[ISSUE]: Okada RVC Kaggle Environment is Not Running

[bennett-krubakar]

Voice Changer Version

Not sure

Operational System

Kaggle

GPU

P100

Read carefully and check the options

• [ ] I've tried to Clear Settings
• [ ] Sample/Default Models are working
• [ ] I've tried to change the Chunk Size
• [ ] GUI was successfully launched
• [ ] I've read the tutorial
• [ ] I've tried to extract to another folder (or re-extract) the .zip file

Model Type

RVC

Issue Description

Description: The Okada RVC environment on Kaggle has been non-functional for the past few days. It seems to be encountering issues with starting the server. The error messages have changed recently, indicating that there might be ongoing maintenance or changes to the environment.

Current Behavior: The Okada RVC Kaggle environment is not running, and it fails to start the server.

Additional Information: It's worth noting that the error messages recently changed, which suggests ongoing updates or maintenance. Any information on these changes or a timeline for when the environment will be functional again would be greatly appreciated.

Labels: bug, kaggle, environment, server

Requesting help from: @Rafacasari

Application Screenshot

Screenshots/Logs:

Recent error message(s) while trying to start the server.

Error messages encountered during the installation of dependencies for reference.

Logs on console

/kaggle/working/voice-changer/server
<frozen importlib._bootstrap>:241: RuntimeWarning: scipy._lib.messagestream.MessageStream size changed, may indicate binary incompatibility. Expected 56 from C header, got 64 from PyObject
^C
Traceback (most recent call last):
  File "/kaggle/working/voice-changer/server/MMVCServerSIO.py", line 19, in <module>
    from sio.MMVC_SocketIOApp import MMVC_SocketIOApp
  File "/kaggle/working/voice-changer/server/sio/MMVC_SocketIOApp.py", line 1, in <module>
    import socketio
  File "/opt/conda/lib/python3.10/site-packages/socketio/__init__.py", line 3, in <module>
    from .client import Client
  File "/opt/conda/lib/python3.10/site-packages/socketio/client.py", line 7, in <module>
    import engineio
  File "/opt/conda/lib/python3.10/site-packages/engineio/__init__.py", line 8, in <module>
    from .asyncio_client import AsyncClient
  File "/opt/conda/lib/python3.10/site-packages/engineio/asyncio_client.py", line 7, in <module>
    import aiohttp
  File "/opt/conda/lib/python3.10/site-packages/aiohttp/__init__.py", line 6, in <module>
    from .client import (
  File "/opt/conda/lib/python3.10/site-packages/aiohttp/client.py", line 32, in <module>
    import attr
  File "/opt/conda/lib/python3.10/site-packages/attr/__init__.py", line 26, in <module>
    from ._version_info import VersionInfo
  File "<frozen importlib._bootstrap>", line 1027, in _find_and_load
  File "<frozen importlib._bootstrap>", line 1006, in _find_and_load_unlocked
  File "<frozen importlib._bootstrap>", line 688, in _load_unlocked
  File "<frozen importlib._bootstrap_external>", line 879, in exec_module
  File "<frozen importlib._bootstrap_external>", line 975, in get_code
  File "<frozen importlib._bootstrap_external>", line 1074, in get_data
KeyboardInterrupt

[Rafacasari]

I think Colab and Kaggle banned WebUIs/tunnels like Gradio, Ngrok, but not sure

[bennett-krubakar]

@Rafacasari do you think Amazon SageMaker Studio lab will still work for WebUIs/tunnels?

[Rafacasari]

I think Amazon is paid, no? The best paid option is paperspace, I think.

But idk about it too, I've never used amazon studio

[bennett-krubakar]

Sagemake is free. We get 4 hrs of GPU time in a day, and 8 hrs of CPU time in a day with a runtime of max 4 hrs per session. With Paperspace I think you get 8 hrs per day. But overall, I've heard Sagemaker GPU is stronger than Paperspace. But if we can get Okada RVC running in either of them, that would be great.

[bennett-krubakar]

@Rafacasari I have checked out both Sagemaker and Paperspace. From what I have found, Paperspace is technically free but practically, it is not. Paperspace only has one free GPU option and is unavailable primarily due to its high demand. While Paperspace may provide more GPU runtime hours per day than Sagemaker, paperspace is not free(Practically). Sagemaker has an odd interface compared to other notebooks but it's not complicated. I think Sagemaker is the best option we have at hand.

I would like to request your help to create a notebook for Sagemaker to run okada-RVC.

[Rafacasari]

I have tried for some minutes to run the notebooks that we already have, but seems that Amazon will have some specific issues that I'll have to fix. I will try again later.

[bennett-krubakar]

@Rafacasari Thank you for looking into this. If you have any updates on Sagemaker, I would love to know about it.

[DaUnknow1]

Amazon_SageMaker_Studio(Lab)_RealtimeVoiceChanger.zip I made one that works with Amazon SageMaker Studio Lab but I don't know how to merge it into master if you can merge it thanks.

[bennett-krubakar]

I don't think what I've done is right but give it a gander and see what I should be changing for this to work.

[DaUnknow1]

It looks like you got your directory wrong. I think your directory is /home/studio-lab-user/sagemaker-studiolab-notebooks/voice-changer/server if it doesn't work try and find where the server folder is and right-click and copy the path then use %cd /home/studio-lab-user/[Copied path]. Also, you need the !python3 MMVCServerSIO.py and the ngrok.disconnect(ngrokConnection.public_url) needs to be in the same cell as the above. If you need an easier way just download the zip and unzip the file click upload file put in your token and just run every cell and should run properly FYI next time you start it you can just run the last cell Amazon_SageMaker_Studio(Lab)_RealtimeVoiceChanger.zip

[kreindo]

I can't get this to work for some reason

[DaUnknow1]

Send a screenshot of your problem

[razzytest61]

https://www.kaggle.com/code/razzytest/public-w-okada-voice-changer

100 working voice changers