github/ebook2audiobook
1
1
Fork 0
mirror of https://github.com/DrewThomasson/ebook2audiobook.git synced 2026-01-08 21:38:12 -05:00
Code Issues Packages Projects Releases Wiki Activity

...

Browse Source
This commit is contained in:
unknown
2025-12-25 10:41:23 -08:00
parent 002e4f415b
commit b00bf24782
1 changed files with 10 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
lib/classes/tts_engines/common/utils.py
View File

@@ -207,7 +207,16 @@ class TTSUtils:
speaker_embedding=speaker_embedding,
**fine_tuned_params,
)
audio_sentence = result.get('wav') if isinstance(result, dict) else None
if isinstance(audio_sentence, torch.Tensor):
audio_tensor = audio_sentence.detach().cpu().unsqueeze(0)
elif isinstance(audio_sentence, np.ndarray):
audio_tensor = torch.from_numpy(audio_sentence).unsqueeze(0)
elif isinstance(audio_sentence, (list, tuple)):
audio_tensor = torch.tensor(audio_sentence, dtype=torch.float32).unsqueeze(0)
else:
error = f"Unsupported XTTSv2 wav type: {type(audio_sentence)}"
print(error)
return False
if audio_sentence is not None:
audio_sentence = audio_sentence.tolist()
sourceTensor = self._tensor_type(audio_sentence)