# Audio-visual Zero-Shot Classification on VGGSound GZSL (test)

Source: [Audio-visual Zero-Shot Classification on VGGSound GZSL (test)](<https://www.sota2.com/research/sota/audio-visual-zero-shot-classification-on-vggsound-gzsl-test>)

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

Data updated: 2026-02-27T09:10:06.994Z

Evaluation of Audio-visual Zero-Shot Classification models on the test split of the VGGSound dataset configured for Generalized Zero-Shot Learning (GZSL). Metrics typically include S, U, HM, and ZSL 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 | Audio-visual Zero-Shot Classification |
| Dataset (including any stated split) | VGGSound GZSL (test) |
| Metric | S Score |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 38 of 38 |
| 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

- [S Score](<https://www.sota2.com/research/sota/audio-visual-zero-shot-classification-on-vggsound-gzsl-test.md?metric=S%20Score>): Higher is better; unit: PERCENT.

- [U Score](<https://www.sota2.com/research/sota/audio-visual-zero-shot-classification-on-vggsound-gzsl-test.md?metric=U%20Score>): Higher is better; unit: PERCENT.

- [HM Score](<https://www.sota2.com/research/sota/audio-visual-zero-shot-classification-on-vggsound-gzsl-test.md?metric=HM%20Score>): Higher is better; unit: PERCENT.

- [ZSL Score](<https://www.sota2.com/research/sota/audio-visual-zero-shot-classification-on-vggsound-gzsl-test.md?metric=ZSL%20Score>): Higher is better; unit: PERCENT.

- [Accuracy](<https://www.sota2.com/research/sota/audio-visual-zero-shot-classification-on-vggsound-gzsl-test.md?metric=Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | S Score | Result date | Report paper |
| --- | --- | --- | --- |
| DEVISE | 29.96 | Not recorded | 68fa6c341e209131cea32cb4 |
| DEVISE | 29.96 | Not recorded | 68fa72f61e209131cea59938 |
| ALE | 26.13 | Not recorded | 68fa6c341e209131cea32cb4 |
| AVGZSLNet | 18.05 | Not recorded | 68fa71791e209131cea5176b |
| SJE | 16.94 | Not recorded | 68fa6c341e209131cea32cb4 |
| SJE | 16.94 | Not recorded | 68fa72f61e209131cea59938 |
| Hyper-multiple | 15.62 | Not recorded | 68fa71791e209131cea5176b |
| Hyper-multiple | 15.62 | Not recorded | 68fa72f61e209131cea59938 |
| AVGZSLNet | 15.02 | Not recorded | 68fa6d191e209131cea37ba4 |
| AVGZSLNet | 15.02 | Not recorded | 68fa71791e209131cea5176b |
| Hyper-multiple | 15.02 | Not recorded | 68fa71791e209131cea5176b |
| AVGZSLNet | 15.02 | Not recorded | 68fa6c341e209131cea32cb4 |
| AVGZSLNet | 15.02 | Not recorded | 68fa72f61e209131cea59938 |
| AVCA | 14.9 | Not recorded | 68fa71791e209131cea5176b |
| KDA | 14.64 | Not recorded | 68fa72f61e209131cea59938 |
| Attention Fusion | 14.13 | Not recorded | 68fa6d191e209131cea37ba4 |
| Perceiver | 13.25 | Not recorded | 68fa6d191e209131cea37ba4 |
| Hyper-alignment | 13.22 | Not recorded | 68fa71791e209131cea5176b |
| ACFS | 12.87 | Not recorded | 68fa72f61e209131cea59938 |
| AVCA | 12.63 | Not recorded | 68fa6d191e209131cea37ba4 |
| TCAF | 12.63 | Not recorded | 68fa6d191e209131cea37ba4 |
| TCAF | 12.63 | Not recorded | 68fa6d191e209131cea37ba4 |
| AVCA | 12.63 | Not recorded | 68fa71791e209131cea5176b |
| AVCA | 12.63 | Not recorded | 68fa6c341e209131cea32cb4 |
| AVCA | 12.63 | Not recorded | 68fa72f61e209131cea59938 |
| TCaF | 12.63 | Not recorded | 68fa72f61e209131cea59938 |
| Hyper-single | 12.56 | Not recorded | 68fa71791e209131cea5176b |
| Hyper-alignment | 12.5 | Not recorded | 68fa71791e209131cea5176b |
| CJME | 10.86 | Not recorded | 68fa6d191e209131cea37ba4 |
| CJME | 10.86 | Not recorded | 68fa71791e209131cea5176b |
| CJME | 10.86 | Not recorded | 68fa6c341e209131cea32cb4 |
| CJME | 10.86 | Not recorded | 68fa72f61e209131cea59938 |
| Hyper-single | 9.79 | Not recorded | 68fa71791e209131cea5176b |
| CJME | 8.69 | Not recorded | 68fa71791e209131cea5176b |
| APN | 6.46 | Not recorded | 68fa6c341e209131cea32cb4 |
| APN | 6.46 | Not recorded | 68fa72f61e209131cea59938 |
| TCAF-audio | 5.11 | Not recorded | 68fa6d191e209131cea37ba4 |
| TCAF-visual | 3.97 | Not recorded | 68fa6d191e209131cea37ba4 |

