# Image Classification on CIFAR-100 (test) (Accuracy, Calibration, and Detection Metrics)

Source: [Image Classification on CIFAR-100 (test) (Accuracy, Calibration, and Detection Metrics)](<https://www.sota2.com/research/sota/image-classification-on-cifar-100-test-accuracy-calibration-and-detection-metrics>)

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

Data updated: 2026-02-26T05:29:22.060Z

Evaluation of Image Classification performance on the CIFAR-100 test set, reporting standard accuracy, negative log-likelihood (NLL), expected calibration error (ECE), AUPR, AUC, and FPR95.

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) | CIFAR-100 (test) |
| Metric | Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 14 of 14 |
| 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](<https://www.sota2.com/research/sota/image-classification-on-cifar-100-test-accuracy-calibration-and-detection-metrics.md?metric=Accuracy>): Higher is better; unit: PERCENT.

- [NLL](<https://www.sota2.com/research/sota/image-classification-on-cifar-100-test-accuracy-calibration-and-detection-metrics.md?metric=NLL>): Lower is better; unit: SCALAR.

- [ECE](<https://www.sota2.com/research/sota/image-classification-on-cifar-100-test-accuracy-calibration-and-detection-metrics.md?metric=ECE>): Lower is better; unit: PERCENT.

- [AUPR](<https://www.sota2.com/research/sota/image-classification-on-cifar-100-test-accuracy-calibration-and-detection-metrics.md?metric=AUPR>): Higher is better; unit: PERCENT.

- [AUC](<https://www.sota2.com/research/sota/image-classification-on-cifar-100-test-accuracy-calibration-and-detection-metrics.md?metric=AUC>): Higher is better; unit: PERCENT.

- [FPR@95](<https://www.sota2.com/research/sota/image-classification-on-cifar-100-test-accuracy-calibration-and-detection-metrics.md?metric=FPR%4095>): Lower is better; unit: PERCENT.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| Deep Ensembles | 82.5 | Not recorded | 68fa71ad1e209131cea527f6 |
| BatchEnsemble | 82.3 | Not recorded | 68fa71ad1e209131cea527f6 |
| Deep Ensembles | 80.9 | Not recorded | 68fa71ad1e209131cea527f6 |
| ABNN | 80.4 | Not recorded | 68fa71ad1e209131cea527f6 |
| Single Model | 80.3 | Not recorded | 68fa71ad1e209131cea527f6 |
| MIMO (ρ = 1) | 80.2 | Not recorded | 68fa71ad1e209131cea527f6 |
| Laplace | 80.1 | Not recorded | 68fa71ad1e209131cea527f6 |
| LPBNN | 79.7 | Not recorded | 68fa71ad1e209131cea527f6 |
| MIMO (ρ = 1) | 79 | Not recorded | 68fa71ad1e209131cea527f6 |
| ABNN | 78.9 | Not recorded | 68fa71ad1e209131cea527f6 |
| LPBNN | 78.5 | Not recorded | 68fa71ad1e209131cea527f6 |
| Single Model | 78.3 | Not recorded | 68fa71ad1e209131cea527f6 |
| Laplace | 78.2 | Not recorded | 68fa71ad1e209131cea527f6 |
| BatchEnsemble | 66.6 | Not recorded | 68fa71ad1e209131cea527f6 |

### Deep Ensembles

Accuracy: 82.5. Result ID: 6991db4d3a6c3148490b07d1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71ad1e209131cea527f6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63bc1e209131cea0084f.md>)

Recorded evaluation configuration: {"Backbone":"WideResNet-28×10","ΔParam":109.64,"Time (h)":16.6,"M":4}. This record does not assert independent reproduction.

### BatchEnsemble

Accuracy: 82.3. Result ID: 6991db4d3a6c3148490b07ce.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71ad1e209131cea527f6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa671c1e209131cea1455e.md>)

Recorded evaluation configuration: {"Backbone":"WideResNet-28×10","ΔParam":0.1,"Time (h)":25.6,"M":4}. This record does not assert independent reproduction.

