# Contrastive Text-Audio Retrieval on CompA

Source: [Contrastive Text-Audio Retrieval on CompA](<https://www.sota2.com/research/sota/contrastive-text-audio-retrieval-on-compa>)

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

Data updated: 2026-06-17T06:52:03.227Z

Evaluation of Contrastive Text-Audio Retrieval performance on the CompA dataset, measured by attribute and order accuracy for both text and audio retrieval components.

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 | Contrastive Text-Audio Retrieval |
| Dataset (including any stated split) | CompA |
| Metric | Attribute Accuracy (Text) |
| 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

- [Attribute Accuracy (Text)](<https://www.sota2.com/research/sota/contrastive-text-audio-retrieval-on-compa.md?metric=Attribute%20Accuracy%20(Text)>): Higher is better; unit: PERCENT.

- [Attribute Accuracy (Audio)](<https://www.sota2.com/research/sota/contrastive-text-audio-retrieval-on-compa.md?metric=Attribute%20Accuracy%20(Audio)>): Higher is better; unit: PERCENT.

- [Order Accuracy (Text)](<https://www.sota2.com/research/sota/contrastive-text-audio-retrieval-on-compa.md?metric=Order%20Accuracy%20(Text)>): Higher is better; unit: PERCENT.

- [Order Accuracy (Audio)](<https://www.sota2.com/research/sota/contrastive-text-audio-retrieval-on-compa.md?metric=Order%20Accuracy%20(Audio)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Attribute Accuracy (Text) | Result date | Report paper |
| --- | --- | --- | --- |
| CLAPScoreMS | 23.9 | Not recorded | 6a322a36d33dfb23e783ec2d |
| ELSA | 23.4 | Not recorded | 6a322a36d33dfb23e783ec2d |
| CLAPScoreLAION | 19.3 | Not recorded | 6a322a36d33dfb23e783ec2d |
| CLAPScoreHuman | 17.3 | Not recorded | 6a322a36d33dfb23e783ec2d |

### CLAPScoreMS

Attribute Accuracy (Text): 23.9. Result ID: 6a3243fafbc373592c2f305e.

[Reporting paper](<https://www.sota2.com/research/paper/6a322a36d33dfb23e783ec2d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6cb71e209131cea35874.md>)

Recorded evaluation configuration: {"CLAP model":"MS-CLAP"}. This record does not assert independent reproduction.

### ELSA

Attribute Accuracy (Text): 23.4. Result ID: 6a3243fafbc373592c2f3061.

[Reporting paper](<https://www.sota2.com/research/paper/6a322a36d33dfb23e783ec2d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a322a36d33dfb23e783ec2d.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### CLAPScoreLAION

Attribute Accuracy (Text): 19.3. Result ID: 6a3243fafbc373592c2f305f.

[Reporting paper](<https://www.sota2.com/research/paper/6a322a36d33dfb23e783ec2d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa729e1e209131cea57959.md>)

Recorded evaluation configuration: {"CLAP model":"LAION-CLAP"}. This record does not assert independent reproduction.

### CLAPScoreHuman

Attribute Accuracy (Text): 17.3. Result ID: 6a3243fafbc373592c2f3060.

[Reporting paper](<https://www.sota2.com/research/paper/6a322a36d33dfb23e783ec2d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79fe1e209131cea8088b.md>)

Recorded evaluation configuration: {"CLAP model":"Human-CLAP"}. This record does not assert independent reproduction.

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