fishaudio/fish-speech · critical · TypeError

Expected GenerateResponse, got {type(wrapped_result.response

Error message

Expected GenerateResponse, got {type(wrapped_result.response).__name__}

What it means

InferenceEngine.inference wraps the underlying model's streaming responses and asserts each response is a GenerateResponse instance; any other type coming through the response channel is treated as a protocol violation between the engine layers and raises a TypeError.

Source

Thrown at fish_speech/inference_engine/__init__.py:103

        while True:
            # Get the response from the LLAMA model
            wrapped_result: WrappedGenerateResponse = response_queue.get()
            if wrapped_result.status == "error":
                yield InferenceResult(
                    code="error",
                    audio=None,
                    error=(
                        wrapped_result.response
                        if isinstance(wrapped_result.response, Exception)
                        else Exception("Unknown error")
                    ),
                )
                break

            # Check the response type
            if not isinstance(wrapped_result.response, GenerateResponse):
                raise TypeError(
                    f"Expected GenerateResponse, got {type(wrapped_result.response).__name__}"
                )

            result: GenerateResponse = wrapped_result.response
            if result.action != "next":
                segment = self.get_audio_segment(result)

                if req.streaming:  # Used only by the API server
                    yield InferenceResult(
                        code="segment",
                        audio=(sample_rate, segment),
                        error=None,
                    )
                segments.append(segment)
            else:
                break

        # Clean up the memory

View on GitHub (pinned to befe400174)

Solutions

Pin/align client and server to the same fish-speech version (reinstall both sides)
If using a custom or mocked backend, make it return GenerateResponse instances exactly
Update the code — this check exists precisely to catch protocol drift, so don't suppress it; fix the source of the wrong type

Defensive patterns

Strategy: validation

Validate before calling

from fish_speech import GenerateResponse  # adjust import to actual module
resp = engine.inference(...)
# pre-check versions before inferring
import fish_speech
assert fish_speech.__version__ == SERVER_VERSION

Type guard

from fish_speech import GenerateResponse

def is_generate_response(r) -> bool:
    return isinstance(r, GenerateResponse)

Try / catch

try:
    for chunk in engine.inference(req):
        ...
except TypeError as e:
    if "Expected GenerateResponse" in str(e):
        raise RuntimeError("client/server version mismatch — align fish-speech versions") from e
    raise

Prevention

Pin identical fish-speech versions on client and server
Add a version/protocol handshake at connection setup
Don't mock backends with plain dicts; return real GenerateResponse objects

When it happens

Trigger: Calling engine.inference()/inference_async()/tts()/warm_up() when the wrapped backend (model server or local model) yields a response object of an unexpected type — e.g. after a version mismatch where the backend returns a raw dataclass/dict instead of GenerateResponse.

Common situations: Mixing incompatible versions of fish-speech client and server (schema drift in the RPC/message protocol), a mocked or custom backend not returning the expected class, or deserialization producing a different class with the same shape.

Related errors

AI-assisted analysis of fishaudio/fish-speech@befe400174 (2026-08-27). Data as JSON: /api/errors/142685dd5dc7cb86. Report an issue: GitHub.