# Image Classification on ImageNet (BD-Rate and BD-Acc)

Source: [Image Classification on ImageNet (BD-Rate and BD-Acc)](<https://www.sota2.com/research/sota/image-classification-on-imagenet-bd-rate-and-bd-acc>)

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

Data updated: 2026-04-14T07:50:43.730Z

Evaluation of Image Classification performance on the standard ImageNet dataset, focusing on metrics related to BD-Rate and BD-Acc.

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 | Image Classification |
| Dataset (including any stated split) | ImageNet |
| Metric | BD-Acc |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 7 of 7 |
| 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

- [BD-Acc](<https://www.sota2.com/research/sota/image-classification-on-imagenet-bd-rate-and-bd-acc.md?metric=BD-Acc>): Higher is better; unit: SCALAR.

- [BD-Rate](<https://www.sota2.com/research/sota/image-classification-on-imagenet-bd-rate-and-bd-acc.md?metric=BD-Rate>): Lower is better; unit: SCALAR.

## Recorded results

| Method | BD-Acc | Result date | Report paper |
| --- | --- | --- | --- |
| full fine-tuning | 17.688 | Not recorded | 69ddc9ed7d0172140ec79852 |
| Ours | 17.351 | Not recorded | 69ddc9ed7d0172140ec79852 |
| Adapt-ICMH | 16.901 | Not recorded | 69ddc9ed7d0172140ec79852 |
| TransTIC | 9.956 | Not recorded | 69ddc9ed7d0172140ec79852 |
| SVD-LoRA | 7.92 | Not recorded | 69ddc9ed7d0172140ec79852 |
| Channel Selection | 6.278 | Not recorded | 69ddc9ed7d0172140ec79852 |
| ICMH-Net | 3.36 | Not recorded | 69ddc9ed7d0172140ec79852 |

### full fine-tuning

BD-Acc: 17.688. Result ID: 69ddf11447f34ce924c51b19.

[Reporting paper](<https://www.sota2.com/research/paper/69ddc9ed7d0172140ec79852.md>)

Recorded evaluation configuration: {"Codec":"Lu2022-TIC","Trainable Params":"7.51 (100%)"}. This record does not assert independent reproduction.

### Ours

BD-Acc: 17.351. Result ID: 69ddf11447f34ce924c51b1f.

[Reporting paper](<https://www.sota2.com/research/paper/69ddc9ed7d0172140ec79852.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69ddc9ed7d0172140ec79852.md>)

Recorded evaluation configuration: {"Codec":"Lu2022-TIC","Trainable Params":"0.42 (5.59%)"}. This record does not assert independent reproduction.

### Adapt-ICMH

BD-Acc: 16.901. Result ID: 69ddf11447f34ce924c51b1d.

[Reporting paper](<https://www.sota2.com/research/paper/69ddc9ed7d0172140ec79852.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74261e209131cea60072.md>)

Recorded evaluation configuration: {"Codec":"Lu2022-TIC","Trainable Params":"0.29 (3.86%)"}. This record does not assert independent reproduction.

### TransTIC

BD-Acc: 9.956. Result ID: 69ddf11447f34ce924c51b1b.

[Reporting paper](<https://www.sota2.com/research/paper/69ddc9ed7d0172140ec79852.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7fee1e209131cea9b821.md>)

Recorded evaluation configuration: {"Codec":"Lu2022-TIC","Trainable Params":"1.62 (21.57%)"}. This record does not assert independent reproduction.

### SVD-LoRA

BD-Acc: 7.92. Result ID: 69ddf11447f34ce924c51b1e.

[Reporting paper](<https://www.sota2.com/research/paper/69ddc9ed7d0172140ec79852.md>)

Recorded evaluation configuration: {"Codec":"Lu2022-TIC","Trainable Params":"0.09 (1.20%)"}. This record does not assert independent reproduction.

### Channel Selection

BD-Acc: 6.278. Result ID: 69ddf11447f34ce924c51b1a.

[Reporting paper](<https://www.sota2.com/research/paper/69ddc9ed7d0172140ec79852.md>)

Recorded evaluation configuration: {"Codec":"Lu2022-TIC","Trainable Params":"0.92 (12.25%)"}. This record does not assert independent reproduction.

### ICMH-Net

BD-Acc: 3.36. Result ID: 69ddf11447f34ce924c51b1c.

[Reporting paper](<https://www.sota2.com/research/paper/69ddc9ed7d0172140ec79852.md>)

Recorded evaluation configuration: {"Codec":"Lu2022-TIC","Trainable Params":"3.98 (53.00%)"}. This record does not assert independent reproduction.

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