0.1.5

vector attention for multi-head attention variant

0.1.4

fix nearest neighbors w/ multihead

0.1.2

fix multihead point transformer layer

0.1.1

add multi-head point transformer layer

0.1.0

__

0.0.3

add ability to attend to k-nearest neighbors, to cutdown on attention…

0.0.2

Merge pull request #3 from lucidrains/modulate-feature-dimension

different attention map for each element of feature dimension

0.0.1

Create python-publish.yml