Unable to load RVC models

[Chrisyk]

I am running a WSL2 conda environment on Windows and the model upload always stops at 99% and gives me this error.

[Voice Changer] RVC loading... slot: 0
EXCEPTION during loading hubert/contentvec model Model file not found: weights/hubert_base.pt
[Voice Changer] Model Load Error! Check your model is valid. 'config'

This is my weights directory

Launch command:

python3 MMVCServerSIO.py -p 18888 --https true   --content_vec_500 weights/checkpoint_best_legacy_500.pt   --hubert_base weights/hubert_base.pt   --hubert_soft weights/hubert-soft-0d54a1f4.pt   --nsf_hifigan weights/nsf_hifigan/model 

[w-okada]

I don’t know windows about conda envirnment, but, is path OK with "/" delimiter?

[Chrisyk]

Path seems to work fine using "/" when using cd. I tried changing the delimiter to \ and I got this error:

python3 MMVCServerSIO.py -p 18888 --https true   --content_vec_500 weights\checkpoint_best_legacy_500.pt   --hubert_base weights\hubert_base.pt   --hubert_soft weights\hubert-soft-0d54a1f4.pt   --nsf_hifigan weights\nsf_hifigan\model

[Voice Changer] RVC loading... slot: 0
EXCEPTION during loading hubert/contentvec model Model file not found: weightshubert_base.pt
[Voice Changer] Model Load Error! Check your model is valid. 'config'
INFO:     127.0.0.1:58310 - "POST /load_model HTTP/1.1" 200 OK
INFO:     127.0.0.1:58310 - "GET /info HTTP/1.1" 200 OK

It seems to be better but I'm not sure why it didn't work with "/"

[w-okada]

Sorry, I think "/" is better. Because, "EXCEPTION during loading hubert/contentvec model Model file not found: weightshubert_base.pt" doesn't contain correct file path "weights/hubert_base.pt".
By the way, which version you use? V.1.5.2.3a include some bugs. Please use v.1.5.2.4a.

[Chrisyk]

I found out the issue with weights was just a simple misspelling. I wrote 'weigths' by accident.
I don't get that error anymore, but it still can't load the model after updating to V.1.5.2.4a. This is the error I get now:

[Voice Changer] RVC loading... slot: 0
[Voice Changer] Prepare Model of slot: 0
[Voice Changer] Model Load Error! Check your model is valid. 'config'
INFO:     127.0.0.1:34112 - "POST /load_model HTTP/1.1" 200 OK
INFO:     127.0.0.1:34112 - "GET /info HTTP/1.1" 200 OK

I tried 3 different pt models from hugging face, but none seem to work. This is how I'm loading them:

[w-okada]

Maybe you upload wrong file. Please use the .pth file in weights folder. (not G_xxx.pth)

[Tybost]

I tried 3 different pt models from hugging face, but none seem to work. This is how I'm loading them:

Make sure you're using custom voice models trained by RVC: https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI

ie. https://booth.pm/en/items/4706454

[Chrisyk]

That seems to have worked. The model can be uploaded now, but I'm getting a separate issue when starting the program:

[Voice Changer] RVC loading... slot: 0
[Voice Changer] Prepare Model of slot: 0
gin_channels: 256 self.spk_embed_dim: 109
INFO:     127.0.0.1:53016 - "POST /load_model HTTP/1.1" 200 OK
INFO:     127.0.0.1:53016 - "GET /info HTTP/1.1" 200 OK
Generated Strengths: for prev:(1024,), for cur:(1024,)
/home/chris/anaconda3/envs/vcclient-dev/lib/python3.10/site-packages/scipy/signal/_signaltools.py:1551: UserWarning: kernel_size exceeds volume extent: the volume will be zero-padded.
  warnings.warn('kernel_size exceeds volume extent: the volume will be '
Could not load library libcudnn_cnn_infer.so.8. Error: libcuda.so: cannot open shared object file: No such file or directory

[w-okada]

Maybe you failed to install cuda.

[Chrisyk]

The issue has been fixed. libcuda.so was not on the wsl lib/cache which is why it gave me that error.

[w-okada]

Thanks for your reporting!

[BenLikeCode]

Le problème a été résolu. libcuda.so n'était pas sur la lib/cache wsl, c'est pourquoi il m'a donné cette erreur.

can you give the steps for installing this library please?

it seems that i have already the lib:

(vcclient-dev) ┌──(ben㉿BEN)-[/usr/lib/wsl/lib]
└─$ ll
total 220708
-r-xr-xr-x 1 root root 10593576 Oct 20 01:52 libcudadebugger.so.1
-r-xr-xr-x 1 root root 158304 Dec 7 00:05 libcuda.so
-r-xr-xr-x 1 root root 158304 Dec 7 00:05 libcuda.so.1
-r-xr-xr-x 1 root root 158304 Dec 7 00:05 libcuda.so.1.1
-r-xr-xr-x 1 root root 6880344 Oct 20 01:13 libd3d12core.so
-r-xr-xr-x 1 root root 801840 Oct 20 01:13 libd3d12.so
-r-xr-xr-x 1 root root 829248 Jun 1 2022 libdxcore.so
-r-xr-xr-x 1 root root 11123896 Dec 7 00:05 libnvcuvid.so
-r-xr-xr-x 1 root root 11123896 Dec 7 00:05 libnvcuvid.so.1
-r-xr-xr-x 1 root root 113395360 Dec 7 00:05 libnvdxdlkernels.so
-r-xr-xr-x 1 root root 555624 Dec 7 00:05 libnvidia-encode.so
-r-xr-xr-x 1 root root 555624 Dec 7 00:05 libnvidia-encode.so.1
-r-xr-xr-x 1 root root 234568 Oct 20 01:52 libnvidia-ml.so.1
-r-xr-xr-x 1 root root 362960 Dec 7 00:05 libnvidia-opticalflow.so
-r-xr-xr-x 1 root root 362960 Dec 7 00:05 libnvidia-opticalflow.so.1
lrwxrwxrwx 1 root root 15 Feb 11 10:58 libnvoptix_loader.so.1 -> libnvoptix.so.1
-r-xr-xr-x 1 root root 68552 Oct 20 01:52 libnvoptix.so.1
-r-xr-xr-x 1 root root 67924736 Dec 7 00:05 libnvwgf2umx.so
-r-xr-xr-x 1 root root 686416 Dec 7 00:05 nvidia-smi

but i get this message:
Could not load library libcudnn_cnn_infer.so.8. Error: libcuda.so: cannot open shared object file: No such fil e or directory