# Fine grained classification on Oxford-IIIT Pets

Source: [Fine grained classification on Oxford-IIIT Pets](<https://www.sota2.com/research/sota/fine-grained-classification-on-oxford-iiit-pets>)

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

Data updated: 2026-05-13T11:27:49.968Z

Evaluation of Fine Grained Classification performance on the Oxford-IIIT Pets dataset.

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 | Fine grained classification |
| Dataset (including any stated split) | Oxford-IIIT Pets |
| Metric | Top-1 Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 37 of 37 |
| 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

- [Top-1 Accuracy](<https://www.sota2.com/research/sota/fine-grained-classification-on-oxford-iiit-pets.md?metric=Top-1%20Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Top-1 Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| OmniVec2 | 99.6 | Not recorded | 68fa7a181e209131cea811ca |
| OmniVec | 99.2 | Not recorded | 68fa7a181e209131cea811ca |
| DINOv3-B | 96.54 | Not recorded | 6a0405eadbc836b87d8279bb |
| DINOv2-Reg-B | 95.97 | Not recorded | 6a0405eadbc836b87d8279bb |
| AmoebaNet-B | 95.9 | Not recorded | 68fa67391e209131cea15048 |
| VECA-B | 95.67 | Not recorded | 6a0405eadbc836b87d8279bb |
| DINOv2-B | 95.61 | Not recorded | 6a0405eadbc836b87d8279bb |
| VECA-B | 95.12 | Not recorded | 6a0405eadbc836b87d8279bb |
| Assemble-ResNet-FGVC-50 | 94.3 | Not recorded | 68fa67391e209131cea15048 |
| ResNet-50 (ReLabel-trained) | 94.28 | Not recorded | 68fa6a751e209131cea2842b |
| ResNet-50 (Baseline) | 93.92 | Not recorded | 68fa6a751e209131cea2842b |
| DFN2B-CLIP-B/16 | 93.59 | Not recorded | 6a0405eadbc836b87d8279bb |
| C-RADIOv3-B | 93.54 | Not recorded | 6a0405eadbc836b87d8279bb |
| SigLIP2-B/16-256 | 92.89 | Not recorded | 6a0405eadbc836b87d8279bb |
| CLIP-B/16 | 92.72 | Not recorded | 6a0405eadbc836b87d8279bb |
| ResNet-50 | 92.5 | Not recorded | 68fa67391e209131cea15048 |
| OpenCLIP-B/16 | 91.91 | Not recorded | 6a0405eadbc836b87d8279bb |
| CoCoOp | 90.5 | Not recorded | 68fa7b941e209131cea89411 |
| TransCLIP | 89.3 | Not recorded | 68fa7c701e209131cea8e17f |
| CoOp | 89.1 | Not recorded | 68fa7b941e209131cea89411 |
| CLIP-VIT-B/16 | 88.3 | Not recorded | 68fa7b941e209131cea89411 |
| ADAPT | 87.95 | Not recorded | 68fa7c701e209131cea8e17f |
| TPT | 87.8 | Not recorded | 68fa7b941e209131cea89411 |
| StatA | 87.7 | Not recorded | 68fa7c701e209131cea8e17f |
| Dota | 87.33 | Not recorded | 68fa7c701e209131cea8e17f |
| Ensemble | 86.9 | Not recorded | 68fa7b941e209131cea89411 |
| DMN | 86.78 | Not recorded | 68fa7c701e209131cea8e17f |
| ADAPT | 86.43 | Not recorded | 68fa7c701e209131cea8e17f |
| TDA | 86.18 | Not recorded | 68fa7c701e209131cea8e17f |
| DPE | 85.97 | Not recorded | 68fa7c701e209131cea8e17f |
| BCA | 85.58 | Not recorded | 68fa7c701e209131cea8e17f |
| TPT | 84.49 | Not recorded | 68fa7c701e209131cea8e17f |
| TPS | 84.4 | Not recorded | 68fa7c701e209131cea8e17f |
| C-TPT | 84.1 | Not recorded | 68fa7c701e209131cea8e17f |
| CLIP | 83.57 | Not recorded | 68fa7c701e209131cea8e17f |
| DiffTPT | 83.4 | Not recorded | 68fa7c701e209131cea8e17f |
| APM | 81.6 | Not recorded | 68fa7b941e209131cea89411 |

### OmniVec2

Top-1 Accuracy: 99.6. Result ID: 698b6d317a967eddb76e784a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a181e209131cea811ca.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a181e209131cea811ca.md>)

Recorded evaluation configuration: {"Modality":"Image"}. This record does not assert independent reproduction.