### DEVISE

S Score: 29.96. Result ID: 69a1437e35dfcabb00e5deba.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c341e209131cea32cb4.md>)

Recorded evaluation configuration: {"Method type":"ZSL"}. This record does not assert independent reproduction.

### DEVISE

S Score: 29.96. Result ID: 69a15f6935dfcabb00e69d8c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72f61e209131cea59938.md>)

Recorded evaluation configuration: {"Venue":"NeurIPS’13"}. This record does not assert independent reproduction.

### ALE

S Score: 26.13. Result ID: 69a1437e35dfcabb00e5deb8.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c341e209131cea32cb4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa62d31e209131ce9fb5fb.md>)

[Implementation](<https://code.google.com/p/word2vec/>)

Recorded evaluation configuration: {"Method type":"ZSL"}. This record does not assert independent reproduction.

### AVGZSLNet

S Score: 18.05. Result ID: 6991d0bd3a6c3148490ac1bc.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71791e209131cea5176b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa68c11e209131cea1e112.md>)

[Implementation](<https://github.com/krantiparida/AudioSetZSL>)

Recorded evaluation configuration: {"Feature type":"main"}. This record does not assert independent reproduction.

### SJE

S Score: 16.94. Result ID: 69a1437e35dfcabb00e5deb9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c341e209131cea32cb4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa62641e209131ce9f90f8.md>)

Recorded evaluation configuration: {"Method type":"ZSL"}. This record does not assert independent reproduction.

### SJE

S Score: 16.94. Result ID: 69a15f6935dfcabb00e69d8d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72f61e209131cea59938.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa62641e209131ce9f90f8.md>)

Recorded evaluation configuration: {"Venue":"CVPR’20"}. This record does not assert independent reproduction.

### Hyper-multiple

S Score: 15.62. Result ID: 6991d0bd3a6c3148490ac1ae.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71791e209131cea5176b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71791e209131cea5176b.md>)

Recorded evaluation configuration: {"Feature Type":"cls feature","Number of adaptive curvatures (Ne)":3}. This record does not assert independent reproduction.

### Hyper-multiple

S Score: 15.62. Result ID: 69a15f6935dfcabb00e69d94.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72f61e209131cea59938.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71791e209131cea5176b.md>)

Recorded evaluation configuration: {"Venue":"ICCV’23"}. This record does not assert independent reproduction.

### AVGZSLNet

S Score: 15.02. Result ID: 698f7dd53a6c314849f92f33.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d191e209131cea37ba4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa68c11e209131cea1e112.md>)

[Implementation](<https://github.com/krantiparida/AudioSetZSL>)

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

### AVGZSLNet

S Score: 15.02. Result ID: 6991d0bd3a6c3148490ac1aa.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71791e209131cea5176b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa68c11e209131cea1e112.md>)

[Implementation](<https://github.com/krantiparida/AudioSetZSL>)

Recorded evaluation configuration: {"Feature Type":"cls feature"}. This record does not assert independent reproduction.

### Hyper-multiple

S Score: 15.02. Result ID: 6991d0bd3a6c3148490ac1c0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71791e209131cea5176b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71791e209131cea5176b.md>)

Recorded evaluation configuration: {"Feature type":"main","Number of adaptive curvatures (Ne)":2}. This record does not assert independent reproduction.

