Normalization for AdaPoinTr finetuning on Projected ShapeNet #185
Description
What is the normalization scheme used for training AdaPoinTr on Projected ShapeNet? I notice it's not implemented in the Dataset class as it is with regular ShapeNet.
Is the dataset already normalized before training? What normalization should I use with my own dataset for fine tuning?