odusseys
odusseys
I'm not sure it is. https://github.com/pubkey/broadcast-channel#close-the-channel-if-you-do-not-need-it-anymore Now, on browsers which support BrodcastChannel, the lib uses the native, API. This also requires the channel to be closed to allow for garbage...
this would be extremely helpful. It is really tough right now to properly build our own datasets as we have no way of filtering for the objects we want (especially...
This is because of the scheduler used in the example code. Just remove it and the results will be correct.
unfortunately not, this leads to undefined symbol errors
You could try to train an upscaler on top but without a diffusion prior I am not optimistic that you will get anything interesting. And if you're using diffusion you...
This is still an issue. I think you might want to remove the quantizations in the provided configs so that new users (like poor me) don't train for days to...
You are going to have a much easier time using models dedicated to face embeddings. DinoV2 is quite slow and gives extremely high dimensional outputs which are not really conducive...
same issue in 2.5.0 - but it even fails when defining the trainer and kills the kernel