mirror of
https://github.com/SWivid/F5-TTS.git
synced 2025-12-05 20:40:12 -08:00
fix speech_edit.py
This commit is contained in:
@@ -154,7 +154,7 @@ for part in parts_to_edit:
|
||||
dim=-1,
|
||||
)
|
||||
offset = end * target_sample_rate
|
||||
# audio = torch.cat((audio_, audio[:, round(offset):]), dim = -1)
|
||||
audio = torch.cat((audio_, audio[:, round(offset) :]), dim=-1)
|
||||
edit_mask = F.pad(edit_mask, (0, audio.shape[-1] // hop_length - edit_mask.shape[-1] + 1), value=True)
|
||||
audio = audio.to(device)
|
||||
edit_mask = edit_mask.to(device)
|
||||
|
||||
Reference in New Issue
Block a user