfinite(y).all():
             print(" [!] Waveform is not finite everywhere. Skipping the GL.")

TTS/utils/manage.py

@@ -287,6 +287,7 @@ class ModelManager(object):
                 "author": "fairseq",
                 "description": "this model is released by Meta under Fairseq repo. Visit https://github.com/facebookresearch/fairseq/tree/main/examples/mms for more info.",
             }
+            model_item["model_name"] = model_name
         else:
             # get model from models.json
             model_item = self.models_dict[model_type][lang][dataset][model]

TTS/vc/models/freevc.py

@@ -794,8 +794,8 @@ class FreeVCConfig(BaseVCConfig):
 
     model: str = "freevc"
     # model specific params
-    model_args: FreeVCArgs = FreeVCArgs()
-    audio: FreeVCAudioConfig = FreeVCAudioConfig()
+    model_args: FreeVCArgs = field(default_factory=FreeVCArgs)
+    audio: FreeVCAudioConfig = field(default_factory=FreeVCAudioConfig)
 
     # optimizer
     # TODO with training support

requirements.txt

@@ -1,6 +1,6 @@
 # core deps
-numpy==1.22.0
-numpy==1.22.0
+numpy==1.22.0;python_version<="3.10"
+numpy==1.24.3;python_version>"3.10"
 cython==0.29.30
 scipy>=1.4.0
 torch>=1.7
@@ -8,7 +8,7 @@ torchaudio
 soundfile
 librosa==0.10.0.*
 numba==0.55.1;python_version<"3.9"
-numba==0.56.4;python_version>="3.9"
+numba==0.57.0;python_version>="3.9"
 inflect==5.6.0
 tqdm
 anyascii