# Adversarial Robustness Evaluation on Stanford Dogs (Image Classification) Source: [Adversarial Robustness Evaluation on Stanford Dogs (Image Classification)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:39:16.850Z Evaluation of Image Classification robustness on the Stanford Dogs dataset, reporting metrics such as Clean Accuracy, PGD-10 attack success rate, and Adversarial Accuracy (AA). 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) | Stanford Dogs | | Metric | Clean Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 18 of 18 | | 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 ASR](): Lower is better; unit: PERCENT. - [Adversarial Accuracy (AA)](): Higher is better; unit: PERCENT. ## Recorded results | Method | Clean Accuracy | Result date | Report paper | | --- | --- | --- | --- | | RoLI - Adapter | 92.02 | Not recorded | 68fa72bf1e209131cea585bb | | RanLI - Linear | 91.55 | Not recorded | 68fa72bf1e209131cea585bb | | RoLI - Full-FT | 91.55 | Not recorded | 68fa72bf1e209131cea585bb | | RoLI - Bias | 91.52 | Not recorded | 68fa72bf1e209131cea585bb | | RoLI - LoRA | 91.2 | Not recorded | 68fa72bf1e209131cea585bb | | RanLI - Adapter | 87.98 | Not recorded | 68fa72bf1e209131cea585bb | | RanLI - Bias | 87.41 | Not recorded | 68fa72bf1e209131cea585bb | | RanLI - VPT | 86.34 | Not recorded | 68fa72bf1e209131cea585bb | | RanLI - LoRA | 84.63 | Not recorded | 68fa72bf1e209131cea585bb | | TWINS-AT | 66.12 | Not recorded | 68fa6ed11e209131cea4265b | | RanLI - Full-FT | 65.85 | Not recorded | 68fa72bf1e209131cea585bb | | TWINS-AT | 64.98 | Not recorded | 68fa6ed11e209131cea4265b | | TWINS-AT | 64.98 | Not recorded | 68fa72bf1e209131cea585bb | | TWINS-TRADES | 63.96 | Not recorded | 68fa6ed11e209131cea4265b | | AutoLoRa | 62.33 | Not recorded | 68fa72bf1e209131cea585bb | | TWINS-TRADES | 60.58 | Not recorded | 68fa6ed11e209131cea4265b | | AT | 60.09 | Not recorded | 68fa6ed11e209131cea4265b | | TRADES | 59.99 | Not recorded | 68fa6ed11e209131cea4265b | ### RoLI - Adapter Clean Accuracy: 92.02. Result ID: 6992775a3a6c3148490f27be. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Swin Transformer","Initialization":"Robust Linear Initialization","Finetuning Protocol":"Adapter"}. This record does not assert independent reproduction. ### RanLI - Linear Clean Accuracy: 91.55. Result ID: 6992775a3a6c3148490f27bc. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Swin Transformer","Initialization":"Random Initialization","Finetuning Protocol":"Linear Probing"}. This record does not assert independent reproduction. ### RoLI - Full-FT Clean Accuracy: 91.55. Result ID: 6992775a3a6c3148490f27bd. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Swin Transformer","Initialization":"Robust Linear Initialization","Finetuning Protocol":"Full Finetuning"}. This record does not assert independent reproduction. ### RoLI - Bias Clean Accuracy: 91.52. Result ID: 6992775a3a6c3148490f27c0. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Swin Transformer","Initialization":"Robust Linear Initialization","Finetuning Protocol":"Bias-only"}. This record does not assert independent reproduction. ### RoLI - LoRA Clean Accuracy: 91.2. Result ID: 6992775a3a6c3148490f27bf. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Swin Transformer","Initialization":"Robust Linear Initialization","Finetuning Protocol":"LoRA"}. This record does not assert independent reproduction. ### RanLI - Adapter Clean Accuracy: 87.98. Result ID: 6992775a3a6c3148490f27b8. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Swin Transformer","Initialization":"Random Initialization","Finetuning Protocol":"Adapter"}. This record does not assert independent reproduction. ### RanLI - Bias Clean Accuracy: 87.41. Result ID: 6992775a3a6c3148490f27ba. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Swin Transformer","Initialization":"Random Initialization","Finetuning Protocol":"Bias-only"}. This record does not assert independent reproduction. ### RanLI - VPT Clean Accuracy: 86.34. Result ID: 6992775a3a6c3148490f27bb. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Swin Transformer","Initialization":"Random Initialization","Finetuning Protocol":"Visual Prompt Tuning"}. This record does not assert independent reproduction. ### RanLI - LoRA Clean Accuracy: 84.63. Result ID: 6992775a3a6c3148490f27b9. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Swin Transformer","Initialization":"Random Initialization","Finetuning Protocol":"LoRA"}. This record does not assert independent reproduction. ### TWINS-AT Clean Accuracy: 66.12. Result ID: 699060623a6c314849001b06. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"warmup":"true"}. This record does not assert independent reproduction. ### RanLI - Full-FT Clean Accuracy: 65.85. Result ID: 6992775a3a6c3148490f27b7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Swin Transformer","Initialization":"Random Initialization","Finetuning Protocol":"Full Finetuning"}. This record does not assert independent reproduction. ### TWINS-AT Clean Accuracy: 64.98. Result ID: 699060623a6c314849001b05. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"warmup":"false"}. This record does not assert independent reproduction. ### TWINS-AT Clean Accuracy: 64.98. Result ID: 6992775a3a6c3148490f27b5. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"ResNet-50"}. This record does not assert independent reproduction. ### TWINS-TRADES Clean Accuracy: 63.96. Result ID: 699060623a6c314849001b09. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"warmup":"true"}. This record does not assert independent reproduction. ### AutoLoRa Clean Accuracy: 62.33. Result ID: 6992775a3a6c3148490f27b6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"ResNet-50"}. This record does not assert independent reproduction. ### TWINS-TRADES Clean Accuracy: 60.58. Result ID: 699060623a6c314849001b08. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"warmup":"false"}. This record does not assert independent reproduction. ### AT Clean Accuracy: 60.09. Result ID: 699060623a6c314849001b04. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### TRADES Clean Accuracy: 59.99. Result ID: 699060623a6c314849001b07. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.