How common is it for Cloud Lambda to be out of capacity on everything?
It’s been like this for a while, my guess is that it’s due to the coming deadline of NeurIPS 2023 (May 17). Previously, it’s because ICCV 2023 deadline (March 8). Everyone’s holding onto the machines in case they need to run some extra experiments before submission.
I hope the situation gets better after May 17, we’ll see.