dsh-voice-input-qwen-asr

Usage

1. Open Settings → Voice, press Install all (or run the four steps individually): clone runtime → clone model → create virtualenv → install deps. Watch the install log; the deps step downloads torch and may take a while. pip mirror can be configured on the same page.
2. Press Start service and wait until the state chip turns "Running" (device shows cuda:0 or cpu).
3. In any chat, click the mic button left of the send button, speak, then press Done, insert — the recognized text is appended to the input draft. "Cancel" discards the recording.