Speech Emotion Recognition on RAVDESS 8 classes (test)
84.72Weighted AccuracyPhysioSER
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| PhysioSERBackbone=WavLM2026.02 | 84.72 | 84.71 | 84.02 | 27.76 | |
| PhysioSERBackbone=BEATs2026.02 | 82.99 | 83.05 | 82.03 | 1.7 | |
| PhysioSERBackbone=Emotion2Vec2026.02 | 81.6 | 81.1 | 81.45 | 31.3 | |
| BEATsEvaluation Mode=Backbone-only2026.02 | 81.6 | 78.89 | 79.07 | — | |
| PhysioSERBackbone=CLAP2026.02 | 68.75 | 67.42 | 67.21 | 19.27 | |
| WavLMEvaluation Mode=Backbone-only2026.02 | 66.31 | 64.19 | 64.02 | — | |
| PhysioSERBackbone=HuBERT2026.02 | 65.28 | 65.12 | 64.34 | 10.59 | |
| Emotion2VecEvaluation Mode=Backbone-only2026.02 | 62.15 | 58.1 | 57.57 | — | |
| HuBERTEvaluation Mode=Backbone-only2026.02 | 59.03 | 56.45 | 56.63 | — | |
| CLAPEvaluation Mode=Backbone-only2026.02 | 57.64 | 53.63 | 53.83 | — | |
| PhysioSERBackbone=Wav2vec22026.02 | 48.61 | 47.46 | 47.03 | 9.38 | |
| Wav2vec2Evaluation Mode=Backbone-only2026.02 | 44.44 | 42.48 | 40.16 | — |