# Image Classification on CIFAR-10 1.0 (test) (Adversarial Robustness) Source: [Image Classification on CIFAR-10 1.0 (test) (Adversarial Robustness)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:39:16.001Z Evaluation of Image Classification model robustness on the CIFAR-10 1.0 test set, specifically reporting performance under PGD adversarial attacks (10 and 20 steps) and clean accuracy. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Image Classification | | Dataset (including any stated split) | CIFAR-10 1.0 (test) | | Metric | Clean Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 34 of 34 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [Clean Accuracy](): Higher is better; unit: PERCENT. - [PGD-10 Accuracy (RA)](): Higher is better; unit: PERCENT. - [PGD-20 Accuracy (RA)](): Higher is better; unit: PERCENT. - [FGSM Accuracy](): Higher is better; unit: PERCENT. - [PGD Accuracy (General)](): Higher is better; unit: PERCENT. - [CW-Inf Accuracy](): Higher is better; unit: PERCENT. - [AP Accuracy](): Higher is better; unit: PERCENT. - [DLR Score](): Higher is better; unit: PERCENT. - [AA Accuracy](): Higher is better; unit: PERCENT. ## Recorded results | Method | Clean Accuracy | Result date | Report paper | | --- | --- | --- | --- | | MART | 86.5 | Not recorded | 68fa6eb41e209131cea419ed | | TRADES | 86.3 | Not recorded | 68fa6eb41e209131cea419ed | | ADV + CAFE | 85.7 | Not recorded | 68fa6eb41e209131cea419ed | | MART + CAFE | 85.7 | Not recorded | 68fa6eb41e209131cea419ed | | BN | 84.84 | Not recorded | 68fa6b101e209131cea2be1f | | AWP + CAFE | 84.6 | Not recorded | 68fa6eb41e209131cea419ed | | ADV | 84.3 | Not recorded | 68fa6eb41e209131cea419ed | | SaAuxBN | 84.08 | Not recorded | 68fa6b101e209131cea2be1f | | HELP | 83.8 | Not recorded | 68fa6eb41e209131cea419ed | | TRADES + CAFE | 83.7 | Not recorded | 68fa6eb41e209131cea419ed | | AWP | 83.7 | Not recorded | 68fa6eb41e209131cea419ed | | MART | 83.5 | Not recorded | 68fa6eb41e209131cea419ed | | AuxBN | 83.42 | Not recorded | 68fa6b101e209131cea2be1f | | ModeNorm | 83.28 | Not recorded | 68fa6b101e209131cea2be1f | | HELP + CAFE | 83.1 | Not recorded | 68fa6eb41e209131cea419ed | | TRADES | 83 | Not recorded | 68fa6eb41e209131cea419ed | | ADV + CAFE | 82.6 | Not recorded | 68fa6eb41e209131cea419ed | | MART + CAFE | 82.1 | Not recorded | 68fa6eb41e209131cea419ed | | ADV | 82 | Not recorded | 68fa6eb41e209131cea419ed | | AWP + CAFE | 81.5 | Not recorded | 68fa6eb41e209131cea419ed | | AWP | 81.2 | Not recorded | 68fa6eb41e209131cea419ed | | TRADES + CAFE | 80.7 | Not recorded | 68fa6eb41e209131cea419ed | | HELP + CAFE | 80.6 | Not recorded | 68fa6eb41e209131cea419ed | | HELP | 80.5 | Not recorded | 68fa6eb41e209131cea419ed | | MART | 79.7 | Not recorded | 68fa6eb41e209131cea419ed | | TRADES | 79.5 | Not recorded | 68fa6eb41e209131cea419ed | | ADV | 78.5 | Not recorded | 68fa6eb41e209131cea419ed | | ADV + CAFE | 78.4 | Not recorded | 68fa6eb41e209131cea419ed | | MART + CAFE | 78.3 | Not recorded | 68fa6eb41e209131cea419ed | | AWP | 78 | Not recorded | 68fa6eb41e209131cea419ed | | AWP + CAFE | 77.4 | Not recorded | 68fa6eb41e209131cea419ed | | HELP | 77.4 | Not recorded | 68fa6eb41e209131cea419ed | | TRADES + CAFE | 77 | Not recorded | 68fa6eb41e209131cea419ed | | HELP + CAFE | 75.6 | Not recorded | 68fa6eb41e209131cea419ed | ### MART Clean Accuracy: 86.5. Result ID: 699053c63a6c314849ffc0ca. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"WideResNet-34-10"}. This record does not assert independent reproduction. ### TRADES Clean Accuracy: 86.3. Result ID: 699053c63a6c314849ffc0c8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"WideResNet-34-10"}. This record does not assert independent reproduction. ### ADV + CAFE Clean Accuracy: 85.7. Result ID: 699053c63a6c314849ffc0c7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"WideResNet-34-10"}. This record does not assert independent reproduction. ### MART + CAFE Clean Accuracy: 85.7. Result ID: 699053c63a6c314849ffc0cb. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"WideResNet-34-10"}. This record does not assert independent reproduction. ### BN Clean Accuracy: 84.84. Result ID: 698e4fdbffc3901de09c882d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"ResNet-18"}. This record does not assert independent reproduction. ### AWP + CAFE Clean Accuracy: 84.6. Result ID: 699053c63a6c314849ffc0cd. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"WideResNet-34-10"}. This record does not assert independent reproduction. ### ADV Clean Accuracy: 84.3. Result ID: 699053c63a6c314849ffc0c6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"WideResNet-34-10"}. This record does not assert independent reproduction. ### SaAuxBN Clean Accuracy: 84.08. Result ID: 698e4fdbffc3901de09c8830. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"ResNet-18","Evaluation Branch":"adversarial branch"}. This record does not assert independent reproduction. ### HELP Clean Accuracy: 83.8. Result ID: 699053c63a6c314849ffc0ce. