generate_audio2audio: FE process use loop_record thread and loop_play thread; BE process use loop_asr_llm_generate thread and loop_tts_synthesize thread with text_buffer queue. FE process and BE process with pipe connection (unix socket)
fix log: add is_file and is_console set write file or std.out
add verbose info control for llm_llamacpp and whisper_timestamped_asr;
set TQDM_DISABLE=True to disable tqdm for ChatTTS infer
weedge
commented
3 months ago
Optimize local-terminal-chat：
TQDM_DISABLE=Trueto disabletqdmfor ChatTTS infer