# Visual Question Answering on TextVQA Clean (test)

Source: [Visual Question Answering on TextVQA Clean (test)](<https://www.sota2.com/research/sota/visual-question-answering-on-textvqa-clean-test>)

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

Data updated: 2026-03-31T08:00:16.129Z

Evaluation of model performance on the test split of the TextVQA Clean dataset for visual question answering, typically measured by 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 | Visual Question Answering |
| Dataset (including any stated split) | TextVQA Clean (test) |
| Metric | VQA Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 10 of 10 |
| 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

- [VQA Accuracy](<https://www.sota2.com/research/sota/visual-question-answering-on-textvqa-clean-test.md?metric=VQA%20Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | VQA Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| CLIP | 37.1 | Not recorded | 69cb54e10636e5bf4afc9cff |
| CLIP | 37.1 | Not recorded | 69cb54e10636e5bf4afc9cff |
| FARE | 31.9 | Not recorded | 69cb54e10636e5bf4afc9cff |
| FARE | 31.9 | Not recorded | 69cb54e10636e5bf4afc9cff |
| FARE | 27.6 | Not recorded | 69cb54e10636e5bf4afc9cff |
| FARE | 27.4 | Not recorded | 69cb54e10636e5bf4afc9cff |
| TeCoA | 24.2 | Not recorded | 69cb54e10636e5bf4afc9cff |
| TeCoA | 24.1 | Not recorded | 69cb54e10636e5bf4afc9cff |
| TeCoA | 21 | Not recorded | 69cb54e10636e5bf4afc9cff |
| TeCoA | 20.7 | Not recorded | 69cb54e10636e5bf4afc9cff |

### CLIP

VQA Accuracy: 37.1. Result ID: 69cb7e7b47f34ce924b39a58.

[Reporting paper](<https://www.sota2.com/research/paper/69cb54e10636e5bf4afc9cff.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69cb54e10636e5bf4afc9cff.md>)

Recorded evaluation configuration: {"ET3 Defense":"false"}. This record does not assert independent reproduction.

### CLIP

VQA Accuracy: 37.1. Result ID: 69cb7e7b47f34ce924b39a59.

[Reporting paper](<https://www.sota2.com/research/paper/69cb54e10636e5bf4afc9cff.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69cb54e10636e5bf4afc9cff.md>)

Recorded evaluation configuration: {"ET3 Defense":"true"}. This record does not assert independent reproduction.

### FARE

VQA Accuracy: 31.9. Result ID: 69cb7e7b47f34ce924b39a5c.

[Reporting paper](<https://www.sota2.com/research/paper/69cb54e10636e5bf4afc9cff.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69cb54e10636e5bf4afc9cff.md>)

Recorded evaluation configuration: {"Adversarial Training Strength (epsilon\_t)":"2/255","ET3 Defense":"false"}. This record does not assert independent reproduction.

### FARE

VQA Accuracy: 31.9. Result ID: 69cb7e7b47f34ce924b39a5d.

[Reporting paper](<https://www.sota2.com/research/paper/69cb54e10636e5bf4afc9cff.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69cb54e10636e5bf4afc9cff.md>)

Recorded evaluation configuration: {"Adversarial Training Strength (epsilon\_t)":"2/255","ET3 Defense":"true"}. This record does not assert independent reproduction.

### FARE

VQA Accuracy: 27.6. Result ID: 69cb7e7b47f34ce924b39a60.

[Reporting paper](<https://www.sota2.com/research/paper/69cb54e10636e5bf4afc9cff.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69cb54e10636e5bf4afc9cff.md>)

Recorded evaluation configuration: {"Adversarial Training Strength (epsilon\_t)":"4/255","ET3 Defense":"false"}. This record does not assert independent reproduction.

### FARE

VQA Accuracy: 27.4. Result ID: 69cb7e7b47f34ce924b39a61.

[Reporting paper](<https://www.sota2.com/research/paper/69cb54e10636e5bf4afc9cff.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69cb54e10636e5bf4afc9cff.md>)

Recorded evaluation configuration: {"Adversarial Training Strength (epsilon\_t)":"4/255","ET3 Defense":"true"}. This record does not assert independent reproduction.

### TeCoA

VQA Accuracy: 24.2. Result ID: 69cb7e7b47f34ce924b39a5b.

[Reporting paper](<https://www.sota2.com/research/paper/69cb54e10636e5bf4afc9cff.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69cb54e10636e5bf4afc9cff.md>)

Recorded evaluation configuration: {"Adversarial Training Strength (epsilon\_t)":"2/255","ET3 Defense":"true"}. This record does not assert independent reproduction.

### TeCoA

VQA Accuracy: 24.1. Result ID: 69cb7e7b47f34ce924b39a5a.

[Reporting paper](<https://www.sota2.com/research/paper/69cb54e10636e5bf4afc9cff.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69cb54e10636e5bf4afc9cff.md>)

Recorded evaluation configuration: {"Adversarial Training Strength (epsilon\_t)":"2/255","ET3 Defense":"false"}. This record does not assert independent reproduction.

### TeCoA

VQA Accuracy: 21. Result ID: 69cb7e7b47f34ce924b39a5f.

[Reporting paper](<https://www.sota2.com/research/paper/69cb54e10636e5bf4afc9cff.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69cb54e10636e5bf4afc9cff.md>)

Recorded evaluation configuration: {"Adversarial Training Strength (epsilon\_t)":"4/255","ET3 Defense":"true"}. This record does not assert independent reproduction.

### TeCoA

VQA Accuracy: 20.7. Result ID: 69cb7e7b47f34ce924b39a5e.

[Reporting paper](<https://www.sota2.com/research/paper/69cb54e10636e5bf4afc9cff.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69cb54e10636e5bf4afc9cff.md>)

Recorded evaluation configuration: {"Adversarial Training Strength (epsilon\_t)":"4/255","ET3 Defense":"false"}. This record does not assert independent reproduction.

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