ellonde
ellonde
You have to use a gcp bucket path like `gs://` for saving the models and the data
I have same issue. Did you manage to resolve it?
Im having the same issue it takes a very long time to encode images. Im getting an average of 830 ms for my q5_0 model on a rather old(2019, i7)...
I would presume it works for any task where you have access to your logits
Okay cool :) So what are the checkpoints that are available now?