# Image Classification on CIFAR-100 (test) (Instance Accuracy Metrics)

Source: [Image Classification on CIFAR-100 (test) (Instance Accuracy Metrics)](<https://www.sota2.com/research/sota/image-classification-on-cifar-100-test-instance-accuracy-metrics>)

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

Data updated: 2026-02-26T05:34:32.510Z

Evaluation of Image Classification performance on the CIFAR-100 test set, focusing specifically on accuracy under conditions defined by 20%, 30%, and 40% instance-level perturbations or noise.

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 (ins-20%) |
| Unit | PERCENT |
| 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

- [Accuracy (ins-20%)](<https://www.sota2.com/research/sota/image-classification-on-cifar-100-test-instance-accuracy-metrics.md?metric=Accuracy%20(ins-20%25)>): Higher is better; unit: PERCENT.

- [Accuracy (ins-30%)](<https://www.sota2.com/research/sota/image-classification-on-cifar-100-test-instance-accuracy-metrics.md?metric=Accuracy%20(ins-30%25)>): Higher is better; unit: PERCENT.

- [Accuracy (ins-40%)](<https://www.sota2.com/research/sota/image-classification-on-cifar-100-test-instance-accuracy-metrics.md?metric=Accuracy%20(ins-40%25)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy (ins-20%) | Result date | Report paper |
| --- | --- | --- | --- |
| Co-learning+HCR | 70.03 | Not recorded | 68fa6cb81e209131cea358e5 |
| Co-learning | 69.42 | Not recorded | 68fa6cb81e209131cea358e5 |
| Standard CE | 55.45 | Not recorded | 68fa6cb81e209131cea358e5 |
| Co-teaching | 55.16 | Not recorded | 68fa6cb81e209131cea358e5 |
| JoCoR | 54.21 | Not recorded | 68fa6cb81e209131cea358e5 |
| Decoupling | 52.2 | Not recorded | 68fa6cb81e209131cea358e5 |
| Co-teaching+ | 50.37 | Not recorded | 68fa6cb81e209131cea358e5 |

### Co-learning+HCR

Accuracy (ins-20%): 70.03. Result ID: 698f35413a6c314849f7020f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6cb81e209131cea358e5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6cb81e209131cea358e5.md>)

Recorded evaluation configuration: {"Backbone":"ResNet-18","Number of trials":5,"Evaluation Protocol":"Average test accuracy over last 10 epochs"}. This record does not assert independent reproduction.

### Co-learning

Accuracy (ins-20%): 69.42. Result ID: 698f35413a6c314849f7020e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6cb81e209131cea358e5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b2e1e209131cea2ca11.md>)

Recorded evaluation configuration: {"Backbone":"ResNet-18","Number of trials":5,"Evaluation Protocol":"Average test accuracy over last 10 epochs"}. This record does not assert independent reproduction.

### Standard CE

Accuracy (ins-20%): 55.45. Result ID: 698f35413a6c314849f70209.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6cb81e209131cea358e5.md>)

Recorded evaluation configuration: {"Backbone":"ResNet-18","Number of trials":5,"Evaluation Protocol":"Average test accuracy over last 10 epochs"}. This record does not assert independent reproduction.

### Co-teaching

Accuracy (ins-20%): 55.16. Result ID: 698f35413a6c314849f7020b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6cb81e209131cea358e5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa64fe1e209131cea07aa4.md>)

[Implementation](<https://github.com/bhanML/Co-teaching>)

Recorded evaluation configuration: {"Backbone":"ResNet-18","Number of trials":5,"Evaluation Protocol":"Average test accuracy over last 10 epochs"}. This record does not assert independent reproduction.

### JoCoR

Accuracy (ins-20%): 54.21. Result ID: 698f35413a6c314849f7020d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6cb81e209131cea358e5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67711e209131cea164c7.md>)

Recorded evaluation configuration: {"Backbone":"ResNet-18","Number of trials":5,"Evaluation Protocol":"Average test accuracy over last 10 epochs"}. This record does not assert independent reproduction.

### Decoupling

Accuracy (ins-20%): 52.2. Result ID: 698f35413a6c314849f7020a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6cb81e209131cea358e5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa642f1e209131cea0344e.md>)

Recorded evaluation configuration: {"Backbone":"ResNet-18","Number of trials":5,"Evaluation Protocol":"Average test accuracy over last 10 epochs"}. This record does not assert independent reproduction.

### Co-teaching+

Accuracy (ins-20%): 50.37. Result ID: 698f35413a6c314849f7020c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6cb81e209131cea358e5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa65bb1e209131cea0bed4.md>)

Recorded evaluation configuration: {"Backbone":"ResNet-18","Number of trials":5,"Evaluation Protocol":"Average test accuracy over last 10 epochs"}. This record does not assert independent reproduction.

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