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"WideResNet-34-10"}. This record does not assert independent reproduction. ### TRADES + CAFE Clean Accuracy: 83.7. Result ID: 699053c63a6c314849ffc0c9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"WideResNet-34-10"}. This record does not assert independent reproduction. ### AWP Clean Accuracy: 83.7. Result ID: 699053c63a6c314849ffc0cc. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"WideResNet-34-10"}. This record does not assert independent reproduction. ### MART Clean Accuracy: 83.5. Result ID: 699053c63a6c314849ffc0c0. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"ResNet-18"}. This record does not assert independent reproduction. ### AuxBN Clean Accuracy: 83.42. Result ID: 698e4fdbffc3901de09c882f. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"ResNet-18","Evaluation Branch":"adversarial branch"}. This record does not assert independent reproduction. ### ModeNorm Clean Accuracy: 83.28. Result ID: 698e4fdbffc3901de09c882e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"ResNet-18"}. This record does not assert independent reproduction. ### HELP + CAFE Clean Accuracy: 83.1. Result ID: 699053c63a6c314849ffc0cf. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"WideResNet-34-10"}. This record does not assert independent reproduction. ### TRADES Clean Accuracy: 83. Result ID: 699053c63a6c314849ffc0be. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"ResNet-18"}. This record does not assert independent reproduction. ### ADV + CAFE Clean Accuracy: 82.6. Result ID: 699053c63a6c314849ffc0bd. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"ResNet-18"}. This record does not assert independent reproduction. ### MART + CAFE Clean Accuracy: 82.1. Result ID: 699053c63a6c314849ffc0c1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"ResNet-18"}. This record does not assert independent reproduction. ### ADV Clean Accuracy: 82. Result ID: 699053c63a6c314849ffc0bc. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"ResNet-18"}. This record does not assert independent reproduction. ### AWP + CAFE Clean Accuracy: 81.5. Result ID: 699053c63a6c314849ffc0c3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"ResNet-18"}. This record does not assert independent reproduction. ### AWP Clean Accuracy: 81.2. Result ID: 699053c63a6c314849ffc0c2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"ResNet-18"}. This record does not assert independent reproduction. ### TRADES + CAFE Clean Accuracy: 80.7. Result ID: 699053c63a6c314849ffc0bf. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"ResNet-18"}. This record does not assert independent reproduction. ### HELP + CAFE Clean Accuracy: 80.6. Result ID: 699053c63a6c314849ffc0c5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"ResNet-18"}. This record does not assert independent reproduction. ### HELP Clean Accuracy: 80.5. Result ID: 699053c63a6c314849ffc0c4. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"ResNet-18"}. This record does not assert independent reproduction. ### MART Clean Accuracy: 79.7. Result ID: 699053c63a6c314849ffc0b6. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"VGG-16"}. This record does not assert independent reproduction. ### TRADES Clean Accuracy: 79.5. Result ID: 699053c63a6c314849ffc0b4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"VGG-16"}. This record does not assert independent reproduction. ### ADV Clean Accuracy: 78.5. Result ID: 699053c63a6c314849ffc0b2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"VGG-16"}. This record does not assert independent reproduction. ### ADV + CAFE Clean Accuracy: 78.4. Result ID: 699053c63a6c314849ffc0b3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"VGG-16"}. This record does not assert independent reproduction. ### MART + CAFE Clean Accuracy: 78.3. Result ID: 699053c63a6c314849ffc0b7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"VGG-16"}. This record does not assert independent reproduction. ### AWP Clean Accuracy: 78. Result ID: 699053c63a6c314849ffc0b8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"VGG-16"}. This record does not assert independent reproduction. ### AWP + CAFE Clean Accuracy: 77.4. Result ID: 699053c63a6c314849ffc0b9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"VGG-16"}. This record does not assert independent reproduction. ### HELP Clean Accuracy: 77.4. Result ID: 699053c63a6c314849ffc0ba. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"VGG-16"}. This record does not assert independent reproduction. ### TRADES + CAFE Clean Accuracy: 77. Result ID: 699053c63a6c314849ffc0b5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"VGG-16"}. This record does not assert independent reproduction. ### HELP + CAFE Clean Accuracy: 75.6. Result ID: 699053c63a6c314849ffc0bb. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"VGG-16"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.