### OmniVec

Top-1 Accuracy: 99.2. Result ID: 698b6d317a967eddb76e784b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a181e209131cea811ca.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71201e209131cea4f58e.md>)

Recorded evaluation configuration: {"Modality":"Image"}. This record does not assert independent reproduction.

### DINOv3-B

Top-1 Accuracy: 96.54. Result ID: 6a04601847f34ce924eb3dea.

[Reporting paper](<https://www.sota2.com/research/paper/6a0405eadbc836b87d8279bb.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a931e209131cea83fc5.md>)

Recorded evaluation configuration: {"Evaluation protocol":"frozen-backbone linear probing","Crop size":"256 x 256"}. This record does not assert independent reproduction.

### DINOv2-Reg-B

Top-1 Accuracy: 95.97. Result ID: 6a04601847f34ce924eb3deb.

[Reporting paper](<https://www.sota2.com/research/paper/6a0405eadbc836b87d8279bb.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72e01e209131cea5919d.md>)

Recorded evaluation configuration: {"Evaluation protocol":"frozen-backbone linear probing","Crop size":"224 x 224"}. This record does not assert independent reproduction.

### AmoebaNet-B

Top-1 Accuracy: 95.9. Result ID: 698d22dc8b517eb4331f4685.

[Reporting paper](<https://www.sota2.com/research/paper/68fa67391e209131cea15048.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66121e209131cea0e102.md>)

Recorded evaluation configuration: {"Model":"AmoebaNet-B"}. This record does not assert independent reproduction.

### VECA-B

Top-1 Accuracy: 95.67. Result ID: 6a04601847f34ce924eb3df2.

[Reporting paper](<https://www.sota2.com/research/paper/6a0405eadbc836b87d8279bb.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a0405eadbc836b87d8279bb.md>)

Recorded evaluation configuration: {"Evaluation protocol":"frozen-backbone linear probing","Crop size":"256 x 256","Core budget (C)":"64"}. This record does not assert independent reproduction.

### DINOv2-B

Top-1 Accuracy: 95.61. Result ID: 6a04601847f34ce924eb3dec.

[Reporting paper](<https://www.sota2.com/research/paper/6a0405eadbc836b87d8279bb.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72181e209131cea546c3.md>)

[Implementation](<https://github.com/facebookresearch/dinov2>)

Recorded evaluation configuration: {"Evaluation protocol":"frozen-backbone linear probing","Crop size":"224 x 224"}. This record does not assert independent reproduction.

### VECA-B

Top-1 Accuracy: 95.12. Result ID: 6a04601847f34ce924eb3df3.

[Reporting paper](<https://www.sota2.com/research/paper/6a0405eadbc836b87d8279bb.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a0405eadbc836b87d8279bb.md>)

Recorded evaluation configuration: {"Evaluation protocol":"frozen-backbone linear probing","Crop size":"256 x 256","Core budget (C)":"8"}. This record does not assert independent reproduction.

### Assemble-ResNet-FGVC-50

Top-1 Accuracy: 94.3. Result ID: 698d22dc8b517eb4331f4687.

[Reporting paper](<https://www.sota2.com/research/paper/68fa67391e209131cea15048.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67391e209131cea15048.md>)

[Implementation](<https://github.com/clovaai/assembled-cnn>)

Recorded evaluation configuration: {"Model":"Assemble-ResNet-FGVC-50"}. This record does not assert independent reproduction.

### ResNet-50 (ReLabel-trained)

Top-1 Accuracy: 94.28. Result ID: 698e3724ffc3901de09bba24.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6a751e209131cea2842b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a751e209131cea2842b.md>)

[Implementation](<https://github.com/naver-ai/relabel_imagenet>)

Recorded evaluation configuration: {"Backbone":"ResNet-50","Pre-training strategy":"ReLabel"}. This record does not assert independent reproduction.

### ResNet-50 (Baseline)

Top-1 Accuracy: 93.92. Result ID: 698e3724ffc3901de09bba23.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6a751e209131cea2842b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa62321e209131ce9f8328.md>)

Recorded evaluation configuration: {"Backbone":"ResNet-50","Pre-training strategy":"Standard"}. This record does not assert independent reproduction.

### DFN2B-CLIP-B/16

Top-1 Accuracy: 93.59. Result ID: 6a04601847f34ce924eb3dee.

[Reporting paper](<https://www.sota2.com/research/paper/6a0405eadbc836b87d8279bb.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71221e209131cea4f63c.md>)

Recorded evaluation configuration: {"Evaluation protocol":"frozen-backbone linear probing","Crop size":"224 x 224"}. This record does not assert independent reproduction.

### C-RADIOv3-B

Top-1 Accuracy: 93.54. Result ID: 6a04601847f34ce924eb3def.

