Parcollet Titouan

@AsuMagic interested in fixing? :D

I am not sure to see what @popcornell is mentioning here. encode_batch is calling the forward function of TransformerASR (if properly defined in the yaml, as it is the case...

you can use whatever dataset you want with this work ;) Datasets are just Numpy arrays.

That's quite late to answer, but yes it certainly is true. The reason is that we rely on HF functions here, and back to when we wrote this code, I...

Hello @whettenr, a few major changes before we go into the details of the code. You know that SB follows a dataset-task-oriented structure. Hence you will need to refactor the...

Hi, we could use some help to think and design this feature. Important works will restart very soon on SpeechBrain as we will push a major version by the end...

@poonehmousavi @pplantinga @mravanelli @Adel-Moumen @asumagic what should we do here?

Right, it's most likely a hardware issue on my side. It still happens, dunno why.

@hbredin may want to have a look.

Hi, I don't see any reason for this type of mismatch, could you give more details on what types are involved here? Could you open a PR?