Adog64 b4a1aee0db Trained CIFAR-10 CNN, 60% accuracy, switching to DLA

2024-04-24 21:24:48 -04:00

..

__pycache__

Trained CIFAR-10 CNN, 60% accuracy, switching to DLA

2024-04-24 21:24:48 -04:00

data

Loop over all filters and serialize results

2024-04-23 13:48:33 -04:00

images

got data for fgsm mnist, working on displaying it

2024-04-24 12:16:09 -04:00

results

Trained CIFAR-10 CNN, 60% accuracy, switching to DLA

2024-04-24 21:24:48 -04:00

wiki

Figure saving with proper size

2024-04-18 13:51:31 -04:00

cifar10.py

Trained CIFAR-10 CNN, 60% accuracy, switching to DLA

2024-04-24 21:24:48 -04:00

cifar_net.pth

Trained CIFAR-10 CNN, 60% accuracy, switching to DLA

2024-04-24 21:24:48 -04:00

DatasetImage.png

Moved to separate project on Gitea

2024-02-22 13:39:09 -05:00

defense_filters.py

Trained CIFAR-10 CNN, 60% accuracy, switching to DLA

2024-04-24 21:24:48 -04:00

difference.png

Moved to separate project on Gitea

2024-02-22 13:39:09 -05:00

display_results.py

Converted old results to JSON format with metadata

2024-04-24 18:12:24 -04:00

dla.py

Trained CIFAR-10 CNN, 60% accuracy, switching to DLA

2024-04-24 21:24:48 -04:00

Filter_Performance_Against_FGSM_Attack_With_Snapped.png

Implemented reduced color space (snapped color) filter

2024-04-05 17:18:25 -04:00

Filter_Performance_Against_FGSM_Attack.png

Implemented reduced color space (snapped color) filter

2024-04-05 17:18:25 -04:00

kuwahara_FGSM_performance.png

Tested models with filtered and unfiltered training data

2024-03-05 13:31:22 -05:00

mnist_cnn_bilateral.pt

Models trained with various filters; kuwahara filter defense

2024-04-04 13:50:35 -04:00

mnist_cnn_filtered.pt

Tested models with filtered and unfiltered training data

2024-03-05 13:31:22 -05:00

mnist_cnn_gaussian_blur.pt

Models trained with various filters; kuwahara filter defense

2024-04-04 13:50:35 -04:00

mnist_cnn_unfiltered.pt

Tested models with filtered and unfiltered training data

2024-03-05 13:31:22 -05:00

mnist.py

Models trained with various filters; kuwahara filter defense

2024-04-04 13:50:35 -04:00

New_Snapped_Performace.png

New implementation of color snapping

2024-04-06 22:31:31 -04:00

Plurality_Vote_Accuracy_Agaisnt_Individual_Filters_0.025_epsilon_step.png

Decreased epsilon step size to 0.025

2024-04-10 12:23:55 -04:00

Plurality_Vote_Accuracy_Agaisnt_Individual_Filters.png

1-Bit, color snapping, and pluality vote accuracies

2024-04-07 12:04:46 -04:00

reformat_data.py

Converted old results to JSON format with metadata

2024-04-24 18:12:24 -04:00

test_defenses.py

Trained CIFAR-10 CNN, 60% accuracy, switching to DLA

2024-04-24 21:24:48 -04:00

unfiltered_FGSM_performance.png

Tested models with filtered and unfiltered training data

2024-03-05 13:31:22 -05:00

whale_full.png

Moved to separate project on Gitea

2024-02-22 13:39:09 -05:00

Whale-Attacked.png

Moved to separate project on Gitea

2024-02-22 13:39:09 -05:00

Whale-Original.png

Moved to separate project on Gitea

2024-02-22 13:39:09 -05:00