# Visual Question Answering on VQA v2 (Detailed Metric Breakdown)

Source: [Visual Question Answering on VQA v2 (Detailed Metric Breakdown)](<https://www.sota2.com/research/sota/visual-question-answering-on-vqa-v2-detailed-metric-breakdown>)

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

Data updated: 2026-06-11T07:45:20.363Z

Evaluation of model performance on the VQA v2 dataset for Visual Question Answering, reporting detailed accuracy breakdowns across 12 specific categories (Rec., Loc., Jud., Com., Cou., Act., Col., Typ., Sub., Cau., AP, SF).

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 (Rec.) |
| 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 (Rec.)](<https://www.sota2.com/research/sota/visual-question-answering-on-vqa-v2-detailed-metric-breakdown.md?metric=Accuracy%20(Rec.)>): Higher is better; unit: PERCENT.

- [Accuracy (Loc.)](<https://www.sota2.com/research/sota/visual-question-answering-on-vqa-v2-detailed-metric-breakdown.md?metric=Accuracy%20(Loc.)>): Higher is better; unit: PERCENT.

- [Accuracy (Jud.)](<https://www.sota2.com/research/sota/visual-question-answering-on-vqa-v2-detailed-metric-breakdown.md?metric=Accuracy%20(Jud.)>): Higher is better; unit: PERCENT.

- [Accuracy (Com.)](<https://www.sota2.com/research/sota/visual-question-answering-on-vqa-v2-detailed-metric-breakdown.md?metric=Accuracy%20(Com.)>): Higher is better; unit: PERCENT.

- [Accuracy (Cou.)](<https://www.sota2.com/research/sota/visual-question-answering-on-vqa-v2-detailed-metric-breakdown.md?metric=Accuracy%20(Cou.)>): Higher is better; unit: PERCENT.

- [Accuracy (Act.)](<https://www.sota2.com/research/sota/visual-question-answering-on-vqa-v2-detailed-metric-breakdown.md?metric=Accuracy%20(Act.)>): Higher is better; unit: PERCENT.

- [Accuracy (Col.)](<https://www.sota2.com/research/sota/visual-question-answering-on-vqa-v2-detailed-metric-breakdown.md?metric=Accuracy%20(Col.)>): Higher is better; unit: PERCENT.

- [Accuracy (Typ.)](<https://www.sota2.com/research/sota/visual-question-answering-on-vqa-v2-detailed-metric-breakdown.md?metric=Accuracy%20(Typ.)>): Higher is better; unit: PERCENT.

- [Accuracy (Sub.)](<https://www.sota2.com/research/sota/visual-question-answering-on-vqa-v2-detailed-metric-breakdown.md?metric=Accuracy%20(Sub.)>): Higher is better; unit: PERCENT.

- [Accuracy (Cau.)](<https://www.sota2.com/research/sota/visual-question-answering-on-vqa-v2-detailed-metric-breakdown.md?metric=Accuracy%20(Cau.)>): Higher is better; unit: PERCENT.

- [AP Score](<https://www.sota2.com/research/sota/visual-question-answering-on-vqa-v2-detailed-metric-breakdown.md?metric=AP%20Score>): Higher is better; unit: PERCENT.

- [Accuracy (SF)](<https://www.sota2.com/research/sota/visual-question-answering-on-vqa-v2-detailed-metric-breakdown.md?metric=Accuracy%20(SF)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy (Rec.) | Result date | Report paper |
| --- | --- | --- | --- |
| LoRAMoE + Muon | 57.72 | Not recorded | 6a28f32190966ade919c163c |
| LoRAMoE + FOGO | 57.72 | Not recorded | 6a28f32190966ade919c163c |
| LoRAMoE + AdamW | 55.88 | Not recorded | 6a28f32190966ade919c163c |
| CL-MoE | 55.74 | Not recorded | 6a28f32190966ade919c163c |

### LoRAMoE + Muon

Accuracy (Rec.): 57.72. Result ID: 6a2a65e8fbc373592c272dfb.

[Reporting paper](<https://www.sota2.com/research/paper/6a28f32190966ade919c163c.md>)

Recorded evaluation configuration: {"Architecture":"LoRAMoE","Optimizer":"Muon"}. This record does not assert independent reproduction.

### LoRAMoE + FOGO

Accuracy (Rec.): 57.72. Result ID: 6a2a65e8fbc373592c272dfc.

[Reporting paper](<https://www.sota2.com/research/paper/6a28f32190966ade919c163c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a28f32190966ade919c163c.md>)

Recorded evaluation configuration: {"Architecture":"LoRAMoE","Optimizer":"FOGO"}. This record does not assert independent reproduction.

### LoRAMoE + AdamW

Accuracy (Rec.): 55.88. Result ID: 6a2a65e8fbc373592c272dfa.

[Reporting paper](<https://www.sota2.com/research/paper/6a28f32190966ade919c163c.md>)

Recorded evaluation configuration: {"Architecture":"LoRAMoE","Optimizer":"AdamW"}. This record does not assert independent reproduction.

### CL-MoE

Accuracy (Rec.): 55.74. Result ID: 6a2a65e8fbc373592c272df9.

[Reporting paper](<https://www.sota2.com/research/paper/6a28f32190966ade919c163c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77421e209131cea720fd.md>)

Recorded evaluation configuration: {"Architecture":"CL-MoE"}. This record does not assert independent reproduction.

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