what a great work! I have a question about dataset.

On your paper, you used Instavariety, posetrack, pennaction, human3.6m, 3dpw, and 3dhp as training dataset.

but in config.yaml file, you only used insta, h36m, 3dpw.

which one is correct? and I want to know about what dataset was used when learning the best_model.pt file.