# Categorical Visual Question Answering on VizWiz

Source: [Categorical Visual Question Answering on VizWiz](<https://www.sota2.com/research/sota/categorical-visual-question-answering-on-vizwiz>)

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

Data updated: 2026-03-12T09:14:11.390Z

Evaluation of model performance on the VizWiz dataset specifically for categorical visual question answering tasks, reporting Accuracy and F1 scores.

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 | Categorical Visual Question Answering |
| Dataset (including any stated split) | VizWiz |
| 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/categorical-visual-question-answering-on-vizwiz.md?metric=Accuracy>): Higher is better; unit: PERCENT.

- [F1 Score](<https://www.sota2.com/research/sota/categorical-visual-question-answering-on-vizwiz.md?metric=F1%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| SN | 81 | Not recorded | 69b248b6d030084c22e2fb15 |
| Qwen-VL-4b-Instruct | 78.3 | Not recorded | 69b248b6d030084c22e2fb15 |
| SN | 71.9 | Not recorded | 69b248b6d030084c22e2fb15 |
| LLaVA-v1.5-7b | 64.8 | Not recorded | 69b248b6d030084c22e2fb15 |

### SN

Accuracy: 81. Result ID: 69b282f06dbb7075b95d5cdf.

[Reporting paper](<https://www.sota2.com/research/paper/69b248b6d030084c22e2fb15.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69b248b6d030084c22e2fb15.md>)

Recorded evaluation configuration: {"Base Model":"Qwen-VL-4b-Instruct","Probing Set Size":"942","Configuration":"Best probed"}. This record does not assert independent reproduction.

### Qwen-VL-4b-Instruct

Accuracy: 78.3. Result ID: 69b282f06dbb7075b95d5cde.

[Reporting paper](<https://www.sota2.com/research/paper/69b248b6d030084c22e2fb15.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6927db54d6b0dd3342930f66.md>)

Recorded evaluation configuration: {"Probing Set Size":"942"}. This record does not assert independent reproduction.

### SN

Accuracy: 71.9. Result ID: 69b282f06dbb7075b95d5cdd.

[Reporting paper](<https://www.sota2.com/research/paper/69b248b6d030084c22e2fb15.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69b248b6d030084c22e2fb15.md>)

Recorded evaluation configuration: {"Base Model":"LLaVA-v1.5-7b","Probing Set Size":"942","Configuration":"Best probed"}. This record does not assert independent reproduction.

### LLaVA-v1.5-7b

Accuracy: 64.8. Result ID: 69b282f06dbb7075b95d5cdc.

[Reporting paper](<https://www.sota2.com/research/paper/69b248b6d030084c22e2fb15.md>)

Recorded evaluation configuration: {"Probing Set Size":"942"}. This record does not assert independent reproduction.

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