Closed Fly-Pluche closed 1 month ago
I added the flow block in the PointFlow to my model. Then, the bug happened. I tried
torch.cuda.empty_cache()
anddel recon_loss
, but not work.
Have you solved this problem? I met the same problem.
Have you visualize the number of solver steps? It could be the server steps increasing as the training go, which can happen as the points flow in more and more curve trajectories.
I added the flow block in the PointFlow to my model. Then, the bug happened. I tried
torch.cuda.empty_cache()
anddel recon_loss
, but not work.