speech_recognition/FunASR
1
0
Fork 0
mirror of https://github.com/modelscope/FunASR synced 2025-09-15 14:48:36 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

vad bug fix

Browse Source
This commit is contained in:
凌匀 2023-04-21 21:40:11 +08:00
parent 7ae8df96fc
commit 7a7ead00bc
2 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
funasr/bin/vad_inference.py
View File

@ -109,7 +109,7 @@ class Speech2VadSegment:
fbanks, fbanks_len = self.frontend.forward_fbank(speech, speech_lengths)
feats, feats_len = self.frontend.forward_lfr_cmvn(fbanks, fbanks_len)
fbanks = to_device(fbanks, device=self.device)
# feats = to_device(feats, device=self.device)
feats = to_device(feats, device=self.device)
feats_len = feats_len.int()
else:
raise Exception("Need to extract feats first, please configure frontend configuration")
@ -131,7 +131,7 @@ class Speech2VadSegment:
"in_cache": in_cache
}
# a. To device
batch = to_device(batch, device=self.device)
#batch = to_device(batch, device=self.device)
segments_part, in_cache = self.vad_model(**batch)
if segments_part:
for batch_num in range(0, self.batch_size):

2
funasr/models/frontend/wav_frontend.py
View File

@ -34,7 +34,7 @@ def load_cmvn(cmvn_file):
means = np.array(means_list).astype(np.float)
vars = np.array(vars_list).astype(np.float)
cmvn = np.array([means, vars])
cmvn = torch.as_tensor(cmvn, dype=torch.float32)
cmvn = torch.as_tensor(cmvn, dtype=torch.float32)
return cmvn