The proportion of sparse patterns during training

Hello, thank you for your wonderful work!

I would like to ask what are the proportions of the sparse patterns _(sparse points (SfM, LiDAR, Extremely sparse), low-resolution (captured, x8, x16), and missing areas (Range, Shape, Object))_ in the training?

