unslothai/unsloth · info · RuntimeError

Audio generation cancelled

Error message

Audio generation cancelled

What it means

First of three cancellation checkpoints in generate_audio: before acquiring the generation lock, the routine checks cancel_event and raises RuntimeError('Audio generation cancelled'). This gives cancellation a fast exit before prompt-neutralization and codec work, using the same sentinel message as the in-lock and post-generation checks.

Source

Thrown at studio/backend/core/inference/inference.py:2079

        """
        if not self.active_model_name:
            raise RuntimeError("No active model")

        model_info = self.models[self.active_model_name]
        audio_type = model_info.get("audio_type")
        model = model_info["model"]
        tokenizer = model_info.get("tokenizer")

        if not audio_type:
            raise RuntimeError(f"Model {self.active_model_name} is not an audio model")

        top_k = self._normalize_top_k(top_k)
        # Every codec below concatenates its prompt instead of templating it, so this
        # is the one choke point for all four (#7066).
        text = neutralize_tts_prompt_text(text, audio_type)

        if cancel_event is not None and cancel_event.is_set():
            raise RuntimeError("Audio generation cancelled")
        with self._generation_lock:
            if cancel_event is not None and cancel_event.is_set():
                raise RuntimeError("Audio generation cancelled")
            if use_adapter is not None:
                self._apply_adapter_state(use_adapter)
            stopping_criteria = self._cancel_stopping_criteria(cancel_event)

            if audio_type == "snac":
                result = self._generate_snac(
                    model,
                    tokenizer,
                    text,
                    temperature,
                    top_p,
                    max_new_tokens,
                    repetition_penalty,
                    stopping_criteria = stopping_criteria,
                    cancel_event = cancel_event,

View on GitHub (pinned to 203007d190)

Solutions

1. Handle this as normal cancellation (return 'cancelled' status), not an error
2. If unexpected, find who shares/sets this cancel_event (watchdog, sibling request)
3. Do not auto-retry; re-issue the TTS request if the user still wants the audio

if cancel_event is not None and cancel_event.is_set():
    return {"status": "cancelled"}  # skip the call entirely

try:
    wav, sr = engine.generate_audio(text, cancel_event=ev)
except RuntimeError as e:
    if str(e) == "Audio generation cancelled":
        return {"status": "cancelled"}
    raise

When it happens

Trigger: Client sets the cancel event (stop button / request abort) while the audio request is queued or just entering generate_audio, before the generation lock is acquired.

Common situations: User cancels a long TTS job immediately; deduplication logic cancelling superseded requests; client disconnect handlers flipping the event.

Related errors

• No valid audio codes found after START_OF_SPEECH token
• No bicodec_semantic tokens found in generated output
• No DAC code tokens (c1/c2) found in generated output
• Diffusion generation was cancelled.
• Model {self.active_model_name} is not an audio model

AI-assisted analysis of unslothai/unsloth@203007d190 (2026-08-15). Data as JSON: /api/errors/d977f4a38928ca58. Report an issue: GitHub.