# Targeted Adversarial Attack on ImageNet (test) - Model Suite VGG/ResNet/Transformer Mix

Source: [Targeted Adversarial Attack on ImageNet (test) - Model Suite VGG/ResNet/Transformer Mix](<https://www.sota2.com/research/sota/targeted-adversarial-attack-on-imagenet-test-model-suite-vgg-resnet-transformer-mix>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-03-17T10:57:07.835Z

Evaluation of targeted adversarial attack effectiveness on the ImageNet test set, specifically measuring performance against a diverse suite of models including VGG16, VGG19, RN152, DN201, SE154, ConViT-B, TNT-S, Visformer-S, IncV4, and IncRes.

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 | Targeted Adversarial Attack |
| Dataset (including any stated split) | ImageNet (test) |
| Metric | Model Performance: VGG16 |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 8 of 8 |
| 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

- [Model Performance: VGG16](<https://www.sota2.com/research/sota/targeted-adversarial-attack-on-imagenet-test-model-suite-vgg-resnet-transformer-mix.md?metric=Model%20Performance%3A%20VGG16>): Higher is better; unit: SCALAR.

- [Model Performance: VGG19](<https://www.sota2.com/research/sota/targeted-adversarial-attack-on-imagenet-test-model-suite-vgg-resnet-transformer-mix.md?metric=Model%20Performance%3A%20VGG19>): Higher is better; unit: SCALAR.

- [Model Performance: RN152](<https://www.sota2.com/research/sota/targeted-adversarial-attack-on-imagenet-test-model-suite-vgg-resnet-transformer-mix.md?metric=Model%20Performance%3A%20RN152>): Higher is better; unit: SCALAR.

- [Model Performance: DN201](<https://www.sota2.com/research/sota/targeted-adversarial-attack-on-imagenet-test-model-suite-vgg-resnet-transformer-mix.md?metric=Model%20Performance%3A%20DN201>): Higher is better; unit: SCALAR.

- [Model Performance: SE154](<https://www.sota2.com/research/sota/targeted-adversarial-attack-on-imagenet-test-model-suite-vgg-resnet-transformer-mix.md?metric=Model%20Performance%3A%20SE154>): Higher is better; unit: SCALAR.

- [Model Performance: ConViT-B](<https://www.sota2.com/research/sota/targeted-adversarial-attack-on-imagenet-test-model-suite-vgg-resnet-transformer-mix.md?metric=Model%20Performance%3A%20ConViT-B>): Higher is better; unit: SCALAR.

- [Model Performance: TNT-S](<https://www.sota2.com/research/sota/targeted-adversarial-attack-on-imagenet-test-model-suite-vgg-resnet-transformer-mix.md?metric=Model%20Performance%3A%20TNT-S>): Higher is better; unit: SCALAR.

- [Model Performance: Visformer-S](<https://www.sota2.com/research/sota/targeted-adversarial-attack-on-imagenet-test-model-suite-vgg-resnet-transformer-mix.md?metric=Model%20Performance%3A%20Visformer-S>): Higher is better; unit: SCALAR.

- [Model Performance: IncV4](<https://www.sota2.com/research/sota/targeted-adversarial-attack-on-imagenet-test-model-suite-vgg-resnet-transformer-mix.md?metric=Model%20Performance%3A%20IncV4>): Higher is better; unit: SCALAR.

- [Model Performance: IncRes](<https://www.sota2.com/research/sota/targeted-adversarial-attack-on-imagenet-test-model-suite-vgg-resnet-transformer-mix.md?metric=Model%20Performance%3A%20IncRes>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Model Performance: VGG16 | Result date | Report paper |
| --- | --- | --- | --- |
| CFM+SGM | 84.52 | Not recorded | 68fa75361e209131cea66753 |
| CFM | 81.86 | Not recorded | 68fa75361e209131cea66753 |
| Logit-Margin+SGM | 71.69 | Not recorded | 68fa75361e209131cea66753 |
| Logit+SGM | 63.49 | Not recorded | 68fa75361e209131cea66753 |
| Logit-Margin | 47.86 | Not recorded | 68fa75361e209131cea66753 |
| Logit | 37.9 | Not recorded | 68fa75361e209131cea66753 |
| FFT+SGM | 16.99 | Not recorded | 68fa75361e209131cea66753 |
| FFT | 9.72 | Not recorded | 68fa75361e209131cea66753 |

### CFM+SGM

Model Performance: VGG16: 84.52. Result ID: 69b9336f47f34ce924a2b837.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75361e209131cea66753.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75361e209131cea66753.md>)

Recorded evaluation configuration: {"Source Model":"ResNet-50","Number of Iterations":"300"}. This record does not assert independent reproduction.

### CFM

Model Performance: VGG16: 81.86. Result ID: 69b9336f47f34ce924a2b836.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75361e209131cea66753.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f6e1e209131cea45835.md>)

[Implementation](<https://github.com/dreamflake/CFM>)

Recorded evaluation configuration: {"Source Model":"ResNet-50","Number of Iterations":"300"}. This record does not assert independent reproduction.

### Logit-Margin+SGM

Model Performance: VGG16: 71.69. Result ID: 69b9336f47f34ce924a2b839.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75361e209131cea66753.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75361e209131cea66753.md>)

Recorded evaluation configuration: {"Source Model":"ResNet-50","Number of Iterations":"300"}. This record does not assert independent reproduction.

### Logit+SGM

Model Performance: VGG16: 63.49. Result ID: 69b9336f47f34ce924a2b833.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75361e209131cea66753.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75361e209131cea66753.md>)

Recorded evaluation configuration: {"Source Model":"ResNet-50","Number of Iterations":"300"}. This record does not assert independent reproduction.

### Logit-Margin

Model Performance: VGG16: 47.86. Result ID: 69b9336f47f34ce924a2b838.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75361e209131cea66753.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6eb71e209131cea41bcb.md>)

Recorded evaluation configuration: {"Source Model":"ResNet-50","Number of Iterations":"300"}. This record does not assert independent reproduction.

### Logit

Model Performance: VGG16: 37.9. Result ID: 69b9336f47f34ce924a2b832.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75361e209131cea66753.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b161e209131cea2c0bb.md>)

Recorded evaluation configuration: {"Source Model":"ResNet-50","Number of Iterations":"300"}. This record does not assert independent reproduction.

### FFT+SGM

Model Performance: VGG16: 16.99. Result ID: 69b9336f47f34ce924a2b835.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75361e209131cea66753.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75361e209131cea66753.md>)

Recorded evaluation configuration: {"Source Model":"ResNet-50","Number of Iterations":"300"}. This record does not assert independent reproduction.

### FFT

Model Performance: VGG16: 9.72. Result ID: 69b9336f47f34ce924a2b834.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75361e209131cea66753.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71ce1e209131cea530e4.md>)

Recorded evaluation configuration: {"Source Model":"ResNet-50","Number of Iterations":"300"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.