Julie-tang00/Point-BERT

Looking forward to the code of "Transformer+OcCo"

cuge1995 opened this issue · 1 comments

Looking forward to the code of "Transformer+OcCo"

Hi, we follow PoinTr to implement 'Transformer+OcCo', replacing the encoder of PoinTr with Standard Transformer, and use the training set of ShapeNet to pretrain the Transformer.