Initialization Failed • 'NoneType' object is not callable

[ageofalgo]

When launching this repo I get this message in the frontend UI: "Initialization Failed • 'NoneType' object is not callable". Before getting the message, this other message appears briefly, for a few seconds: "Connection Error • Unable to connect to API" (I guess that's just the backend failing and stopping communication for a bit).

This is the backend health status: "{"status":"error","model_loaded":false,"device":"cpu","config":{"max_chunk_length":280,"max_total_length":3000,"voice_sample_path":"/app/voice-sample.mp3","default_exaggeration":0.5,"default_cfg_weight":0.5,"default_temperature":0.8},"memory_info":{"cpu_memory_mb":6331.9140625,"cpu_memory_percent":80.79174383808258},"initialization_state":"error","initialization_progress":"Failed: 'NoneType' object is not callable","initialization_error":"'NoneType' object is not callable"}%

I am running on cpu/mps. Currently DEVICE=auto but I tried DEVICE=mps and DEVICE=cpu and didn't change.
I mounted both containers with either uv and standard command.
I added some debug prints in tts_model.py and the initialization fails on "Loading Chatterbox Multilingual TTS model". As I saw that, in the .env I set USE_MULTILINGUAL_MODEL=false but still I keep running into the same error.
The voice-sample.mp3 was located in /root and it's now located in app/voice-sample.mp3.
As I don't use Chinese I don't have pkuseg installed, but I don't think that has anything to do with this error.

How can I solve this? Thank you

[Mikec78660]

I was getting the same error using the uv install method. Didn't matter if I used uv run uvicorn app.main:app --host 0.0.0.0 --port 4123 or uv run main.py. I switched to the pip method (option 2) and everything works fine. Not 100% sure what the issue was with the uv method. The error indicates the models wasn't loaded properly. I tried clearing the huggingface cache a few times to redownload the model files and still got the issue.

I was using device=cuda so it doesn't seem like mps is the issue.

So currently running in Debian13 LXC with Python 3.10 using pip, no venv and everything runs as it should.

[travisvn]

@ageofalgo I personally have not had good success getting things working on my MacBook Pro (M1 Pro) that I develop on

Last I checked, chatterbox basically broke the patched support for MPS once multilingual was released.

So, you might want to use the stable branch of this repository and follow those instructions / build with that code

https://github.com/travisvn/chatterbox-tts-api/tree/stable

This is a savepoint of sorts for pre-multilingual where I can confirm MPS worked

FWIW, I'd been hoping that Resemble AI would've issued a fix for Apple Silicon by this point... 😕

[travisvn]

I was getting the same error using the uv install method. Didn't matter if I used uv run uvicorn app.main:app --host 0.0.0.0 --port 4123 or uv run main.py. I switched to the pip method (option 2) and everything works fine. Not 100% sure what the issue was with the uv method. The error indicates the models wasn't loaded properly. I tried clearing the huggingface cache a few times to redownload the model files and still got the issue.

I was using device=cuda so it doesn't seem like mps is the issue.

So currently running in Debian13 LXC with Python 3.10 using pip, no venv and everything runs as it should.

Hey @Mikec78660 , you're running chatterbox-tts-api using pip and no virtual environment? Are you using conda?

[Mikec78660]

Hey @Mikec78660 , you're running chatterbox-tts-api using pip and no virtual environment? Are you using conda?

No, no conda, just installed 3.10 and used update-alternative to set it as default. I was having some issues using 3.13 that comes with debian13. There is nothing else running in the LXC so didn't have to worry about it breaking anything.

[Mikec78660]

Oh and on the issue. I tried https://github.com/travisvn/chatterbox-tts-api/tree/stable as well but still could not get it to work using uv. Funny I also tried what the OP did about changing multi language support because the errors I was getting also mentioned something about Chinese as well. But didn't change anything. Still got the 'NoneType' object is not callable issue. I actually wanted to use the pip installation anyway but only tried the uv install first for a quick setup to test it out so in the end all worked out for me. But still strange about the error. Seems like uv was installing all the same python libraries.

[travisvn]

Oh and on the issue. I tried https://github.com/travisvn/chatterbox-tts-api/tree/stable as well but still could not get it to work using uv. Funny I also tried what the OP did about changing multi language support because the errors I was getting also mentioned something about Chinese as well. But didn't change anything. Still got the 'NoneType' object is not callable issue. I actually wanted to use the pip installation anyway but only tried the uv install first for a quick setup to test it out so in the end all worked out for me. But still strange about the error. Seems like uv was installing all the same python libraries.

uv generally works great. In fact, I'd never heard of it until creating this project as it ended up being the easiest way to get chatterbox working across different environments.

