# Image Classification on ImageNet-1K (test val) (ECE Calibration)

Source: [Image Classification on ImageNet-1K (test val) (ECE Calibration)](<https://www.sota2.com/research/sota/image-classification-on-imagenet-1k-test-val-ece-calibration>)

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

Data updated: 2026-02-26T05:41:22.366Z

Evaluation of Image Classification model calibration on the ImageNet-1K test/validation split, reporting Expected Calibration Error (ECE) before and after temperature scaling.

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-1K (test val) |
| Metric | ECE |
| Unit | SCALAR |
| Direction | Lower is better |
| Results loaded | 3 of 3 |
| 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

- [ECE](<https://www.sota2.com/research/sota/image-classification-on-imagenet-1k-test-val-ece-calibration.md?metric=ECE>): Lower is better; unit: SCALAR.

- [ECE (TS)](<https://www.sota2.com/research/sota/image-classification-on-imagenet-1k-test-val-ece-calibration.md?metric=ECE%20(TS)>): Lower is better; unit: SCALAR.

## Recorded results

| Method | ECE | Result date | Report paper |
| --- | --- | --- | --- |
| BSCE-GRA | 0.0263 | Not recorded | 68fa77bd1e209131cea750a1 |
| Focal Loss (FL) | 0.0328 | Not recorded | 68fa77bd1e209131cea750a1 |
| Cross Entropy (CE) | 0.0369 | Not recorded | 68fa77bd1e209131cea750a1 |

### BSCE-GRA

ECE: 0.0263. Result ID: 699564053badfaffe656e5bd.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77bd1e209131cea750a1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77bd1e209131cea750a1.md>)

[Implementation](<https://github.com/Jinxu-Lin/BSCE-GRA>)

Recorded evaluation configuration: {"Loss":"BSCE-GRA","Backbone":"ResNet-50","Pre-training":"Pre-trained on IN-1K","Epochs":"90","Optimizer":"Adam","Learning Rate Scheduler":"Cosine Learning Rate Scheduler"}. This record does not assert independent reproduction.

### Focal Loss (FL)

ECE: 0.0328. Result ID: 699564053badfaffe656e5bc.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77bd1e209131cea750a1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa640c1e209131cea02556.md>)

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

Recorded evaluation configuration: {"Loss":"Focal Loss (FL)","Backbone":"ResNet-50","Pre-training":"Pre-trained on IN-1K","Epochs":"90","Optimizer":"Adam","Learning Rate Scheduler":"Cosine Learning Rate Scheduler"}. This record does not assert independent reproduction.

### Cross Entropy (CE)

ECE: 0.0369. Result ID: 699564053badfaffe656e5bb.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77bd1e209131cea750a1.md>)

Recorded evaluation configuration: {"Loss":"Cross Entropy (CE)","Backbone":"ResNet-50","Pre-training":"Pre-trained on IN-1K","Epochs":"90","Optimizer":"Adam","Learning Rate Scheduler":"Cosine Learning Rate Scheduler"}. This record does not assert independent reproduction.

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