[Reporting paper](<https://www.sota2.com/research/paper/6a0405eadbc836b87d8279bb.md>)

Recorded evaluation configuration: {"Evaluation protocol":"frozen-backbone linear probing","Crop size":"256 x 256"}. This record does not assert independent reproduction.

### SigLIP2-B/16-256

Top-1 Accuracy: 92.89. Result ID: 6a04601847f34ce924eb3ded.

[Reporting paper](<https://www.sota2.com/research/paper/6a0405eadbc836b87d8279bb.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa770a1e209131cea70ad4.md>)

Recorded evaluation configuration: {"Evaluation protocol":"frozen-backbone linear probing","Crop size":"256 x 256"}. This record does not assert independent reproduction.

### CLIP-B/16

Top-1 Accuracy: 92.72. Result ID: 6a04601847f34ce924eb3df0.

[Reporting paper](<https://www.sota2.com/research/paper/6a0405eadbc836b87d8279bb.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69581e209131cea21ad0.md>)

[Implementation](<https://github.com/OpenAI/CLIP>)

Recorded evaluation configuration: {"Evaluation protocol":"frozen-backbone linear probing","Crop size":"224 x 224"}. This record does not assert independent reproduction.

### ResNet-50

Top-1 Accuracy: 92.5. Result ID: 698d22dc8b517eb4331f4686.

[Reporting paper](<https://www.sota2.com/research/paper/68fa67391e209131cea15048.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa62321e209131ce9f8328.md>)

Recorded evaluation configuration: {"Model":"ResNet-50"}. This record does not assert independent reproduction.

### OpenCLIP-B/16

Top-1 Accuracy: 91.91. Result ID: 6a04601847f34ce924eb3df1.

[Reporting paper](<https://www.sota2.com/research/paper/6a0405eadbc836b87d8279bb.md>)

Recorded evaluation configuration: {"Evaluation protocol":"frozen-backbone linear probing","Crop size":"224 x 224"}. This record does not assert independent reproduction.

### CoCoOp

Top-1 Accuracy: 90.5. Result ID: 698c5a7b37894ef7d9cdc709.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b941e209131cea89411.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d8c1e209131cea3a831.md>)

[Implementation](<https://github.com/KaiyangZhou/CoOp>)

Recorded evaluation configuration: {"Pre-trained on ImageNet":true}. This record does not assert independent reproduction.

### TransCLIP

Top-1 Accuracy: 89.3. Result ID: 69ba574147f34ce924a3e518.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c701e209131cea8e17f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73791e209131cea5cc32.md>)

[Implementation](<https://github.com/MaxZanella/transduction-for-vlms>)

Recorded evaluation configuration: {"Backbone":"CLIP-RN50","BP-free":"true","Evaluation Protocol":"Transductive"}. This record does not assert independent reproduction.

### CoOp

Top-1 Accuracy: 89.1. Result ID: 698c5a7b37894ef7d9cdc708.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b941e209131cea89411.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d8b1e209131cea3a814.md>)

[Implementation](<https://github.com/KaiyangZhou/CoOp>)

Recorded evaluation configuration: {"Pre-trained on ImageNet":true}. This record does not assert independent reproduction.

### CLIP-VIT-B/16

Top-1 Accuracy: 88.3. Result ID: 698c5a7b37894ef7d9cdc70a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b941e209131cea89411.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69581e209131cea21ad0.md>)

[Implementation](<https://github.com/OpenAI/CLIP>)

Recorded evaluation configuration: {"Pre-trained on ImageNet":false}. This record does not assert independent reproduction.

### ADAPT

Top-1 Accuracy: 87.95. Result ID: 69ba574147f34ce924a3e51a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c701e209131cea8e17f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c701e209131cea8e17f.md>)

Recorded evaluation configuration: {"Backbone":"CLIP-RN50","BP-free":"true","Evaluation Protocol":"Transductive"}. This record does not assert independent reproduction.

### TPT

Top-1 Accuracy: 87.8. Result ID: 698c5a7b37894ef7d9cdc70c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b941e209131cea89411.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d691e209131cea399f9.md>)

[Implementation](<https://tinyurl.com/yr3zmhma>)

Recorded evaluation configuration: {"Pre-trained on ImageNet":false}. This record does not assert independent reproduction.

### StatA

Top-1 Accuracy: 87.7. Result ID: 69ba574147f34ce924a3e519.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c701e209131cea8e17f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76791e209131cea6d3c6.md>)

[Implementation](<https://github.com/MaxZanella/StatA>)

Recorded evaluation configuration: {"Backbone":"CLIP-RN50","BP-free":"true","Evaluation Protocol":"Transductive"}. This record does not assert independent reproduction.

