Text-to-text retrieval on AudioCaps
50.3Recall@1OEA-Qwen7B
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| OEA-Qwen7BModel Category=OEA Models, Contrastive Learning (+Cl)=false2026.04 | 50.3 | 72.31 | 82.05 | |
| OEA-Qwen7BModel Category=OEA Models, Contrastive Learning (+Cl)=true2026.04 | 49.93 | 71.79 | 81.05 | |
| OEA-Qwen3BModel Category=OEA Models, Contrastive Learning (+Cl)=true2026.04 | 49.64 | 71.94 | 81.01 | |
| OEA-Qwen3BModel Category=OEA Models, Contrastive Learning (+Cl)=false2026.04 | 49.42 | 72.27 | 81.76 | |
| OEA-Nemo3BModel Category=OEA Models, Contrastive Learning (+Cl)=false2026.04 | 48.94 | 72.62 | 82.19 | |
| OEA-Nemo3BModel Category=OEA Models, Contrastive Learning (+Cl)=true2026.04 | 48.94 | 72.21 | 81.64 | |
| MGA-CLAPModel Category=CLAP Models2026.04 | 47.73 | 71.63 | 80.96 | |
| M2D-CLAPModel Category=CLAP Models2026.04 | 47.51 | 70.03 | 79.96 | |
| LAION-CLAPModel Category=CLAP Models2026.04 | 41.11 | 63.94 | 74.19 | |
| Robust-CLAPModel Category=CLAP Models2026.04 | 41.09 | 63.96 | 74.22 | |
| Nemotron-3BModel Category=Vanilla LALMs, Retrieval Training=false2026.04 | 33.17 | 52.27 | 62.13 | |
| Qwen2.5-Omni-7BModel Category=Vanilla LALMs, Retrieval Training=false2026.04 | 24.06 | 37.19 | 44.62 | |
| Qwen2.5-Omni-3BModel Category=Vanilla LALMs, Retrieval Training=false2026.04 | 22.63 | 34.48 | 41.39 |