Speech Emotion Recognition on CREMA-D 6 classes (test)
75.2Weighted Accuracy (WA)PhysioSER
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| PhysioSERBackbone=WavLM2026.02 | 75.2 | 75.62 | 75.63 | 7.91 | |
| PhysioSERBackbone=BEATs2026.02 | 75.08 | 75.59 | 75.07 | 4.99 | |
| PhysioSERBackbone=Emotion2Vec2026.02 | 72.98 | 73.34 | 72.99 | 5.74 | |
| BEATsEvaluation Mode=Backbone-only2026.02 | 71.51 | 71.78 | 71.67 | — | |
| PhysioSERBackbone=HuBERT2026.02 | 69.89 | 70.33 | 69.51 | 5.69 | |
| WavLMEvaluation Mode=Backbone-only2026.02 | 69.69 | 70.22 | 69.79 | — | |
| Emotion2VecEvaluation Mode=Backbone-only2026.02 | 69.02 | 69.52 | 69.16 | — | |
| HuBERTEvaluation Mode=Backbone-only2026.02 | 66.13 | 66.68 | 66.33 | — | |
| PhysioSERBackbone=Wav2vec22026.02 | 64.38 | 64.84 | 64.13 | 11.81 | |
| PhysioSERBackbone=CLAP2026.02 | 59.81 | 59.84 | 59.35 | 10.96 | |
| Wav2vec2Evaluation Mode=Backbone-only2026.02 | 57.58 | 58.13 | 56.98 | — | |
| CLAPEvaluation Mode=Backbone-only2026.02 | 53.9 | 54.19 | 53.72 | — |