Speech Emotion Captioning on EMOSpeech (test)
71.95SIMSECap + STMIL + SCCL
Evaluation Results
| Method | Links | |
|---|---|---|
| SECap + STMIL + SCCLEncoder=HuBERT, Decoder=Chinese-LLaMA2-7B2026.01 | 71.95 | |
| SLAM-LLMEncoder=emotion2vec, Decoder=Vicuna-v1.5-7B2026.01 | 71.1 | |
| SECap + STMILEncoder=HuBERT, Decoder=Chinese-LLaMA2-7B2026.01 | 68.75 | |
| SECapEncoder=HuBERT, Decoder=Chinese-LLaMA2-7B2026.01 | 67.29 | |
| BaselineEncoder=HTSAT, Decoder=BART2026.01 | 63.95 |