Speech Emotion Recognition on SAVEE
99.47WAStudent
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Studentdepth=3, num head=5, FLOPs=0.32G2024.03 | 99.47 | — | — | |
| Studentdepth=3, num head=5, FLOPs=0.32G2024.03 | 98.95 | — | — | |
| Teacherdepth=12, num head=12, FLOPs=3.58G2024.03 | 98.17 | — | — | |
| Teacherdepth=6, num head=5, FLOPs=1.43G2024.03 | 97.39 | — | — | |
| PL-DistillType=Student Models2026.02 | 92.5 | 91.43 | 91.36 | |
| Gong et al.depth=12, num head=12, Pre-train=ImageNet, FLOPs=3.32G2024.03 | 89.32 | — | — | |
| LLaVA-KDType=Student Models2026.02 | 87.5 | 85.71 | 85.55 | |
| Ristea et al.depth=6, num head=5, FLOPs=175.81G2024.03 | 87.23 | — | — | |
| Forward KLType=Student Models2026.02 | 85 | 83.7 | 83.07 | |
| Reverse KLType=Student Models2026.02 | 81.67 | 79.05 | 77.7 | |
| SFTType=Student Models2026.02 | 80.83 | 78.1 | 76.7 | |
| data2vec 2.0 largeType=Pretrained Models2026.02 | 78.59 | 75.75 | 78.24 | |
| WavLM largeType=Pretrained Models2026.02 | 78.25 | 75.65 | 78.38 | |
| Whisper large v3Type=Pretrained Models2026.02 | 77.24 | 74.07 | 75.31 | |
| Qwen2-AudioType=Teacher Model2026.02 | 58.33 | 61.43 | 53.92 | |
| RQA+IS10Model=LR2018.11 | 54 | 53.8 | — | |
| RQA+IS10Model=SVM2018.11 | 52.5 | 50.6 | — | |
| LLDs StatsModel=ESR2018.11 | 51.5 | 49.3 | — | |
| WSFHM+IS10Model=SVM2018.11 | 50 | — | — | |
| IS10Model=LR2018.11 | 48.5 | 43.1 | — | |
| RQAModel=LR2018.11 | 47.7 | 42.3 | — | |
| IS10Model=SVM2018.11 | 47.5 | 45.6 | — | |
| RQAModel=SVM2018.11 | 45.6 | 41.1 | — |