# Few-shot Image Classification on ImageNet Transfer Learning Suite (20-Shot, test)

Source: [Few-shot Image Classification on ImageNet Transfer Learning Suite (20-Shot, test)](<https://www.sota2.com/research/sota/few-shot-image-classification-on-imagenet-transfer-learning-suite-20-shot-test>)

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

Data updated: 2026-02-26T05:34:06.229Z

Evaluation of Few-shot Image Classification performance using a 20-shot setting across various downstream classification tasks included in the ImageNet Transfer Learning Suite (e.g., Aircraft, CIFAR-10, SUN). Results are typically reported as the average accuracy across these tasks.

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 | Few-shot Image Classification |
| Dataset (including any stated split) | ImageNet Transfer Learning Suite 20-Shot (test) |
| Metric | Accuracy (Aircraft) |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 6 of 6 |
| 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

- [Accuracy (Aircraft)](<https://www.sota2.com/research/sota/few-shot-image-classification-on-imagenet-transfer-learning-suite-20-shot-test.md?metric=Accuracy%20(Aircraft)>): Higher is better; unit: PERCENT.

- [Accuracy (Caltech)](<https://www.sota2.com/research/sota/few-shot-image-classification-on-imagenet-transfer-learning-suite-20-shot-test.md?metric=Accuracy%20(Caltech)>): Higher is better; unit: PERCENT.

- [Accuracy (Cars)](<https://www.sota2.com/research/sota/few-shot-image-classification-on-imagenet-transfer-learning-suite-20-shot-test.md?metric=Accuracy%20(Cars)>): Higher is better; unit: PERCENT.

- [Accuracy (CIFAR-10)](<https://www.sota2.com/research/sota/few-shot-image-classification-on-imagenet-transfer-learning-suite-20-shot-test.md?metric=Accuracy%20(CIFAR-10)>): Higher is better; unit: PERCENT.

- [Accuracy (CIFAR-100)](<https://www.sota2.com/research/sota/few-shot-image-classification-on-imagenet-transfer-learning-suite-20-shot-test.md?metric=Accuracy%20(CIFAR-100)>): Higher is better; unit: PERCENT.

- [Accuracy (DTD)](<https://www.sota2.com/research/sota/few-shot-image-classification-on-imagenet-transfer-learning-suite-20-shot-test.md?metric=Accuracy%20(DTD)>): Higher is better; unit: PERCENT.

- [Accuracy (Flowers)](<https://www.sota2.com/research/sota/few-shot-image-classification-on-imagenet-transfer-learning-suite-20-shot-test.md?metric=Accuracy%20(Flowers)>): Higher is better; unit: PERCENT.

- [Accuracy (Food)](<https://www.sota2.com/research/sota/few-shot-image-classification-on-imagenet-transfer-learning-suite-20-shot-test.md?metric=Accuracy%20(Food)>): Higher is better; unit: PERCENT.

- [Accuracy (Pets)](<https://www.sota2.com/research/sota/few-shot-image-classification-on-imagenet-transfer-learning-suite-20-shot-test.md?metric=Accuracy%20(Pets)>): Higher is better; unit: PERCENT.

- [Accuracy (SUN)](<https://www.sota2.com/research/sota/few-shot-image-classification-on-imagenet-transfer-learning-suite-20-shot-test.md?metric=Accuracy%20(SUN)>): Higher is better; unit: PERCENT.

- [Average Accuracy (20-Shot)](<https://www.sota2.com/research/sota/few-shot-image-classification-on-imagenet-transfer-learning-suite-20-shot-test.md?metric=Average%20Accuracy%20(20-Shot)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy (Aircraft) | Result date | Report paper |
| --- | --- | --- | --- |
| PIRL | 48.15 | Not recorded | 68fa6b2c1e209131cea2c8e9 |
| InsDis | 47.44 | Not recorded | 68fa6b2c1e209131cea2c8e9 |
| MoCo-v1 | 47.08 | Not recorded | 68fa6b2c1e209131cea2c8e9 |
| PCL | 44.72 | Not recorded | 68fa6b2c1e209131cea2c8e9 |
| IP-IRM | 43.92 | Not recorded | 68fa6b2c1e209131cea2c8e9 |
| MoCo-v2 | 42.57 | Not recorded | 68fa6b2c1e209131cea2c8e9 |

### PIRL

Accuracy (Aircraft): 48.15. Result ID: 698e6777ffc3901de09d4db2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b2c1e209131cea2c8e9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66cc1e209131cea12694.md>)

Recorded evaluation configuration: {"Regime":"20-Shot"}. This record does not assert independent reproduction.

### InsDis

Accuracy (Aircraft): 47.44. Result ID: 698e6777ffc3901de09d4db0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b2c1e209131cea2c8e9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa644b1e209131cea03c73.md>)

[Implementation](<http://github.com/zhirongw/lemniscate.pytorch>)

Recorded evaluation configuration: {"Regime":"20-Shot"}. This record does not assert independent reproduction.

### MoCo-v1

Accuracy (Aircraft): 47.08. Result ID: 698e6777ffc3901de09d4db3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b2c1e209131cea2c8e9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa673f1e209131cea15344.md>)

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

Recorded evaluation configuration: {"Regime":"20-Shot"}. This record does not assert independent reproduction.

### PCL

Accuracy (Aircraft): 44.72. Result ID: 698e6777ffc3901de09d4db1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b2c1e209131cea2c8e9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69981e209131cea22eea.md>)

[Implementation](<https://github.com/salesforce/PCL>)

Recorded evaluation configuration: {"Regime":"20-Shot"}. This record does not assert independent reproduction.

### IP-IRM

Accuracy (Aircraft): 43.92. Result ID: 698e6777ffc3901de09d4db5.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b2c1e209131cea2c8e9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b2c1e209131cea2c8e9.md>)

[Implementation](<https://github.com/Wangt-CN/IP-IRM>)

Recorded evaluation configuration: {"Regime":"20-Shot"}. This record does not assert independent reproduction.

### MoCo-v2

Accuracy (Aircraft): 42.57. Result ID: 698e6777ffc3901de09d4db4.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b2c1e209131cea2c8e9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67371e209131cea14e30.md>)

Recorded evaluation configuration: {"Regime":"20-Shot"}. This record does not assert independent reproduction.

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