0.3.10

address https://github.com/lucidrains/equiformer-pytorch/issues/12

0.3.9

oops

0.3.8

make sure to mask out padding tokens in mlp attention, at the attenti…

0.3.7

make sure padding tokens are never considered when ordering by relati…

0.3.6

some cleanup

0.3.5

actually modify relative distances for prioritizing adjacent nodes

0.3.4

small cleanup

0.3.3

port over edges and adjacency matrix logic

0.3.2

0.3.2

0.3.1

zero init the attention and feedforward branch outputs by default, bu…