 wenhongquan
|
0f736e4761
perf(worker): preload VAD at startup, skip Qwen3 when using Mimo ASR
|
há 3 meses atrás |
 wenhongquan
|
c278751721
cleanup: fix Dockerfile COPY path, remove pycache
|
há 3 meses atrás |
|
wenhongquan
|
c92b1edf2c
cleanup: merge worker files, remove simplified version
|
há 3 meses atrás |
|
wenhongquan
|
120b17c900
fix(dispatcher): import Worker from worker (not conversation_worker)
|
há 3 meses atrás |
|
wenhongquan
|
9bf8fc264b
fix: switch dispatcher port to 10005 (10002 occupied by dshai)
|
há 3 meses atrás |
|
wenhongquan
|
28d10436e7
fix(dispatcher): nginx proxy on port 10002, external URL for Flutter
|
há 3 meses atrás |
|
wenhongquan
|
821da4ce0c
feat(dispatcher): multi-user room isolation + Flutter dispatcher client
|
há 3 meses atrás |
|
wenhongquan
|
77a287dd55
fix(dispatcher): add PUBLIC_LIVEKIT_URL for external app access
|
há 3 meses atrás |
|
wenhongquan
|
c83b05f89a
fix: dispatcher port 9100
|
há 3 meses atrás |
|
wenhongquan
|
baa00cd81f
fix: change dispatcher port to 9001 (9000 occupied)
|
há 3 meses atrás |
|
wenhongquan
|
d1afb91175
build(android): 配置安卓签名和混淆规则，更新默认连接地址
|
há 3 meses atrás |
|
wenhongquan
|
29e3519f87
feat(dispatcher): multi-user room management service
|
há 3 meses atrás |
|
wenhongquan
|
0b3134495e
fix(tts+asr): parallel TTS with ordered enqueue + filter single-char ASR filler
|
há 3 meses atrás |
|
wenhongquan
|
34cb936a5e
fix(tts): ordered TTS generation - collect segments first, generate sequentially
|
há 3 meses atrás |
|
wenhongquan
|
c80cfcdb22
fix(tts): keep audio int16, fix indentation
|
há 3 meses atrás |
|
wenhongquan
|
cf28ff0576
fix(vad): model path aligned with Dockerfile (whisper_asr/silero_vad.onnx)
|
há 3 meses atrás |
|
wenhongquan
|
4b1a2ae8db
fix: restore _send method, move _drain_queue before __main__
|
há 3 meses atrás |
|
wenhongquan
|
beea2f5be1
fix: move _drain_queue to module level after Worker class
|
há 3 meses atrás |
|
wenhongquan
|
f648952a0a
fix: indent _drain_queue and _send inside Worker class
|
há 3 meses atrás |
|
wenhongquan
|
7a5790c994
refactor(tts): single persistent track + queue playback + gen-based interrupt
|
há 3 meses atrás |
|
wenhongquan
|
ca38fff29c
fix(asr): concatenate streaming ASR chunks instead of keeping only last
|
há 3 meses atrás |
|
wenhongquan
|
18e3e4d341
fix(flutter): display streaming AI reply (liveReply) in conversation list
|
há 3 meses atrás |
|
wenhongquan
|
1ba163922d
feat(tts-asr): streaming Mimo TTS (voiceclone) + streaming ASR
|
há 3 meses atrás |
|
wenhongquan
|
a381dc1067
feat(provider): add LLM/ASR provider switch (vllm/mimo, qwen/mimo)
|
há 3 meses atrás |
|
wenhongquan
|
fbe9284c3e
revert: remove partial ASR (CPU thrashing hurt latency)
|
há 3 meses atrás |
|
wenhongquan
|
3d398e2baf
build(asr_agent): add onnxruntime dependency
|
há 3 meses atrás |
|
wenhongquan
|
b3de0309a2
perf(asr): streaming partial ASR during speech + reduce VAD silence to 0.4s
|
há 3 meses atrás |
|
wenhongquan
|
6afa008037
feat(tts): concurrent TTS playback + user interruption (xiaozhi-style abort)
|
há 3 meses atrás |
|
wenhongquan
|
57e325ab76
feat(vad): bundle Silero VAD ONNX model (2.3MB) in asr_agent/
|
há 3 meses atrás |
|
wenhongquan
|
5cec1bb5ae
chore: add .gitignore
|
há 3 meses atrás |
