Parcollet Titouan

Hi, please have a look at our tutorials. There is one dedicated to this.

Thanks for contributing.

@Gastron Do you think this can due to a problem in the definition of the audio pipeline ? @monaabdelazim, could you please post you .py files as well ?

Is the issue solved ?

@monaabdelazim I actually have no idea about that :-( The thing is, num_workers only helps if you have slow reading devices (HDD / SSD) it won't make any difference if...

Can you post the entirety of the YAML that doesn't work ?

To my understanding, the only problem is that lm_model is not on modules in the yaml. Modules is what asks SB to put things on GPUs.

Thanks !!! @KacperKubara !! @Gastron @30stomercury could you guys take a look / review ?

Hi ! I think you are right, I did a mistake like this once if I remember correctly. Do you already know if this has an impact? I don't feel...

Hehehe no idea! The thing being that currently, the batch size impacts decoding (this will change with the next major release). But I think that we should remain closer to...