deepgeek/silero-vad
2
0
Fork 0
mirror of https://github.com/snakers4/silero-vad.git synced 2026-02-05 18:09:22 +08:00
Code Issues Packages Projects Releases Wiki Activity

Fx

Browse Source
This commit is contained in:
snakers41
2020-12-15 12:34:42 +00:00
parent 95111b9535
commit 9f9b0dca7a
1 changed files with 4 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
utils.py
View File

@@ -38,7 +38,7 @@ def read_audio(path: str,
def save_audio(path: str, def save_audio(path: str,
tensor: torch.Tensor, tensor: torch.Tensor,
sr: int): sr: int = 16000):
torchaudio.save(path, tensor, sr) torchaudio.save(path, tensor, sr)
@@ -82,7 +82,7 @@ def get_speech_ts(wav: torch.Tensor,
outs = torch.cat(outs, dim=0) outs = torch.cat(outs, dim=0)
buffer = deque(maxlen=num_steps) # when max queue len is reached, first element is dropped buffer = deque(maxlen=num_steps) # maxlen reached => first element dropped
triggered = False triggered = False
speeches = [] speeches = []
current_speech = {} current_speech = {}
@@ -113,7 +113,7 @@ class VADiterator:
self.num_samples = 4000 self.num_samples = 4000
self.num_steps = num_steps self.num_steps = num_steps
assert self.num_samples % num_steps == 0 assert self.num_samples % num_steps == 0
self.step = int(self.num_samples / num_steps) self.step = int(self.num_samples / num_steps) # 500 samples is good enough
self.prev = torch.zeros(self.num_samples) self.prev = torch.zeros(self.num_samples)
self.last = False self.last = False
self.triggered = False self.triggered = False
@@ -144,7 +144,7 @@ class VADiterator:
self.prev = wav_chunk self.prev = wav_chunk
overlap_chunks = [stacked[i:i+self.num_samples].unsqueeze(0) overlap_chunks = [stacked[i:i+self.num_samples].unsqueeze(0)
for i in range(self.step, self.num_samples+1, self.step)] # 500 sample step is good enough for i in range(self.step, self.num_samples+1, self.step)]
return torch.cat(overlap_chunks, dim=0) return torch.cat(overlap_chunks, dim=0)
def state(self, model_out): def state(self, model_out):