The script encountered an issue while finding the module specification for 'main' and a socket connection problem

[Blacksingh]

It seems that the script encountered an issue while finding the module specification for 'main' and a socket connection problem.

Happens when I try to train

(so-vitts-fork) PS C:\Users\jerme\anaconda3\envs\so-vitts-fork> svc train [18:56:14] INFO [18:56:14] Version: 3.8.0 main.py:22 [18:56:18] INFO [18:56:18] Using strategy: ddp_find_unused_parameters_true train.py:82 INFO: GPU available: True (cuda), used: True INFO [18:56:18] GPU available: True (cuda), used: True rank_zero.py:48 INFO: TPU available: False, using: 0 TPU cores INFO [18:56:18] TPU available: False, using: 0 TPU cores rank_zero.py:48 INFO: IPU available: False, using: 0 IPUs INFO [18:56:18] IPU available: False, using: 0 IPUs rank_zero.py:48 INFO: HPU available: False, using: 0 HPUs INFO [18:56:18] HPU available: False, using: 0 HPUs rank_zero.py:48 WARNING [18:56:18] warnings.py:109 C:\Users\jerme\anaconda3\envs\so-vitts-fork\lib\site-packages\so_vits_svc_fork\modul es\synthesizers.py:81: UserWarning: Unused arguments: {'n_layers_q': 3, 'use_spectral_norm': False} warnings.warn(f"Unused arguments: {kwargs}")

       INFO     [18:56:18] Decoder type: hifi-gan                                                synthesizers.py:100

[18:56:19] WARNING [18:56:19] warnings.py:109 C:\Users\jerme\anaconda3\envs\so-vitts-fork\lib\site-packages\so_vits_svc_fork\utils .py:187: UserWarning: Keys not found in checkpoint state dict:['emb_g.weight'] warnings.warn(f"Keys not found in checkpoint state dict:" f"{not_in_from}")

       INFO     [18:56:19] Loaded checkpoint 'logs\44k\G_0.pth' (iteration 0)                           utils.py:244
       INFO     [18:56:19] Loaded checkpoint 'logs\44k\D_0.pth' (iteration 0)                           utils.py:244

INFO: Initializing distributed: GLOBAL_RANK: 0, MEMBER: 1/2 INFO [18:56:19] Initializing distributed: GLOBAL_RANK: 0, MEMBER: 1/2 distributed.py:244 [W ..\torch\csrc\distributed\c10d\socket.cpp:601] [c10d] The client socket has failed to connect to [kubernetes.docker.internal]:57680 (system error: 10049 - The requested address is not valid in its context.). C:\Users\jerme\anaconda3\envs\so-vitts-fork\python.exe: Error while finding module specification for 'main' (ValueError: main.spec is None)

Tensor board website pops up and says "No dashboards are active for the current data set."

[Blacksingh]

Previously I was trying to install on anaconda in windows 11. My GPU is rtx3090 and needs to run in Linux Ubuntu because of CUDA and NCCL in order to train.

[34j]

@allcontributors add Blacksingh bug

[allcontributors[bot]]

@34j

I've put up a pull request to add @Blacksingh! :tada:

[Byron-Zoomry]

@Blacksingh I'm having this same issue and am unable to train. Were you able to resolve this somehow?

[Blacksingh]

Not really tbh Try fresh with Ubuntu and install Nvidia Driver Nvidia CUDA Nvidia cudnn Python Tensorflow Pytorch Then go from there

[Byron-Zoomry]

I'm on windows. I'm already pretty in over my head trying to figure out all this github stuff to also try to install linux. Has Ubuntu given you better results?

[Blacksingh]

Yea I was in the same boat, in over my head lol But if you have an Nvidia card you’re gonna need to go the Ubuntu route, because I don’t think there’s windows support for some of the stuff I listed there for the cards to run proper for training models.When I did that, it was the closest I got, then I tried to delete n reinstall Ubuntu n everything n then I couldn’t even download PyTorch again properly so idk I had to take a break cuz of the stress it was giving me n being unable to solve Prolly gonna run off an instance my next go around Rent GPUsvast.aiSent from my iPhoneOn May 1, 2023, at 7:43 PM, Byron @.***> wrote: I'm on windows. I'm already pretty in over my head trying to figure out all this github stuff to also try to install linux. Has Ubuntu given you better results?

—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you were mentioned.Message ID: @.***>

[Byron-Zoomry]

I see, I'll try that then. Honestly also in the same boat. It's so stressful running into issues over and over and not being able to figure out why. I also can't even get the colab to work now