Skip to content

Fix Quick Tour Error #216

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

Open
Alexander230 wants to merge 1 commit into
base: master
Choose a base branch
from
Open

Fix Quick Tour Error #216

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
15 changes: 9 additions & 6 deletions README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -103,13 +103,16 @@ Now let's try to load CIFAR-10 and some quite robust CIFAR-10 models from
`eps=8/255`:

```python
import torch
from robustbench.data import load_cifar10

x_test, y_test = load_cifar10(n_examples=50)

from robustbench.utils import load_model

model = load_model(model_name='Carmon2019Unlabeled', dataset='cifar10', threat_model='Linf')
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")

model = load_model(model_name='Carmon2019Unlabeled', dataset='cifar10', threat_model='Linf').to(device)
```

Let's try to evaluate the robustness of this model. We can use any favourite library for this. For example, [FoolBox](https://github.com/bethgelab/foolbox)
Expand All @@ -119,7 +122,7 @@ implements many different attacks. We can start from a simple PGD attack:
import foolbox as fb
fmodel = fb.PyTorchModel(model, bounds=(0, 1))

_, advs, success = fb.attacks.LinfPGD()(fmodel, x_test.to('cuda:0'), y_test.to('cuda:0'), epsilons=[8/255])
_, advs, success = fb.attacks.LinfPGD()(fmodel, x_test.to(device), y_test.to(device), epsilons=[8/255])
print('Robust accuracy: {:.1%}'.format(1 - success.float().mean()))
```
```
Expand All @@ -131,7 +134,7 @@ Let's try to evaluate its robustness with a cheap version [AutoAttack](https://a
```python
# autoattack is installed as a dependency of robustbench so there is not need to install it separately
from autoattack import AutoAttack
adversary = AutoAttack(model, norm='Linf', eps=8/255, version='custom', attacks_to_run=['apgd-ce', 'apgd-dlr'])
adversary = AutoAttack(model, norm='Linf', eps=8/255, version='custom', attacks_to_run=['apgd-ce', 'apgd-dlr'], device=device)
adversary.apgd.n_restarts = 1
x_adv = adversary.run_standard_evaluation(x_test, y_test)
```
Expand All @@ -145,7 +148,7 @@ x_adv = adversary.run_standard_evaluation(x_test, y_test)
>>> robust accuracy: 52.00%
```
Note that for our standardized evaluation of Linf-robustness we use the *full* version of AutoAttack which is slower but
more accurate (for that just use `adversary = AutoAttack(model, norm='Linf', eps=8/255)`).
more accurate (for that just use `adversary = AutoAttack(model, norm='Linf', eps=8/255, device=device)`).

What about other types of perturbations? Is Lp-robustness useful there? We can evaluate the available models on more general perturbations.
For example, let's take images corrupted by fog perturbations from CIFAR-10-C with the highest level of severity (5).
Expand All @@ -160,8 +163,8 @@ x_test, y_test = load_cifar10c(n_examples=1000, corruptions=corruptions, severit

for model_name in ['Standard', 'Engstrom2019Robustness', 'Rice2020Overfitting',
'Carmon2019Unlabeled']:
model = load_model(model_name, dataset='cifar10', threat_model='Linf')
acc = clean_accuracy(model, x_test, y_test)
model = load_model(model_name, dataset='cifar10', threat_model='Linf').to(device)
acc = clean_accuracy(model, x_test, y_test, device=device)
print(f'Model: {model_name}, CIFAR-10-C accuracy: {acc:.1%}')
```
```
Expand Down