Skip to content
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

Pavel/vision #16

Merged
merged 5 commits into from
Dec 18, 2023
Merged

Pavel/vision #16

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
5907f94
Correct syntax
srivhash Dec 17, 2023
0e340b6
Corrected syntax
srivhash Dec 17, 2023
dd560f5
fix lr scheduler
Dec 18, 2023
2611e0b
undo pyproject.toml
Dec 18, 2023
e8e0a8a
Merge branch 'main' into pavel/vision
WuTheFWasThat Dec 18, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 4 additions & 4 deletions vision/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -48,9 +48,9 @@ With the commands above we get the following results (note that the results may
| Model | Top-1 Accuracy |
|-------------------------|----------------|
| AlexNet | 56.6 |
| Dino ResNet50 | 63.7 |
| Dino ViT-B/8 | 74.9 |
| AlexNet → DINO ResNet50 | 60.7 |
| AlexNet → DINO ViT-B/8 | 64.2 |
| Dino ResNet50 | 64.5 |
| Dino ViT-B/8 | 74.0 |
| AlexNet → DINO ResNet50 | 61.9 |
| AlexNet → DINO ViT-B/8 | 66.6 |

You can add new custom models to the `models.py` and new datasets to `data.py`.
4 changes: 3 additions & 1 deletion vision/run_weak_strong.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -66,7 +66,9 @@ def train_logreg(
model = torch.nn.Linear(d, n_classes).cuda()
criterion = torch.nn.CrossEntropyLoss()
optimizer = torch.optim.Adam(model.parameters(), weight_decay=weight_decay, lr=lr)
schedule = torch.optim.lr_scheduler.CosineAnnealingLR(optimizer=optimizer, T_max=n_epochs)
n_batches = len(train_loader)
n_iter = n_batches * n_epochs
schedule = torch.optim.lr_scheduler.CosineAnnealingLR(optimizer=optimizer, T_max=n_iter)

results = {f"{key}_all": [] for key in eval_datasets.keys()}
for epoch in (pbar := tqdm.tqdm(range(n_epochs), desc="Epoch 0")):
Expand Down