# Federated Image Classification on TinyImageNet (95% target accuracy, test)

Source: [Federated Image Classification on TinyImageNet (95% target accuracy, test)](<https://www.sota2.com/research/sota/federated-image-classification-on-tinyimagenet-95-percent-target-accuracy-test>)

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

Data updated: 2026-06-09T09:23:52.634Z

Evaluation of Federated Image Classification performance on the test split of TinyImageNet, specifically measuring efficiency metrics (Latency, Energy, Speedup) required to reach a 95% target accuracy.

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 | Federated Image Classification |
| Dataset (including any stated split) | TinyImageNet 95% target accuracy (test) |
| Metric | Communication Rounds |
| Unit | SCALAR |
| Direction | Lower 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

- [Communication Rounds](<https://www.sota2.com/research/sota/federated-image-classification-on-tinyimagenet-95-percent-target-accuracy-test.md?metric=Communication%20Rounds>): Lower is better; unit: SCALAR.

- [Latency (s)](<https://www.sota2.com/research/sota/federated-image-classification-on-tinyimagenet-95-percent-target-accuracy-test.md?metric=Latency%20(s)>): Lower is better; unit: TIME.

- [Energy (kJ)](<https://www.sota2.com/research/sota/federated-image-classification-on-tinyimagenet-95-percent-target-accuracy-test.md?metric=Energy%20(kJ)>): Lower is better; unit: OTHER.

- [Speedup](<https://www.sota2.com/research/sota/federated-image-classification-on-tinyimagenet-95-percent-target-accuracy-test.md?metric=Speedup>): Higher is better; unit: SCALAR.

- [Accuracy Gain (%)](<https://www.sota2.com/research/sota/federated-image-classification-on-tinyimagenet-95-percent-target-accuracy-test.md?metric=Accuracy%20Gain%20(%25)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Communication Rounds | Result date | Report paper |
| --- | --- | --- | --- |
| Proposed | 1 | Not recorded | 68fa79061e209131cea7bc24 |
| Proposed | 1 | Not recorded | 68fa79061e209131cea7bc24 |
| Average (Baseline) | 7 | Not recorded | 68fa79061e209131cea7bc24 |
| Previous (Baseline) | 11 | Not recorded | 68fa79061e209131cea7bc24 |
| Previous (Baseline) | 14 | Not recorded | 68fa79061e209131cea7bc24 |
| Average (Baseline) | 28 | Not recorded | 68fa79061e209131cea7bc24 |
| Continuous (Baseline) | 55 | Not recorded | 68fa79061e209131cea7bc24 |
| Continuous (Baseline) | 150 | Not recorded | 68fa79061e209131cea7bc24 |

### Proposed

Communication Rounds: 1. Result ID: 6a27db5cfbc373592c24e9c0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79061e209131cea7bc24.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79061e209131cea7bc24.md>)

Recorded evaluation configuration: {"Overlap":"0.2","Dirichlet α":"0.7","FL Alg.":"FedACG"}. This record does not assert independent reproduction.

### Proposed

Communication Rounds: 1. Result ID: 6a27db5cfbc373592c24e9c4.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79061e209131cea7bc24.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79061e209131cea7bc24.md>)

Recorded evaluation configuration: {"Overlap":"0.0","Dirichlet α":"0.3","FL Alg.":"SCAFFOLD"}. This record does not assert independent reproduction.

### Average (Baseline)

Communication Rounds: 7. Result ID: 6a27db5cfbc373592c24e9c3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79061e209131cea7bc24.md>)

Recorded evaluation configuration: {"Overlap":"0.2","Dirichlet α":"0.7","FL Alg.":"FedACG"}. This record does not assert independent reproduction.

### Previous (Baseline)

Communication Rounds: 11. Result ID: 6a27db5cfbc373592c24e9c1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79061e209131cea7bc24.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e761e209131cea40072.md>)

Recorded evaluation configuration: {"Overlap":"0.2","Dirichlet α":"0.7","FL Alg.":"FedACG"}. This record does not assert independent reproduction.

### Previous (Baseline)

Communication Rounds: 14. Result ID: 6a27db5cfbc373592c24e9c5.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79061e209131cea7bc24.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e761e209131cea40072.md>)

Recorded evaluation configuration: {"Overlap":"0.0","Dirichlet α":"0.3","FL Alg.":"SCAFFOLD"}. This record does not assert independent reproduction.

### Average (Baseline)

Communication Rounds: 28. Result ID: 6a27db5cfbc373592c24e9c7.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79061e209131cea7bc24.md>)

Recorded evaluation configuration: {"Overlap":"0.0","Dirichlet α":"0.3","FL Alg.":"SCAFFOLD"}. This record does not assert independent reproduction.

### Continuous (Baseline)

Communication Rounds: 55. Result ID: 6a27db5cfbc373592c24e9c2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79061e209131cea7bc24.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c0d1e209131cea31b3c.md>)

[Implementation](<https://github.com/conditionWang/FCIL>)

Recorded evaluation configuration: {"Overlap":"0.2","Dirichlet α":"0.7","FL Alg.":"FedACG"}. This record does not assert independent reproduction.

### Continuous (Baseline)

Communication Rounds: 150. Result ID: 6a27db5cfbc373592c24e9c6.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79061e209131cea7bc24.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c0d1e209131cea31b3c.md>)

[Implementation](<https://github.com/conditionWang/FCIL>)

Recorded evaluation configuration: {"Overlap":"0.0","Dirichlet α":"0.3","FL Alg.":"SCAFFOLD"}. This record does not assert independent reproduction.

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