0.0.32

make se3 transformer for embedding template sidechain information one…

…-headed (regular non-multiheaded attention)

0.0.31

first pass for incorporating sidechain information of templates, with…

… se3 transformer

0.0.30

encode template residue sequence as well

0.0.29

make sure to color each template with its own positional encoding, in…

… preparation for sidechain information

0.0.28

cleanup, thanks to aced

0.0.26

do axial positional embedding manually

0.0.25

build in template processing

0.0.24

0.0.24

0.0.23

0.0.23

0.0.22

protect sparse attention from row-tied attention