Medix-AI/coqui-tts
Medix-AI
/
coqui-tts
mirror of https://github.com/coqui-ai/TTS.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Bug fix on XTTS get_gpt_cond_latents

This commit is contained in:
Edresson Casanova 2023-11-06 09:15:16 -03:00
parent 3e590503ae
commit 662ee2ba63
1 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
TTS/tts/models/xtts.py
View File

@ -378,8 +378,8 @@ class Xtts(BaseTTS):
length (int): Length of the audio in seconds. Defaults to 3. length (int): Length of the audio in seconds. Defaults to 3.
""" """
if sr != 22050: if sr != 22050:
audio_22k = torchaudio.functional.resample(audio, sr, 22050) audio = torchaudio.functional.resample(audio, sr, 22050)
audio_22k = audio_22k[:, : 22050 * length] audio = audio[:, : 22050 * length]
if self.args.gpt_use_perceiver_resampler: if self.args.gpt_use_perceiver_resampler:
n_fft = 2048 n_fft = 2048
hop_length = 256 hop_length = 256
@ -389,7 +389,7 @@ class Xtts(BaseTTS):
hop_length = 1024 hop_length = 1024
win_length = 4096 win_length = 4096
mel = wav_to_mel_cloning( mel = wav_to_mel_cloning(
audio_22k, audio,
mel_norms=self.mel_stats.cpu(), mel_norms=self.mel_stats.cpu(),
n_fft=n_fft, n_fft=n_fft,
hop_length=hop_length, hop_length=hop_length,