### Dota

Top-1 Accuracy: 87.33. Result ID: 69ba574147f34ce924a3e516.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c701e209131cea8e17f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7ba91e209131cea89b88.md>)

Recorded evaluation configuration: {"Backbone":"CLIP-RN50","BP-free":"true","Evaluation Protocol":"Online"}. This record does not assert independent reproduction.

### Ensemble

Top-1 Accuracy: 86.9. Result ID: 698c5a7b37894ef7d9cdc70b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b941e209131cea89411.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69581e209131cea21ad0.md>)

[Implementation](<https://github.com/OpenAI/CLIP>)

Recorded evaluation configuration: {"Pre-trained on ImageNet":false}. This record does not assert independent reproduction.

### DMN

Top-1 Accuracy: 86.78. Result ID: 69ba574147f34ce924a3e511.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c701e209131cea8e17f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72b31e209131cea58047.md>)

[Implementation](<https://github.com/YBZh/DMN>)

Recorded evaluation configuration: {"Backbone":"CLIP-RN50","BP-free":"false","Evaluation Protocol":"Online"}. This record does not assert independent reproduction.

### ADAPT

Top-1 Accuracy: 86.43. Result ID: 69ba574147f34ce924a3e517.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c701e209131cea8e17f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c701e209131cea8e17f.md>)

Recorded evaluation configuration: {"Backbone":"CLIP-RN50","BP-free":"true","Evaluation Protocol":"Online"}. This record does not assert independent reproduction.

### TDA

Top-1 Accuracy: 86.18. Result ID: 69ba574147f34ce924a3e515.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c701e209131cea8e17f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72b61e209131cea581b5.md>)

Recorded evaluation configuration: {"Backbone":"CLIP-RN50","BP-free":"true","Evaluation Protocol":"Online"}. This record does not assert independent reproduction.

### DPE

Top-1 Accuracy: 85.97. Result ID: 69ba574147f34ce924a3e513.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c701e209131cea8e17f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa753f1e209131cea66cea.md>)

[Implementation](<https://github.com/zhangce01/DPE-CLIP>)

Recorded evaluation configuration: {"Backbone":"CLIP-RN50","BP-free":"false","Evaluation Protocol":"Online"}. This record does not assert independent reproduction.

### BCA

Top-1 Accuracy: 85.58. Result ID: 69ba574147f34ce924a3e514.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c701e209131cea8e17f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa777d1e209131cea738ba.md>)

Recorded evaluation configuration: {"Backbone":"CLIP-RN50","BP-free":"true","Evaluation Protocol":"Online"}. This record does not assert independent reproduction.

### TPT

Top-1 Accuracy: 84.49. Result ID: 69ba574147f34ce924a3e50e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c701e209131cea8e17f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d691e209131cea399f9.md>)

[Implementation](<https://tinyurl.com/yr3zmhma>)

Recorded evaluation configuration: {"Backbone":"CLIP-RN50","BP-free":"false","Evaluation Protocol":"Online"}. This record does not assert independent reproduction.

### TPS

Top-1 Accuracy: 84.4. Result ID: 69ba574147f34ce924a3e512.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c701e209131cea8e17f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa762e1e209131cea6bb46.md>)

Recorded evaluation configuration: {"Backbone":"CLIP-RN50","BP-free":"false","Evaluation Protocol":"Online"}. This record does not assert independent reproduction.

### C-TPT

Top-1 Accuracy: 84.1. Result ID: 69ba574147f34ce924a3e510.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c701e209131cea8e17f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72bc1e209131cea58415.md>)

Recorded evaluation configuration: {"Backbone":"CLIP-RN50","BP-free":"false","Evaluation Protocol":"Online"}. This record does not assert independent reproduction.

### CLIP

Top-1 Accuracy: 83.57. Result ID: 69ba574147f34ce924a3e50d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c701e209131cea8e17f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69581e209131cea21ad0.md>)

[Implementation](<https://github.com/OpenAI/CLIP>)

Recorded evaluation configuration: {"Backbone":"CLIP-RN50","BP-free":"N/A","Evaluation Protocol":"Online"}. This record does not assert independent reproduction.

### DiffTPT

Top-1 Accuracy: 83.4. Result ID: 69ba574147f34ce924a3e50f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c701e209131cea8e17f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70551e209131cea4a4eb.md>)

Recorded evaluation configuration: {"Backbone":"CLIP-RN50","BP-free":"false","Evaluation Protocol":"Online"}. This record does not assert independent reproduction.

### APM

Top-1 Accuracy: 81.6. Result ID: 698c5a7b37894ef7d9cdc70d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b941e209131cea89411.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7b941e209131cea89411.md>)

[Implementation](<https://rajatmodi62.github.io/apm_project_page>)

Recorded evaluation configuration: {"Pre-trained on ImageNet":false}. This record does not assert independent reproduction.

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