admin-valentin/WhisperLiveKit
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Fix regression with faster-whisper encoder_feature

Browse source
This commit is contained in:
Alexander Lindberg 2025-09-09 11:18:55 +03:00
parent f661f21675
commit c7b3bb5e58
1 changed files with 1 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
whisperlivekit/simul_whisper/simul_whisper.py
View file

@ -408,7 +408,7 @@ class PaddedAlignAttWhisper:
content_mel_len = int(audio_length_seconds * 100)//2
mel_padded_2 = self.fw_feature_extractor(waveform=input_segments.numpy(), padding=N_SAMPLES)[None, :]
mel = fw_pad_or_trim(mel_padded_2, N_FRAMES, axis=-1)
encoder_feature_ctranslate = np.array(self.fw_encoder.encode(mel))
encoder_feature_ctranslate = self.fw_encoder.encode(mel)
encoder_feature = torch.as_tensor(encoder_feature_ctranslate, device=self.device)
else:
# mel + padding to 30s