# Long-tailed Image Classification on iNaturalist 2018 (Shot Accuracy Breakdown)

Source: [Long-tailed Image Classification on iNaturalist 2018 (Shot Accuracy Breakdown)](<https://www.sota2.com/research/sota/long-tailed-image-classification-on-inaturalist-2018-shot-accuracy-breakdown>)

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

Data updated: 2026-06-02T13:42:38.463Z

Evaluation of Long-Tailed Image Classification performance on the iNaturalist 2018 dataset, specifically reporting accuracy metrics segmented by the number of training examples per class (Many-shot, Medium-shot, Few-shot).

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 | Long-tailed Image Classification |
| Dataset (including any stated split) | iNaturalist 2018 |
| Metric | Overall Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 10 of 10 |
| 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

- [Overall Accuracy](<https://www.sota2.com/research/sota/long-tailed-image-classification-on-inaturalist-2018-shot-accuracy-breakdown.md?metric=Overall%20Accuracy>): Higher is better; unit: PERCENT.

- [Many-shot Accuracy](<https://www.sota2.com/research/sota/long-tailed-image-classification-on-inaturalist-2018-shot-accuracy-breakdown.md?metric=Many-shot%20Accuracy>): Higher is better; unit: PERCENT.

- [Medium-shot Accuracy](<https://www.sota2.com/research/sota/long-tailed-image-classification-on-inaturalist-2018-shot-accuracy-breakdown.md?metric=Medium-shot%20Accuracy>): Higher is better; unit: PERCENT.

- [Few-shot Accuracy](<https://www.sota2.com/research/sota/long-tailed-image-classification-on-inaturalist-2018-shot-accuracy-breakdown.md?metric=Few-shot%20Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Overall Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| GLMC + SAMN | 72.7 | Not recorded | 6a1e6455dbd7150506f859fe |
| GLMC | 72.2 | Not recorded | 6a1e6455dbd7150506f859fe |
| IWB | 71.5 | Not recorded | 6a1e6455dbd7150506f859fe |
| WD + MaxNorm | 70.2 | Not recorded | 6a1e6455dbd7150506f859fe |
| LogitAdjust | 69.4 | Not recorded | 6a1e6455dbd7150506f859fe |
| DiVE | 69.1 | Not recorded | 6a1e6455dbd7150506f859fe |
| AREA | 68.4 | Not recorded | 6a1e6455dbd7150506f859fe |
| τ-norm | 65.6 | Not recorded | 6a1e6455dbd7150506f859fe |
| cRT | 65.2 | Not recorded | 6a1e6455dbd7150506f859fe |
| CE | 61.7 | Not recorded | 6a1e6455dbd7150506f859fe |

### GLMC + SAMN

Overall Accuracy: 72.7. Result ID: 6a1eddb1fbc373592c1b0abd.

[Reporting paper](<https://www.sota2.com/research/paper/6a1e6455dbd7150506f859fe.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a1e6455dbd7150506f859fe.md>)

Recorded evaluation configuration: {"Backbone":"ResNeXt50"}. This record does not assert independent reproduction.

### GLMC

Overall Accuracy: 72.2. Result ID: 6a1eddb1fbc373592c1b0abc.

[Reporting paper](<https://www.sota2.com/research/paper/6a1e6455dbd7150506f859fe.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f581e209131cea45243.md>)

[Implementation](<https://github.com/ynu-yangpeng/GLMC>)

Recorded evaluation configuration: {"Backbone":"ResNeXt50"}. This record does not assert independent reproduction.

### IWB

Overall Accuracy: 71.5. Result ID: 6a1eddb1fbc373592c1b0abb.

[Reporting paper](<https://www.sota2.com/research/paper/6a1e6455dbd7150506f859fe.md>)

Recorded evaluation configuration: {"Backbone":"ResNeXt50"}. This record does not assert independent reproduction.

### WD + MaxNorm

Overall Accuracy: 70.2. Result ID: 6a1eddb1fbc373592c1b0aba.

[Reporting paper](<https://www.sota2.com/research/paper/6a1e6455dbd7150506f859fe.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c1b1e209131cea32418.md>)

[Implementation](<https://github.com/ShadeAlsha/LTR-weight-balancing>)

Recorded evaluation configuration: {"Backbone":"ResNeXt50"}. This record does not assert independent reproduction.

### LogitAdjust

Overall Accuracy: 69.4. Result ID: 6a1eddb1fbc373592c1b0ab8.

[Reporting paper](<https://www.sota2.com/research/paper/6a1e6455dbd7150506f859fe.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a531e209131cea275d5.md>)

[Implementation](<https://github.com/google-research/google-research/tree/master/logit_adjustment>)

Recorded evaluation configuration: {"Backbone":"ResNeXt50"}. This record does not assert independent reproduction.

### DiVE

Overall Accuracy: 69.1. Result ID: 6a1eddb1fbc373592c1b0ab9.

[Reporting paper](<https://www.sota2.com/research/paper/6a1e6455dbd7150506f859fe.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6acb1e209131cea2a4a4.md>)

Recorded evaluation configuration: {"Backbone":"ResNeXt50"}. This record does not assert independent reproduction.

### AREA

Overall Accuracy: 68.4. Result ID: 6a1eddb1fbc373592c1b0ab7.

[Reporting paper](<https://www.sota2.com/research/paper/6a1e6455dbd7150506f859fe.md>)

Recorded evaluation configuration: {"Backbone":"ResNeXt50"}. This record does not assert independent reproduction.

### τ-norm

Overall Accuracy: 65.6. Result ID: 6a1eddb1fbc373592c1b0ab5.

[Reporting paper](<https://www.sota2.com/research/paper/6a1e6455dbd7150506f859fe.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa671b1e209131cea144a9.md>)

[Implementation](<https://github.com/facebookresearch/classifier-balancing>)

Recorded evaluation configuration: {"Backbone":"ResNeXt50"}. This record does not assert independent reproduction.

### cRT

Overall Accuracy: 65.2. Result ID: 6a1eddb1fbc373592c1b0ab6.

[Reporting paper](<https://www.sota2.com/research/paper/6a1e6455dbd7150506f859fe.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa671b1e209131cea144a9.md>)

[Implementation](<https://github.com/facebookresearch/classifier-balancing>)

Recorded evaluation configuration: {"Backbone":"ResNeXt50"}. This record does not assert independent reproduction.

### CE

Overall Accuracy: 61.7. Result ID: 6a1eddb1fbc373592c1b0ab4.

[Reporting paper](<https://www.sota2.com/research/paper/6a1e6455dbd7150506f859fe.md>)

Recorded evaluation configuration: {"Backbone":"ResNeXt50"}. This record does not assert independent reproduction.

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