christian-byrne / audio-separation-nodes-comfyui

Separate stems (vocals, bass, drums, other) from audio. Recombine, tempo match, slice/crop audio
148 stars 11 forks source link

Got an unexpected keyword argument 'sparse' when executing AudioTempoMatch #5

Closed rodsott closed 2 weeks ago

rodsott commented 1 month ago

When running the remix-songs.json file, it gave me this error:

`Error occurred when executing AudioTempoMatch:

beat_track() got an unexpected keyword argument 'sparse'

File "X:\AI\ComfyUI_windows_portable\ComfyUI\execution.py", line 151, in recursive_execute output_data, output_ui = get_output_data(obj, input_data_all) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "X:\AI\ComfyUI_windows_portable\ComfyUI\execution.py", line 81, in get_output_data return_values = map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "X:\AI\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-0246\utils.py", line 381, in new_func res_value = old_func(final_args, kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "X:\AI\ComfyUI_windows_portable\ComfyUI\execution.py", line 74, in map_node_over_list results.append(getattr(obj, func)(slice_dict(input_data_all, i))) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "X:\AI\ComfyUI_windows_portable\ComfyUI\custom_nodes\audio-separation-nodes-comfyui\src\tempo_match.py", line 32, in main tempo_1 = estimate_tempo(waveform_1, input_sample_rate_1) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "X:\AI\ComfyUI_windows_portable\ComfyUI\custom_nodes\audio-separation-nodes-comfyui\src\utils.py", line 90, in estimatetempo tempo, = librosa.beat.beat_track( ^^^^^^^^^^^^^^^^^^^^^^^^ File "X:\AI\ComfyUI_windows_portable\python_embeded\Lib\site-packages\librosa\util\decorators.py", line 88, in inner_f return f(args, **kwargs) ^^^^^^^^^^^^^^^^^^`

Any idea @christian-byrne ? Please let me know which info you need from my side.

Thanks in advance,

RoD

christian-byrne commented 1 month ago

Hello RoD. Can you verify that you have librosa version 0.10.2 or higher? You can check by running:

pip show librosa

If the version you have is behind 0.10.2 try

pip install librosa

Let me know if that works, thanks.

rodsott commented 1 month ago

Hi @christian-byrne , Thank you for your feedback!

Yes, I was with librosa 0.9.2.dist. Just removed it and installed the latest 0.10.2. Now the workflow worked until the end, but gave me the following error in the console, which wasn`t happening before:

Exception in callback _ProactorBasePipeTransport._call_connection_lost(None) handle: <Handle _ProactorBasePipeTransport._call_connection_lost(None)> Traceback (most recent call last): File "asyncio\events.py", line 84, in _run File "asyncio\proactor_events.py", line 165, in _call_connection_lost ConnectionResetError: [WinError 10054]

Any clue on what could be giving me this?

Thanks in advance!

RoD

christian-byrne commented 1 month ago

Hi @christian-byrne , Thank you for your feedback!

Yes, I was with librosa 0.9.2.dist. Just removed it and installed the latest 0.10.2. Now the workflow worked until the end, but gave me the following error in the console, which wasn`t happening before:

Exception in callback _ProactorBasePipeTransport._call_connection_lost(None) handle: <Handle _ProactorBasePipeTransport._call_connection_lost(None)> Traceback (most recent call last): File "asyncio\events.py", line 84, in _run File "asyncio\proactor_events.py", line 165, in _call_connection_lost ConnectionResetError: [WinError 10054]

Any clue on what could be giving me this?

Thanks in advance!

RoD

No problem, happy to help.

For the new error, it seems to be related to an issue with the aiohttp backend on Windows . Can you try to run just the default workflow and see if it works? Also, are you able to share the full log output?

rodsott commented 1 month ago

Sure, it works but still gives that error in the end..

https://github.com/user-attachments/assets/940534d4-7978-4237-960a-41a25468b329

it give this "An existing connection was forced to be canceled by the remote host". I used to have a similar one when I tried some LLM to connect to local Ollama.

christian-byrne commented 1 month ago

Sure, it works but still gives that error in the end..

NVIDIA_Share_Q676xKkFjp.mp4 it give this "An existing connection was forced to be canceled by the remote host". I used to have a similar one when I tried some LLM to connect to local Ollama.

I think the issue is caused by another extension like Mixlab or ComfyUI-Manager which creates network connections. If you click the Load Default button in the menu and try to run the workflow does it still happen?

rodsott commented 1 month ago

no, running the default it did not give the error

image

Any idea of what I could do to pin point the origin of this error? I just saw that it happened after installing your custom node.

Thank you once again for your time, your node is quite awesome! Last year would have been very useful for a project that I needed to split the tracks, then I used the BandLab Splitter.. now we be able to do locally on Comfy will be great!

rodsott commented 1 month ago

btw, if this can help, I`m running the portable Comfy with the following setup:

Platform: Windows 10 Python version: 3.11.8 pytorch version: 2.3.0+cu118 xformers version: 0.0.26.post1+cu118

With both CUDA Runtime and Development 11.8 modules installed on windows.

The only warning I have on my console, but this is old and never managed to fix is:

_Warn!: From X:\AI\ComfyUI_windows_portable\python_embeded\Lib\site-packages\tf_keras\src\losses.py:2976: The name tf.losses.sparse_softmax_cross_entropy is deprecated. Please use tf.compat.v1.losses.sparse_softmax_crossentropy instead.

I'm using tf_keras 2.16.0, but last week they released the 2.17.0 but didn't updated it yet. =)

Which protobuf you`re using?

christian-byrne commented 1 month ago

btw, if this can help, I`m running the portable Comfy with the following setup:

Platform: Windows 10 Python version: 3.11.8 pytorch version: 2.3.0+cu118 xformers version: 0.0.26.post1+cu118

With both CUDA Runtime and Development 11.8 modules installed on windows.

The only warning I have on my console, but this is old and never managed to fix is:

_Warn!: From X:\AI\ComfyUI_windows_portable\python_embeded\Lib\site-packages\tf_keras\src\losses.py:2976: The name tf.losses.sparse_softmax_cross_entropy is deprecated. Please use tf.compat.v1.losses.sparse_softmax_crossentropy instead.

I'm using tf_keras 2.16.0, but last week they released the 2.17.0 but didn't updated it yet. =)

Which protobuf you`re using?

Thanks for the extra details! It sounds like other users have seen this error too. It seems the connection gets closed, this could be due to:

I'm not sure if this is the main connection or from another extension, or whether the connection is re-established. Are you able to keep using Comfy after that error occurs or does the program crash completely?

christian-byrne commented 2 weeks ago

Other people have said that the ConnectionResetError can happen when a socket connection is closed while still writing preview data. I think it is a minor error that doesn't break the program.

Since all issues related to this repo seem to be resolved, I will close this issue.