### AVGZSLNet

S Score: 15.02. Result ID: 69a1437e35dfcabb00e5debd.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c341e209131cea32cb4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa68c11e209131cea1e112.md>)

[Implementation](<https://github.com/krantiparida/AudioSetZSL>)

Recorded evaluation configuration: {"Method type":"Audio-visual ZSL"}. This record does not assert independent reproduction.

### AVGZSLNet

S Score: 15.02. Result ID: 69a15f6935dfcabb00e69d8f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72f61e209131cea59938.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa68c11e209131cea1e112.md>)

[Implementation](<https://github.com/krantiparida/AudioSetZSL>)

Recorded evaluation configuration: {"Venue":"WACV’21"}. This record does not assert independent reproduction.

### AVCA

S Score: 14.9. Result ID: 6991d0bd3a6c3148490ac1bd.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71791e209131cea5176b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c341e209131cea32cb4.md>)

[Implementation](<https://github.com/ExplainableML/AVCA-GZSL>)

Recorded evaluation configuration: {"Feature type":"main"}. This record does not assert independent reproduction.

### KDA

S Score: 14.64. Result ID: 69a15f6935dfcabb00e69d95.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72f61e209131cea59938.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72f61e209131cea59938.md>)

[Implementation](<https://github.com/chenhaoxing/KDA>)

Recorded evaluation configuration: {"Venue":"Ours"}. This record does not assert independent reproduction.

### Attention Fusion

S Score: 14.13. Result ID: 698f7dd53a6c314849f92f30.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d191e209131cea37ba4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67b11e209131cea17da6.md>)

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

### Perceiver

S Score: 13.25. Result ID: 698f7dd53a6c314849f92f31.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d191e209131cea37ba4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a381e209131cea26a64.md>)

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

### Hyper-alignment

S Score: 13.22. Result ID: 6991d0bd3a6c3148490ac1be.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71791e209131cea5176b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71791e209131cea5176b.md>)

Recorded evaluation configuration: {"Feature type":"main","Curvature (c)":-0.5}. This record does not assert independent reproduction.

### ACFS

S Score: 12.87. Result ID: 69a15f6935dfcabb00e69d93.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72f61e209131cea59938.md>)

Recorded evaluation configuration: {"Venue":"IJCNN’23"}. This record does not assert independent reproduction.

### AVCA

S Score: 12.63. Result ID: 698f7dd53a6c314849f92f34.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d191e209131cea37ba4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c341e209131cea32cb4.md>)

[Implementation](<https://github.com/ExplainableML/AVCA-GZSL>)

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

### TCAF

S Score: 12.63. Result ID: 698f7dd53a6c314849f92f35.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d191e209131cea37ba4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d191e209131cea37ba4.md>)

[Implementation](<https://github.com/ExplainableML/TCAF-GZSL>)

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

### TCAF

S Score: 12.63. Result ID: 698f7dd53a6c314849f92f44.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d191e209131cea37ba4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d191e209131cea37ba4.md>)

[Implementation](<https://github.com/ExplainableML/TCAF-GZSL>)

Recorded evaluation configuration: {"Input features":"audio-visual"}. This record does not assert independent reproduction.

### AVCA

S Score: 12.63. Result ID: 6991d0bd3a6c3148490ac1ab.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71791e209131cea5176b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c341e209131cea32cb4.md>)

[Implementation](<https://github.com/ExplainableML/AVCA-GZSL>)

Recorded evaluation configuration: {"Feature Type":"cls feature"}. This record does not assert independent reproduction.

### AVCA

S Score: 12.63. Result ID: 69a1437e35dfcabb00e5debe.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c341e209131cea32cb4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c341e209131cea32cb4.md>)

[Implementation](<https://github.com/ExplainableML/AVCA-GZSL>)

Recorded evaluation configuration: {"Method type":"Audio-visual ZSL"}. This record does not assert independent reproduction.

### AVCA

S Score: 12.63. Result ID: 69a15f6935dfcabb00e69d91.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72f61e209131cea59938.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c341e209131cea32cb4.md>)

[Implementation](<https://github.com/ExplainableML/AVCA-GZSL>)

Recorded evaluation configuration: {"Venue":"CVPR’22"}. This record does not assert independent reproduction.

