TRACE Training Shape Mismatch

[dylan6thomas]

Hi! I am trying to train trace using just the DynaCam dataset to see if I can get it working before I continue my project. However, I am getting a shape mismatch on some files, though I am not sure why . Any help would be appreciated. I have attached the error and the config file.

I believe there is a disagreement between the world_cam_mask and thetorch.from_numpy(normalize_trans_to_cam_params(trans_world)).float() dimensions, but I couldn't track down what was causing the mismatch.

Thank you for your time.

[jayantsharma]

Facing the exact same issue, when trying to train on only the DynaCam dataset.