I try to train model for detecting keypoints for one class with 9 keypoints. I have an errors like: train: WARNING: Ignoring corrupted image and/or label /content ...
I also have issues installing the package that way, but not with python 3, with ImfTiledMisc.