Function prediction
#135
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
{{ message }}
-
What is the expected way to perform function prediction?
I read the
FunctionTokenizerclass and noticed that if we provide even oneFunctionAnnotation, all other residues without a specified function are labeled with the<none>token. I also observed that the<mask>token is equal to the<pad>token, and when there are no function annotations, all residues are tokenized as<pad>, which is also treated as<mask>.What does this mean? Are we supposed to make function predictions without providing known keyword annotations?
Beta Was this translation helpful? Give feedback.
All reactions