Practical implementation of an astoundingly simple method for self-supervised learning that achieves a new state of the art (surpassing SimCLR) without contrastive learning and having to designate ...
Overview of one stage in PointMLP. Given an input point cloud, PointMLP progressively extracts local features using residual point MLP blocks. In each stage, we first transform the local point using a ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results