### TCaF

S Score: 12.63. Result ID: 69a15f6935dfcabb00e69d92.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72f61e209131cea59938.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d191e209131cea37ba4.md>)

[Implementation](<https://github.com/ExplainableML/TCAF-GZSL>)

Recorded evaluation configuration: {"Venue":"ECCV’22"}. This record does not assert independent reproduction.

### Hyper-single

S Score: 12.56. Result ID: 6991d0bd3a6c3148490ac1ad.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71791e209131cea5176b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71791e209131cea5176b.md>)

Recorded evaluation configuration: {"Feature Type":"cls feature"}. This record does not assert independent reproduction.

### Hyper-alignment

S Score: 12.5. Result ID: 6991d0bd3a6c3148490ac1ac.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71791e209131cea5176b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71791e209131cea5176b.md>)

Recorded evaluation configuration: {"Feature Type":"cls feature","Curvature (c)":-0.1}. This record does not assert independent reproduction.

### CJME

S Score: 10.86. Result ID: 698f7dd53a6c314849f92f32.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d191e209131cea37ba4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66931e209131cea1106d.md>)

[Implementation](<https://www.cse.iitk.ac.in/users/kranti/avzsl.html>)

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

### CJME

S Score: 10.86. Result ID: 6991d0bd3a6c3148490ac1a9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71791e209131cea5176b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66931e209131cea1106d.md>)

[Implementation](<https://www.cse.iitk.ac.in/users/kranti/avzsl.html>)

Recorded evaluation configuration: {"Feature Type":"cls feature"}. This record does not assert independent reproduction.

### CJME

S Score: 10.86. Result ID: 69a1437e35dfcabb00e5debc.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c341e209131cea32cb4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66931e209131cea1106d.md>)

[Implementation](<https://www.cse.iitk.ac.in/users/kranti/avzsl.html>)

Recorded evaluation configuration: {"Method type":"Audio-visual ZSL"}. This record does not assert independent reproduction.

### CJME

S Score: 10.86. Result ID: 69a15f6935dfcabb00e69d8e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72f61e209131cea59938.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66931e209131cea1106d.md>)

[Implementation](<https://www.cse.iitk.ac.in/users/kranti/avzsl.html>)

Recorded evaluation configuration: {"Venue":"WACV’20"}. This record does not assert independent reproduction.

### Hyper-single

S Score: 9.79. Result ID: 6991d0bd3a6c3148490ac1bf.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71791e209131cea5176b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71791e209131cea5176b.md>)

Recorded evaluation configuration: {"Feature type":"main"}. This record does not assert independent reproduction.

### CJME

S Score: 8.69. Result ID: 6991d0bd3a6c3148490ac1bb.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71791e209131cea5176b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66931e209131cea1106d.md>)

[Implementation](<https://www.cse.iitk.ac.in/users/kranti/avzsl.html>)

Recorded evaluation configuration: {"Feature type":"main"}. This record does not assert independent reproduction.

### APN

S Score: 6.46. Result ID: 69a1437e35dfcabb00e5debb.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c341e209131cea32cb4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69eb1e209131cea24ee0.md>)

Recorded evaluation configuration: {"Method type":"ZSL"}. This record does not assert independent reproduction.

### APN

S Score: 6.46. Result ID: 69a15f6935dfcabb00e69d90.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72f61e209131cea59938.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c281e209131cea326a1.md>)

Recorded evaluation configuration: {"Venue":"IJCV’22"}. This record does not assert independent reproduction.

### TCAF-audio

S Score: 5.11. Result ID: 698f7dd53a6c314849f92f42.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d191e209131cea37ba4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d191e209131cea37ba4.md>)

[Implementation](<https://github.com/ExplainableML/TCAF-GZSL>)

Recorded evaluation configuration: {"Input features":"audio"}. This record does not assert independent reproduction.

### TCAF-visual

S Score: 3.97. Result ID: 698f7dd53a6c314849f92f43.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d191e209131cea37ba4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d191e209131cea37ba4.md>)

[Implementation](<https://github.com/ExplainableML/TCAF-GZSL>)

Recorded evaluation configuration: {"Input features":"visual"}. This record does not assert independent reproduction.

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