Add ability for `BenchmarkModel` to have its decoder disabled #101

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Merged

amorehead merged 2 commits into main from amorehead-decoder-disable

Mar 13, 2025

Collaborator

amorehead commented Mar 13, 2025 •

edited

Loading

Adds the ability for BenchmarkModel to have its decoder disabled
This is useful when loading pretrained model weights that are only used to extract encoder representations
For example, it's possible to inject this new decoder.disable argument as follows:

pretrained_config = OmegaConf.load(
    configs.model.encoder.pretrained.config_path
)
pretrained_config.decoder.disable = True
encoder = BenchMarkModel.load_from_checkpoint(
    configs.model.encoder.pretrained.checkpoint_path,
    strict=False,
    cfg=pretrained_config,
)

amorehead added 2 commits

March 13, 2025 11:09


          Update base.py

f18d7e8


          Update CHANGELOG.md

e536f72

amorehead merged commit 340885e into main

2 checks passed

amorehead deleted the amorehead-decoder-disable branch

March 13, 2025 16:23

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet