Peter Plantinga
Peter Plantinga
@ycemsubakan ? What is this "remainder" op?
This change seems reasonable to me, however, I would note that the intended way to change `device` is through the `run_opts` parameter. For example: ```python Pretrained("speechbrain/model", run_opts={"device": "cpu"}) ``` This...
You are correct: `AddBabble` only works on batches, not single examples, and therefore can't be used with `DataPipeline`. Instead, try using `AddBabble` in the `computeForward` method of a `Brain` class.
> @pplantinga is it possible to make an extension to HyperPyYAML that borrows from functools and makes partial function initialisations available to classes which inherit this function from a parent...
I agree that a different name could be more informative. Something like `check_loss_isfinite` or similar.
There's actually quite a few links on the getting started page that don't work: https://github.com/pytorch/serve/blob/master/docs/getting_started.md
Did this issue with decoding ever get resolved? I am running into a similar problem.
Did you solve this issue? I currently have the same issue.
I should have been more clear. I have a similar (but not exactly the same) issue. I am working on a different dataset, and trying to create a new recipe...
For others coming across this issue, my issue seems to be that sometimes noise symbols are ignored, and sometimes they aren't.