diff options
Diffstat (limited to 'converter/tts.py')
| -rw-r--r-- | converter/tts.py | 18 |
1 files changed, 9 insertions, 9 deletions
diff --git a/converter/tts.py b/converter/tts.py index 83284a7..9b54cf4 100644 --- a/converter/tts.py +++ b/converter/tts.py @@ -1,6 +1,6 @@ """Client wrappers for the TTS backends. -QwenTTSClient talks to the Qwen3-TTS Gradio demos (custom voice / voice clone). +QwenTTSClient talks to the Qwen3-TTS demo server (custom voice / voice clone). FasterTTSClient talks to the OpenAI-compatible server from the faster-qwen3-tts repository (voice cloning only; the reference voice is configured server-side — see the "Faster backend" section of the README). @@ -40,10 +40,10 @@ VOICE_MODE_CLONE = "voice_clone" VOICE_MODES = (VOICE_MODE_CUSTOM, VOICE_MODE_CLONE) # TTS backends (re-exported for the CLI and the converter orchestrator). -BACKEND_GRADIO = "gradio" +BACKEND_QWEN = "qwen" BACKEND_FASTER = "faster" BACKEND_AUDIOCPP = "audiocpp" -BACKENDS = (BACKEND_GRADIO, BACKEND_FASTER, BACKEND_AUDIOCPP) +BACKENDS = (BACKEND_AUDIOCPP, BACKEND_QWEN, BACKEND_FASTER) # Languages understood by the Qwen3-TTS API. Display names must match the # demo dropdown exactly (the demo silently falls back to "Auto" for @@ -188,7 +188,7 @@ def _resolve_request_seed() -> int: def speaker_display_name() -> str: - """Return the Gradio display name for the configured custom speaker.""" + """Return the display name for the configured custom speaker.""" return SPEAKER_DISPLAY_NAMES.get( config.SPEAKER.lower(), config.SPEAKER) @@ -343,7 +343,7 @@ class _BaseTTSClient: class QwenTTSClient(_BaseTTSClient): - """Generates audio chunks through a Qwen3-TTS Gradio server.""" + """Generates audio chunks through a Qwen3-TTS demo server.""" def __init__(self, voice_mode: str = "custom_voice", voice_clone_ref_audio: Optional[str] = None, voice_clone_ref_text: Optional[str] = None, skip_transcription: bool = False, @@ -392,7 +392,7 @@ class QwenTTSClient(_BaseTTSClient): except Exception as exc: raise RuntimeError( f"Qwen API initialization failed at {api_url}: {exc}. " - "Make sure the Qwen Gradio server is running and reachable, and that your " + "Make sure the Qwen demo server is running and reachable, and that your " "installed Qwen3-TTS version matches this converter's API expectations " "(voice clone requires the Base-model demo: Qwen/Qwen3-TTS-12Hz-1.7B-Base)." ) from exc @@ -635,7 +635,7 @@ class FasterTTSClient(_BaseTTSClient): repository (examples/openai_server.py). The reference voice (ref audio, ref text) and language are configured on the server itself via --ref-audio/--ref-text or a --voices JSON file; this client only sends - text. Unlike the Gradio demo, the server performs one generation per + text. Unlike the Qwen demo, the server performs one generation per request, so long chunks are sub-chunked client-side. """ @@ -779,7 +779,7 @@ class AudioCppTTSClient(_BaseTTSClient): printed by the CLI. Each response is a complete WAV file, so sub-request audio is - concatenated with the same lossless path used for the Gradio client. + concatenated with the same lossless path used for the Qwen client. """ def __init__(self, voice: Optional[str] = None, language: Optional[str] = None, @@ -796,7 +796,7 @@ class AudioCppTTSClient(_BaseTTSClient): self.language = normalize_language( language if language is not None else config.LANGUAGE) # One seed value per run, reused for every request (see - # _resolve_request_seed). Unlike the Gradio demo, audio.cpp has no + # _resolve_request_seed). Unlike the Qwen demo, audio.cpp has no # negative "randomize" seed, so a negative value means "send no seed # at all" (see _request_wav) and the server randomizes. self._seed = _resolve_request_seed() |