search

vturrisi / solo-learn

solo-learn: a library of self-supervised methods for visual representation learning powered by Pytorch Lightning
MIT License
1.41k stars 182 forks source link

Add Mocov3 #268

Closed vturrisi closed 2 years ago

codecov[bot] commented 2 years ago

Codecov Report

Merging #268 (5b79006) into main (fbbf663) will increase coverage by 0.02%. The diff coverage is 93.75%.

Flag Coverage Δ *Carryforward flag
cpu 80.90% <93.75%> (+0.53%) :arrow_up:
dali 40.66% <80.00%> (ø) Carriedforward from 6f83cef

*This pull request uses carry forward flags. Click here to find out more.

Impacted Files Coverage Δ
solo/methods/byol.py 77.64% <ø> (-8.07%) :arrow_down:
solo/utils/backbones.py 67.05% <0.00%> (ø)
solo/losses/mocov3.py 80.95% <80.95%> (ø)
solo/losses/__init__.py 100.00% <100.00%> (ø)
solo/losses/mocov2plus.py 100.00% <100.00%> (ø)
solo/methods/__init__.py 92.85% <100.00%> (+0.85%) :arrow_up:
solo/methods/base.py 78.42% <100.00%> (-0.64%) :arrow_down:
solo/methods/mocov2plus.py 100.00% <100.00%> (ø)
solo/methods/mocov3.py 100.00% <100.00%> (ø)
solo/losses/moco.py 33.33% <0.00%> (-66.67%) :arrow_down:
... and 4 more
DonkeyShot21 commented 2 years ago

Offline linear eval and cifar experiments are missing. MoCoV3 + link is missing under supported methods.

DonkeyShot21 commented 2 years ago

isnt it better to unify the losses and just pass the labels (diagonal vs first column)?