lucidrains / alphafold2

To eventually become an unofficial Pytorch implementation / replication of Alphafold2, as details of the architecture get released

[todo] BERT loss on MSA

lucidrains opened this issue 3 years ago · comments

Phil Wang commented 3 years ago

needs to be added

Phil Wang commented 3 years ago

need to make noise accurate to the paper (contiguous spans?)

Phil Wang commented 3 years ago

Phil Wang commented 3 years ago

used at inference time.. 🤔

Siqi commented 3 years ago

used at inference time.. 🤔

Yep, I also noticed this part, which seems really weird to me...