libsndfile Error

[gusanmaz]

libsndfile Error

[ChrisShaoMeng]

+1 Same error.

[cbalioglu]

@gusanmaz if you have been able to install libsndfile via your system package manager, could you please try to install seamless_communication in a regular Python virtual environment (i.e. no Conda)?

[ChrisShaoMeng]

@gusanmaz if you have been able to install libsndfile via your system package manager, could you please try to install seamless_communication in a regular Python virtual environment (i.e. no Conda)?

It works for me, thanks!

[cgalisteo-stratio]

@gusanmaz if you have been able to install libsndfile via your system package manager, could you please try to install seamless_communication in a regular Python virtual environment (i.e. no Conda)?

That was the solution for me. I was getting that error using a conda environment, but now it's working with a regular virtualenv.

[gusanmaz]

I solved the problem. It now works in 12GB vram machine with big model and 4GB vram machine with medium model.

[cbalioglu]

@gusanmaz glad to hear! Closing this issue. Let us know if you experience another problem.

[gusanmaz]

Thank you for the advice @cbalioglu.

This time I am getting ERROR: Could not find a version that satisfies the requirement fairseq2n==0.1.0 (from fairseq2) (from versions: none) ERROR: No matching distribution found for fairseq2n==0.1.0 message during installation. It appears like same issue has already been reported under fairseq2 issues: facebookresearch/fairseq2#1 .

I am able to install and run seamless_communication without using conda environment on my another less powerful machine with same OS with no libsndfile error. I am getting CUDA out of memory error. This machine has only 4GB vram whereas the other one has 12GB vram.

[shashank-neuralgarage]

I solved the problem. It now works in 12GB vram machine with big model and 4GB vram machine with medium model.

Can you share what worked for you?

[milo157]

I solved the problem. It now works in 12GB vram machine with big model and 4GB vram machine with medium model.

Can you share what worked for you?

Funny enough making sure I did a conda install of it. You can see my code here deploying it on Cerebrium: https://github.com/CerebriumAI/examples/tree/master/19-meta-seamless