Skip to content
This repository has been archived by the owner on Oct 27, 2022. It is now read-only.

adding pfam samples sweeps and images for pfam and gfp experiments #7

Open
wants to merge 8 commits into
base: master
Choose a base branch
from

Conversation

amirshane
Copy link
Contributor

No description provided.

contextual_lenses/load_transformer.py Outdated Show resolved Hide resolved
contextual_lenses/train_utils.py Outdated Show resolved Hide resolved
from protein_lm import data, models


def load_transformer_model(ckpt_dir, model_cls, domain=None):
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

this function should definitely have test coverage. The easiest way to do this would be to make a new small model, compute the output on some dummy inputs, save the model to disk, make a new model by loading from file, and check that the output on the dummy inputs is the same.

contextual_lenses/tape_gfp/tape_gfp_utils.py Show resolved Hide resolved

optimizer_def = optim.Adam(learning_rate=learning_rate, weight_decay=weight_decay)
optimizer = optimizer_def.create(model)
def create_optimizer(model, learning_rate, weight_decay, layers=None):
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Lots of flax details here. make sure ddohan signs off on this entire file. Also need to have some loading/saving tests.
@dmrd

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

3 participants