ken11o2/tortoise-tts
1
0
Fork 0
forked from mrq/tortoise-tts
Code Pull Requests Activity

load the model on CPU because torch doesn't like loading models directly to GPU (it just follows the default vocoder loading behavior)

Browse Source
This commit is contained in:
mrq 2023-03-03 13:53:21 +00:00
parent 2ba0e056cd
commit 06bdf72b89
1 changed files with 2 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
tortoise/api.py
View File

@ -236,13 +236,6 @@ def classify_audio_clip(clip):
results = F.softmax(classifier(clip), dim=-1)
return results[0][0]
def load_checkpoint(filepath, device):
assert os.path.isfile(filepath)
print("Loading '{}'".format(filepath))
checkpoint_dict = torch.load(filepath, map_location=device)
print("Complete.")
return checkpoint_dict
class TextToSpeech:
"""
Main entry point into Tortoise.
@ -312,10 +305,9 @@ class TextToSpeech:
self.cvvp = None # CVVP model is only loaded if used.
if use_bigvgan:
# credit https://github.com/deviandiceto / https://git.ecker.tech/mrq/ai-voice-cloning/issues/52
# credit to https://github.com/deviandice / https://git.ecker.tech/mrq/ai-voice-cloning/issues/52
self.vocoder = BigVGAN().cpu()
state_dict_bigvgan = load_checkpoint(get_model_path('bigvgan_base_24khz_100band.pth', models_dir), self.device)
self.vocoder.load_state_dict(state_dict_bigvgan['generator'])
self.vocoder.load_state_dict(torch.load(get_model_path('bigvgan_base_24khz_100band.pth', models_dir), map_location=torch.device('cpu'))['generator'])
else:
self.vocoder = UnivNetGenerator().cpu()
self.vocoder.load_state_dict(torch.load(get_model_path('vocoder.pth', models_dir), map_location=torch.device('cpu'))['model_g'])