# Visual Question Answering on VQA v2 (Accuracy & Efficiency)

Source: [Visual Question Answering on VQA v2 (Accuracy & Efficiency)](<https://www.sota2.com/research/sota/visual-question-answering-on-vqa-v2-accuracy-and-efficiency>)

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

Data updated: 2026-02-26T05:32:15.364Z

Evaluation of model performance (Accuracy) and system efficiency metrics (GPU time and communication overhead) on the VQA v2 dataset for visual question answering.

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 | Visual Question Answering |
| Dataset (including any stated split) | VQA v2 |
| Metric | Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 4 of 4 |
| 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/visual-question-answering-on-vqa-v2-accuracy-and-efficiency.md?metric=Accuracy>): Higher is better; unit: PERCENT.

- [Client GPU Time (GPU-h)](<https://www.sota2.com/research/sota/visual-question-answering-on-vqa-v2-accuracy-and-efficiency.md?metric=Client%20GPU%20Time%20(GPU-h)>): Lower is better; unit: TIME.

- [Server GPU Time (GPU-h)](<https://www.sota2.com/research/sota/visual-question-answering-on-vqa-v2-accuracy-and-efficiency.md?metric=Server%20GPU%20Time%20(GPU-h)>): Lower is better; unit: TIME.

- [Communication Overhead (GB)](<https://www.sota2.com/research/sota/visual-question-answering-on-vqa-v2-accuracy-and-efficiency.md?metric=Communication%20Overhead%20(GB)>): Lower is better; unit: SIZE.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| FedUMM | 80.2 | Not recorded | 69730092b2be0ed85a78ca68 |
| FedProx | 79.8 | Not recorded | 69730092b2be0ed85a78ca68 |
| FedOpt | 79.6 | Not recorded | 69730092b2be0ed85a78ca68 |
| FedAvg | 79.5 | Not recorded | 69730092b2be0ed85a78ca68 |

### FedUMM

Accuracy: 80.2. Result ID: 699924cd5698665ec48c46ab.

[Reporting paper](<https://www.sota2.com/research/paper/69730092b2be0ed85a78ca68.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69730092b2be0ed85a78ca68.md>)

Recorded evaluation configuration: {"Training Protocol":"Parameter-Efficient (LoRA)","Number of Clients (K)":8}. This record does not assert independent reproduction.

### FedProx

Accuracy: 79.8. Result ID: 699924cd5698665ec48c46a9.

[Reporting paper](<https://www.sota2.com/research/paper/69730092b2be0ed85a78ca68.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67c71e209131cea18435.md>)

Recorded evaluation configuration: {"Training Protocol":"Full fine-tuning","Number of Clients (K)":8}. This record does not assert independent reproduction.

### FedOpt

Accuracy: 79.6. Result ID: 699924cd5698665ec48c46aa.

[Reporting paper](<https://www.sota2.com/research/paper/69730092b2be0ed85a78ca68.md>)

Recorded evaluation configuration: {"Training Protocol":"Full fine-tuning","Number of Clients (K)":8}. This record does not assert independent reproduction.

### FedAvg

Accuracy: 79.5. Result ID: 699924cd5698665ec48c46a8.

[Reporting paper](<https://www.sota2.com/research/paper/69730092b2be0ed85a78ca68.md>)

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

Recorded evaluation configuration: {"Training Protocol":"Full fine-tuning","Number of Clients (K)":8}. This record does not assert independent reproduction.

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