Hass/xiaozhi-esp32-server
1
0
Fork 0
mirror of https://github.com/xinnan-tech/xiaozhi-esp32-server.git synced 2026-07-28 01:53:53 +08:00
Code Issues Packages Projects Releases Wiki Activity

fix:manager意图识别bug (#762)

Browse Source 
* fix:连接manager后无法使用functioncallbug

* fix:意图识别使用llm无法播放音乐bug

* fix：manager第一图识别使用独立llm无法初始化llm的bug
This commit is contained in:
hrz
2025-04-13 18:10:17 +08:00
committed by GitHub
parent 056659304a
commit de8c762d79
15 changed files with 107 additions and 45 deletions
Download Patch File Download Diff File Expand all files Collapse all files
main/xiaozhi-server/core/providers/vad/silero.py
+2 -1
View File
@@ -42,7 +42,8 @@ class VADProvider(VADProviderBase):
audio_tensor = torch.from_numpy(audio_float32)
# 检测语音活动
speech_prob = self.model(audio_tensor, 16000).item()
with torch.no_grad():
speech_prob = self.model(audio_tensor, 16000).item()
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