Hass/xiaozhi-esp32-server
1
0
Fork 0
mirror of https://github.com/xinnan-tech/xiaozhi-esp32-server.git synced 2026-07-30 05:13:59 +08:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #1393 from wenbindu/main

Browse Source 
修改注释
This commit is contained in:
hrz
2025-05-28 11:30:06 +08:00
committed by GitHub
parent 6a4ed78812 de6b9de341
commit 81689c04a5
1 changed files with 2 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files
main/xiaozhi-server/core/providers/vad/silero.py
+2 -3
View File
@@ -12,13 +12,12 @@ logger = setup_logging()
class VADProvider(VADProviderBase): class VADProvider(VADProviderBase):
def __init__(self, config): def __init__(self, config):
logger.bind(tag=TAG).info("SileroVAD", config) logger.bind(tag=TAG).info("SileroVAD", config)
self.model, self.utils = torch.hub.load( self.model, _ = torch.hub.load(
repo_or_dir=config["model_dir"], repo_or_dir=config["model_dir"],
source="local", source="local",
model="silero_vad", model="silero_vad",
force_reload=False, force_reload=False,
) )
(get_speech_timestamps, _, _, _, _) = self.utils
self.decoder = opuslib_next.Decoder(16000, 1) self.decoder = opuslib_next.Decoder(16000, 1)
@@ -53,7 +52,7 @@ class VADProvider(VADProviderBase):
speech_prob = self.model(audio_tensor, 16000).item() speech_prob = self.model(audio_tensor, 16000).item()
client_have_voice = speech_prob >= self.vad_threshold client_have_voice = speech_prob >= self.vad_threshold
# 如果之前有声音，但本次没有声音，且与上次有声音的时间查已经超过了静默阈值，则认为已经说完一句话 # 如果之前有声音，但本次没有声音，且与上次有声音的时间差已经超过了静默阈值，则认为已经说完一句话
if conn.client_have_voice and not client_have_voice: if conn.client_have_voice and not client_have_voice:
stop_duration = ( stop_duration = (
time.time() * 1000 - conn.client_have_voice_last_time time.time() * 1000 - conn.client_have_voice_last_time