Ethan Harris

@justusschock that issue is for model outputs / predictions, whereas this is for the targets. Currently, all classification metrics require targets that are int or long.

Yeah, I think it could be safe to just do type inference. I would expect that targets in this case would be of the same dtype and shape as the...

One option here (at least in the docs, but import paths would still be limited) would be to use the tag filtering view we have in PL / Flash (e.g....

@SkafteNicki I guess you would break it down by metric. So you can filter all metrics by data-type / task (e.g. find all metrics that can be used for image...

Sounds good, I agree with @awaelchli that it would be nice to keep the base `Accuracy` etc. classes with e.g. the `mode` argument seems reasonable. Flash can be updated to...

Hey @krshrimali Yeah it looks like we're just missing some coverage with the `requires` decorator here: https://github.com/Lightning-AI/lightning-flash/blob/master/flash/audio/classification/input_transform.py Probably we also need to add torchvision to the audio requirements since the...

Hey @Atharva-Phatak thanks for the offer! Please feel free to take a look 😃 I think a great starting point would be to have a model in Flash (trained on...

Hey @derenv @kenohassler cool! Let me know when you have the tag and I'll upload it for review 😄

@derenv @kenohassler New version and legacy update both in review 😃 Thanks!

Hey @derenv @kenohassler Update here, the legacy release was accepted - we have a review for the new version here: https://extensions.gnome.org/review/54304 😄