From 219c2482ab755fbd4e49dfbdee91bf1a8a4ec49a Mon Sep 17 00:00:00 2001
From: 游雁 <zhifu.gzf@alibaba-inc.com>
Date: 星期五, 19 五月 2023 11:33:27 +0800
Subject: [PATCH] websocket 2pass bugfix
---
funasr/models/frontend/wav_frontend.py | 34 ++++++++++++++++------------------
1 files changed, 16 insertions(+), 18 deletions(-)
diff --git a/funasr/models/frontend/wav_frontend.py b/funasr/models/frontend/wav_frontend.py
index f61d7dd..35fab57 100644
--- a/funasr/models/frontend/wav_frontend.py
+++ b/funasr/models/frontend/wav_frontend.py
@@ -34,11 +34,11 @@
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)
+ cmvn = torch.as_tensor(cmvn, dtype=torch.float32)
return cmvn
-def apply_cmvn(inputs, cmvn_file): # noqa
+def apply_cmvn(inputs, cmvn): # noqa
"""
Apply CMVN with mvn data
"""
@@ -47,11 +47,10 @@
dtype = inputs.dtype
frame, dim = inputs.shape
- cmvn = load_cmvn(cmvn_file)
- means = np.tile(cmvn[0:1, :dim], (frame, 1))
- vars = np.tile(cmvn[1:2, :dim], (frame, 1))
- inputs += torch.from_numpy(means).type(dtype).to(device)
- inputs *= torch.from_numpy(vars).type(dtype).to(device)
+ means = cmvn[0:1, :dim]
+ vars = cmvn[1:2, :dim]
+ inputs += means.to(device)
+ inputs *= vars.to(device)
return inputs.type(torch.float32)
@@ -111,6 +110,7 @@
self.dither = dither
self.snip_edges = snip_edges
self.upsacle_samples = upsacle_samples
+ self.cmvn = None if self.cmvn_file is None else load_cmvn(self.cmvn_file)
def output_size(self) -> int:
return self.n_mels * self.lfr_m
@@ -140,8 +140,8 @@
if self.lfr_m != 1 or self.lfr_n != 1:
mat = apply_lfr(mat, self.lfr_m, self.lfr_n)
- if self.cmvn_file is not None:
- mat = apply_cmvn(mat, self.cmvn_file)
+ if self.cmvn is not None:
+ mat = apply_cmvn(mat, self.cmvn)
feat_length = mat.size(0)
feats.append(mat)
feats_lens.append(feat_length)
@@ -194,8 +194,8 @@
mat = input[i, :input_lengths[i], :]
if self.lfr_m != 1 or self.lfr_n != 1:
mat = apply_lfr(mat, self.lfr_m, self.lfr_n)
- if self.cmvn_file is not None:
- mat = apply_cmvn(mat, self.cmvn_file)
+ if self.cmvn is not None:
+ mat = apply_cmvn(mat, self.cmvn)
feat_length = mat.size(0)
feats.append(mat)
feats_lens.append(feat_length)
@@ -423,10 +423,8 @@
reserve_frame_idx = lfr_splice_frame_idxs[0] - minus_frame
# print('reserve_frame_idx: ' + str(reserve_frame_idx))
# print('frame_frame: ' + str(frame_from_waveforms))
- self.reserve_waveforms = self.waveforms[:,
- reserve_frame_idx * self.frame_shift_sample_length:frame_from_waveforms * self.frame_shift_sample_length]
- sample_length = (
- frame_from_waveforms - 1) * self.frame_shift_sample_length + self.frame_sample_length
+ self.reserve_waveforms = self.waveforms[:, reserve_frame_idx * self.frame_shift_sample_length:frame_from_waveforms * self.frame_shift_sample_length]
+ sample_length = (frame_from_waveforms - 1) * self.frame_shift_sample_length + self.frame_sample_length
self.waveforms = self.waveforms[:, :sample_length]
else:
# update self.reserve_waveforms and self.lfr_splice_cache
@@ -488,10 +486,10 @@
for i in range(batch_size):
waveform_length = input_lengths[i]
waveform = input[i][:waveform_length]
- waveform = waveform.unsqueeze(0).numpy()
+ waveform = waveform.numpy()
mat = eend_ola_feature.stft(waveform, self.frame_length, self.frame_shift)
mat = eend_ola_feature.transform(mat)
- mat = mat.splice(mat, context_size=self.lfr_m)
+ mat = eend_ola_feature.splice(mat, context_size=self.lfr_m)
mat = mat[::self.lfr_n]
mat = torch.from_numpy(mat)
feat_length = mat.size(0)
@@ -502,4 +500,4 @@
feats_pad = pad_sequence(feats,
batch_first=True,
padding_value=0.0)
- return feats_pad, feats_lens
+ return feats_pad, feats_lens
\ No newline at end of file
--
Gitblit v1.9.1