### Deep Ensembles

Accuracy: 80.9. Result ID: 6991db4d3a6c3148490b07ca.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71ad1e209131cea527f6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63bc1e209131cea0084f.md>)

Recorded evaluation configuration: {"Backbone":"ResNet-50","ΔParam":71.12,"Time (h)":6.8,"M":4}. This record does not assert independent reproduction.

### ABNN

Accuracy: 80.4. Result ID: 6991db4d3a6c3148490b07d3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71ad1e209131cea527f6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71ad1e209131cea527f6.md>)

Recorded evaluation configuration: {"Backbone":"WideResNet-28×10","ΔParam":0.05,"Time (h)":5,"M":4}. This record does not assert independent reproduction.

### Single Model

Accuracy: 80.3. Result ID: 6991db4d3a6c3148490b07cd.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71ad1e209131cea527f6.md>)

Recorded evaluation configuration: {"Backbone":"WideResNet-28×10","ΔParam":null,"Time (h)":4.2}. This record does not assert independent reproduction.

### MIMO (ρ = 1)

Accuracy: 80.2. Result ID: 6991db4d3a6c3148490b07cf.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71ad1e209131cea527f6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a7d1e209131cea28948.md>)

Recorded evaluation configuration: {"Backbone":"WideResNet-28×10","ΔParam":0.19,"Time (h)":12.6,"M":4}. This record does not assert independent reproduction.

### Laplace

Accuracy: 80.1. Result ID: 6991db4d3a6c3148490b07d2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71ad1e209131cea527f6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6bfe1e209131cea31830.md>)

Recorded evaluation configuration: {"Backbone":"WideResNet-28×10","ΔParam":null,"Time (h)":4.2,"M":4}. This record does not assert independent reproduction.

### LPBNN

Accuracy: 79.7. Result ID: 6991db4d3a6c3148490b07d0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71ad1e209131cea527f6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa699d1e209131cea23158.md>)

Recorded evaluation configuration: {"Backbone":"WideResNet-28×10","ΔParam":0.72,"Time (h)":23.3,"M":4}. This record does not assert independent reproduction.

### MIMO (ρ = 1)

Accuracy: 79. Result ID: 6991db4d3a6c3148490b07c8.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71ad1e209131cea527f6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a7d1e209131cea28948.md>)

Recorded evaluation configuration: {"Backbone":"ResNet-50","ΔParam":0.63,"Time (h)":6.7,"M":4}. This record does not assert independent reproduction.

### ABNN

Accuracy: 78.9. Result ID: 6991db4d3a6c3148490b07cc.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71ad1e209131cea527f6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71ad1e209131cea527f6.md>)

Recorded evaluation configuration: {"Backbone":"ResNet-50","ΔParam":0.16,"Time (h)":2,"M":4}. This record does not assert independent reproduction.

### LPBNN

Accuracy: 78.5. Result ID: 6991db4d3a6c3148490b07c9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71ad1e209131cea527f6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa699d1e209131cea23158.md>)

Recorded evaluation configuration: {"Backbone":"ResNet-50","ΔParam":1.83,"Time (h)":17.2,"M":4}. This record does not assert independent reproduction.

### Single Model

Accuracy: 78.3. Result ID: 6991db4d3a6c3148490b07c6.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71ad1e209131cea527f6.md>)

Recorded evaluation configuration: {"Backbone":"ResNet-50","ΔParam":null,"Time (h)":1.7}. This record does not assert independent reproduction.

### Laplace

Accuracy: 78.2. Result ID: 6991db4d3a6c3148490b07cb.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71ad1e209131cea527f6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6bfe1e209131cea31830.md>)

Recorded evaluation configuration: {"Backbone":"ResNet-50","ΔParam":null,"Time (h)":1.7,"M":4}. This record does not assert independent reproduction.

### BatchEnsemble

Accuracy: 66.6. Result ID: 6991db4d3a6c3148490b07c7.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71ad1e209131cea527f6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa671c1e209131cea1455e.md>)

Recorded evaluation configuration: {"Backbone":"ResNet-50","ΔParam":0.11,"Time (h)":17.2,"M":4}. This record does not assert independent reproduction.

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