# Object Classification on iNaturalist 2018 (Head/Middle/Tail)

Source: [Object Classification on iNaturalist 2018 (Head/Middle/Tail)](<https://www.sota2.com/research/sota/object-classification-on-inaturalist-2018-head-middle-tail>)

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

Data updated: 2026-02-27T06:58:15.472Z

Evaluation of Object Classification performance on the iNaturalist 2018 dataset, reporting Top-1 Accuracy specifically segmented across class frequency groups (Head, Middle, and Tail).

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 | Object Classification |
| Dataset (including any stated split) | iNaturalist 2018 |
| Metric | Top-1 Acc (Head) |
| Unit | PERCENT |
| 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

- [Top-1 Acc (Head)](<https://www.sota2.com/research/sota/object-classification-on-inaturalist-2018-head-middle-tail.md?metric=Top-1%20Acc%20(Head)>): Higher is better; unit: PERCENT.

- [Top-1 Acc (Middle)](<https://www.sota2.com/research/sota/object-classification-on-inaturalist-2018-head-middle-tail.md?metric=Top-1%20Acc%20(Middle)>): Higher is better; unit: PERCENT.

- [Top-1 Acc (Tail)](<https://www.sota2.com/research/sota/object-classification-on-inaturalist-2018-head-middle-tail.md?metric=Top-1%20Acc%20(Tail)>): Higher is better; unit: PERCENT.

- [Top-1 Accuracy](<https://www.sota2.com/research/sota/object-classification-on-inaturalist-2018-head-middle-tail.md?metric=Top-1%20Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Top-1 Acc (Head) | Result date | Report paper |
| --- | --- | --- | --- |
| debLoRA | 72.6 | Not recorded | 68fa75151e209131cea659df |
| LoRA | 71.9 | Not recorded | 68fa75151e209131cea659df |
| Fine-Tune | 66.5 | Not recorded | 68fa75151e209131cea659df |
| Zero-Shot | 36.2 | Not recorded | 68fa75151e209131cea659df |

### debLoRA

Top-1 Acc (Head): 72.6. Result ID: 699373bb3a6c314849160a55.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75151e209131cea659df.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75151e209131cea659df.md>)

[Implementation](<https://github.com/doem97/deblora>)

Recorded evaluation configuration: {"Mode":"debLoRA adaptation"}. This record does not assert independent reproduction.

### LoRA

Top-1 Acc (Head): 71.9. Result ID: 699373bb3a6c314849160a54.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75151e209131cea659df.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b121e209131cea2bf59.md>)

[Implementation](<https://github.com/microsoft/LoRA>)

Recorded evaluation configuration: {"Mode":"Low-Rank Adaptation"}. This record does not assert independent reproduction.

### Fine-Tune

Top-1 Acc (Head): 66.5. Result ID: 699373bb3a6c314849160a53.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75151e209131cea659df.md>)

Recorded evaluation configuration: {"Mode":"Fine-Tuning"}. This record does not assert independent reproduction.

### Zero-Shot

Top-1 Acc (Head): 36.2. Result ID: 699373bb3a6c314849160a52.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75151e209131cea659df.md>)

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

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