Speech Emotion Recognition on ASVP-ESD Mixlingual
0.7136Weighted F1VowelPrompt
Evaluation results
| Method | Links | |
|---|---|---|
| VowelPromptBackbone=Qwen2-7B-Instruct, Protocol=SFT & GRPO2026.02 | 0.7136 | |
| VowelPromptBackbone=Qwen2-7B-Instruct, Protocol=SFT2026.02 | 0.7054 | |
| SpeechCueLLMBackbone=Qwen2-7B-Instruct, Protocol=SFT & GRPO2026.02 | 0.6812 | |
| InstructERCBackbone=Qwen2-7B-Instruct, Protocol=SFT & GRPO2026.02 | 0.6796 | |
| SALMONNBackbone=Qwen2-7B-Instruct, Protocol=SFT & GRPO2026.02 | 0.6785 | |
| SpeechCueLLMBackbone=Qwen2-7B-Instruct, Protocol=SFT2026.02 | 0.6785 | |
| InstructERCBackbone=Qwen2-7B-Instruct, Protocol=SFT2026.02 | 0.6725 | |
| SALMONNBackbone=Qwen2-7B-Instruct, Protocol=SFT2026.02 | 0.671 |