But when multilingual was released, the base package offered by chatterbox would not build properly at all, particularly using uv. It required many adjustments as well as removing the problematic pkuseg package (which required core code modifications, which is why I had to fork the repository and have everyone on this project use that instead of the official repo / PyPI package)

But as to your process of trying out the stable branch and it not working for uv, I am unsure of how to help aside from advising to use docker compose and, specifically, the uv.gpu.yml version.

Unless you were in a fresh environment, you may have forgotten to copy the .env file / some of the initial instructions?

Or it could be the case that the dependencies that stable relied on have updated & cause breaking changes

EDIT: pkuseg is needed for Chinese support (forgot to include this in the message)

[Mikec78660]

I'm really happy with how I have it setup now. I had never used uv before. It seemed to work perfectly and the install went really smooth. Right up until it finishes downloading the models and tries to load them anyway.

installing pkuseg fixed my Chinese language error.

[Ch41r05]

I don't know if this helps, but I second the issue with uv.
I was on a fresh install and I double checked to be sure the .env was correctly populated with cp .env.example.docker .env.

By checking the log I see:

🎵 Using default voice for OpenAI voice 'alloy' (no alias mapping)

INFO:     172.19.0.3:44038 - "POST /v1/audio/speech HTTP/1.0" 200 OK

ERROR:    Exception in ASGI application

  + Exception Group Traceback (most recent call last):

  |   File "/app/.venv/lib/python3.11/site-packages/starlette/_utils.py", line 79, in collapse_excgroups

  |     yield

  |   File "/app/.venv/lib/python3.11/site-packages/starlette/responses.py", line 271, in __call__

  |     async with anyio.create_task_group() as task_group:

  |   File "/app/.venv/lib/python3.11/site-packages/anyio/_backends/_asyncio.py", line 781, in __aexit__

  |     raise BaseExceptionGroup(

  | ExceptionGroup: unhandled errors in a TaskGroup (1 sub-exception)

  +-+---------------- 1 ----------------

    
During handling of the above exception, another exception occurred:


Traceback (most recent call last):

  File "/app/.venv/lib/python3.11/site-packages/starlette/_exception_handler.py", line 42, in wrapped_app

    await app(scope, receive, sender)

  File "/app/.venv/lib/python3.11/site-packages/fastapi/routing.py", line 110, in app

    await response(scope, receive, send)

  File "/app/.venv/lib/python3.11/site-packages/starlette/responses.py", line 270, in __call__

    with collapse_excgroups():

  File "/usr/local/lib/python3.11/contextlib.py", line 158, in __exit__

    self.gen.throw(typ, value, traceback)

  File "/app/.venv/lib/python3.11/site-packages/starlette/_utils.py", line 85, in collapse_excgroups

    raise exc

  File "/app/.venv/lib/python3.11/site-packages/starlette/responses.py", line 274, in wrap

    await func()

  File "/app/.venv/lib/python3.11/site-packages/starlette/responses.py", line 254, in stream_response

    async for chunk in self.body_iterator:

  File "/app/app/api/endpoints/speech.py", line 596, in generate_speech_sse

    raise HTTPException(

fastapi.exceptions.HTTPException: 500: {'error': {'message': 'Model not loaded', 'type': 'model_error'}}


The above exception was the direct cause of the following exception:


Traceback (most recent call last):

  File "/app/.venv/lib/python3.11/site-packages/uvicorn/protocols/http/httptools_impl.py", line 409, in run_asgi

    result = await app(  # type: ignore[func-returns-value]

             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

  File "/app/.venv/lib/python3.11/site-packages/uvicorn/middleware/proxy_headers.py", line 60, in __call__

    return await self.app(scope, receive, send)

           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

  File "/app/.venv/lib/python3.11/site-packages/fastapi/applications.py", line 1133, in __call__

    await super().__call__(scope, receive, send)

  File "/app/.venv/lib/python3.11/site-packages/starlette/applications.py", line 113, in __call__

    await self.middleware_stack(scope, receive, send)

  File "/app/.venv/lib/python3.11/site-packages/starlette/middleware/errors.py", line 186, in __call__

    raise exc

  File "/app/.venv/lib/python3.11/site-packages/starlette/middleware/errors.py", line 164, in __call__

    await self.app(scope, receive, _send)

