Text-to-Audio Retrieval on Clotho (evaluation)
22.87R@1OEA-Qwen3B (+Cl)
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| OEA-Qwen3B (+Cl)Model Type=OEA Models, Backbone=Qwen3B, Additional Training=Clotho2026.04 | 22.87 | 49.78 | 63.25 | |
| OEA-Qwen7B (+Cl)Model Type=OEA Models, Backbone=Qwen7B, Additional Training=Clotho2026.04 | 22.53 | 48.8 | 62.7 | |
| OEA-Nemo3B (+Cl)Model Type=OEA Models, Backbone=Nemotron-3B, Additional Training=Clotho2026.04 | 21.57 | 47.16 | 60.36 | |
| OEA-Qwen7BModel Type=OEA Models, Backbone=Qwen7B2026.04 | 19.77 | 44.78 | 57.65 | |
| OEA-Qwen3BModel Type=OEA Models, Backbone=Qwen3B2026.04 | 19.18 | 42.05 | 55.85 | |
| OEA-Nemo3BModel Type=OEA Models, Backbone=Nemotron-3B2026.04 | 19.04 | 40.57 | 54.24 | |
| MGA-CLAPModel Type=CLAP Models2026.04 | 18.78 | 43.79 | 56.63 | |
| M2D-CLAPModel Type=CLAP Models2026.04 | 17.55 | 42.91 | 55.54 | |
| Robust-CLAPModel Type=CLAP Models2026.04 | 14.83 | 37.89 | 50.05 | |
| LAION-CLAPModel Type=CLAP Models2026.04 | 13.34 | 33.49 | 46.22 | |
| Nemotron-3BModel Type=Vanilla LALMs, Retrieval Training=No2026.04 | 7.2 | 21.57 | 30.12 | |
| Qwen2.5-Omni-3BModel Type=Vanilla LALMs, Retrieval Training=No2026.04 | 0.17 | 0.63 | 1.13 | |
| Qwen2.5-Omni-7BModel Type=Vanilla LALMs, Retrieval Training=No2026.04 | 0.1 | 0.65 | 1.32 |