Jeppe Hallgren

Hey @maverick0004! Correct, currently this only uses the amino acid sequence. However, since the PSSM data it is in the ProteinNet data set it should be quick to include it...

What line are you referring to here?

Hey @TrentBrick , My assumption was that most people would actually want to define their own loss function and overwrite `def compute_loss` in `openprotein.py`. Happy to discuss what default loss...

Processing the masks in the preprocessing was purely done for efficiency reasons (faster that doing it in every training iteration). But of course if a given model needs the masked...

Hey @ecvgit , thanks for flagging this. If this is the case, it is presumably also a problem here https://github.com/OpenProtein/openprotein/blob/e4e2e0c8597f1f113b7074d0e6b223f8d019138e/openprotein.py#L84 ? I've set up some proper unit tests soon so...

For inspiration to model design, you're probably best off by reading https://www.cell.com/cell-systems/fulltext/S2405-4712(19)30076-6 In `preprocessing.py` we're currently converting to angles and back-again to ensure the distance between amino acids is exactly...