Open AlbertoForesti opened 1 year ago
Excuse me.Could you please explain how the binary classification labels were annotated during the puzzlecam_binary experiment? Are they encoded in the same way as other weakly supervised experiments, using one-hot encoding?
Hello, thanks for your precious work!
I am trying to reproduce your weakly supervised experiments, but I get an error when trying to run inference_rw.py. In particular I get: RuntimeError: INDICES element is out of DATA bounds, id=18354 axis_dim=18354 in aff_utils.py, in the edge_to_affinity function I suspect that the CAM size mismatches the size of the edge vector since I got a CAM tensor of size [1, 228, 363] from the previous scripts and a edge tensor of size [1, 128, 128] from affinitynet.
Do you have an idea for why that might happen? Kind regards