  File "/app/.venv/lib/python3.11/site-packages/starlette/middleware/cors.py", line 93, in __call__

    await self.simple_response(scope, receive, send, request_headers=headers)

  File "/app/.venv/lib/python3.11/site-packages/starlette/middleware/cors.py", line 144, in simple_response

    await self.app(scope, receive, send)

  File "/app/.venv/lib/python3.11/site-packages/starlette/middleware/exceptions.py", line 63, in __call__

    await wrap_app_handling_exceptions(self.app, conn)(scope, receive, send)

  File "/app/.venv/lib/python3.11/site-packages/starlette/_exception_handler.py", line 53, in wrapped_app

    raise exc

  File "/app/.venv/lib/python3.11/site-packages/starlette/_exception_handler.py", line 42, in wrapped_app

    await app(scope, receive, sender)

  File "/app/.venv/lib/python3.11/site-packages/fastapi/middleware/asyncexitstack.py", line 18, in __call__

    await self.app(scope, receive, send)

  File "/app/.venv/lib/python3.11/site-packages/starlette/routing.py", line 716, in __call__

    await self.middleware_stack(scope, receive, send)

  File "/app/.venv/lib/python3.11/site-packages/starlette/routing.py", line 736, in app

    await route.handle(scope, receive, send)

  File "/app/.venv/lib/python3.11/site-packages/starlette/routing.py", line 290, in handle

    await self.app(scope, receive, send)

  File "/app/.venv/lib/python3.11/site-packages/fastapi/routing.py", line 123, in app

    await wrap_app_handling_exceptions(app, request)(scope, receive, send)

  File "/app/.venv/lib/python3.11/site-packages/starlette/_exception_handler.py", line 56, in wrapped_app

    raise RuntimeError("Caught handled exception, but response already started.") from exc

RuntimeError: Caught handled exception, but response already started.

EDIT:

Also tried with the docker-compose.cpu.yml to check if uv was the problem, seems it's not:

/app/.venv/lib/python3.11/site-packages/diffusers/models/lora.py:393: FutureWarning: `LoRACompatibleLinear` is deprecated and will be removed in version 1.0.0. Use of `LoRACompatibleLinear` is deprecated. Please switch to PEFT backend by installing PEFT: `pip install peft`.

  deprecate("LoRACompatibleLinear", "1.0.0", deprecation_message)

WARNING:chatterbox.models.tokenizers.tokenizer:pkuseg not available - Chinese segmentation will be skipped

✗ Failed to initialize model: 'NoneType' object is not callable

[yasinarabi]

There is a problem with one of chatterbox depedencies named resemble-perth.
To solve 'NoneType' object is not callable error temporarily just modify the following line in .venv/lib/python3.11/site-packages/chatterbox/mtl_tts.py:

self.watermarker = perth.PerthImplicitWatermarker()

Replace it with

self.watermarker = perth.DummyWatermarker()

[itobey]

For the record: when running a Docker version you can add the fix of @yasinarabi to it this way. Edit the Dockerfile you are using, e.g. docker/Dockerfile.uv.gpu and add this part in the middle here:

 RUN uv pip install git+https://github.com/travisvn/chatterbox-multilingual.git@exp

# Add this:
# Patch resemble-perth NoneType bug: use DummyWatermarker instead of PerthImplicitWatermarker
RUN sed -i 's/self\.watermarker = perth\.PerthImplicitWatermarker()/self.watermarker = perth.DummyWatermarker()/' \
    /app/.venv/lib/python3.11/site-packages/chatterbox/mtl_tts.py

 # Copy application code

[poetter-sebastian]

For the record: when running a Docker version you can add the fix of @yasinarabi to it this way. Edit the Dockerfile you are using, e.g. docker/Dockerfile.uv.gpu and add this part in the middle here:

 RUN uv pip install git+https://github.com/travisvn/chatterbox-multilingual.git@exp

# Add this:
# Patch resemble-perth NoneType bug: use DummyWatermarker instead of PerthImplicitWatermarker
RUN sed -i 's/self\.watermarker = perth\.PerthImplicitWatermarker()/self.watermarker = perth.DummyWatermarker()/' \
    /app/.venv/lib/python3.11/site-packages/chatterbox/mtl_tts.py

 # Copy application code

Thank you! This helped a lot!