Conversational Speech Emotion Recognition on Multidialog
57.54WF1Qwen2-Audio + Titans+LoRA
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Qwen2-Audio + Titans+LoRABackbone=Qwen2-Audio, Adapter=Titans+LoRA2026.06 | 57.54 | 37.94 | |
| Qwen2-Audio + LoRABackbone=Qwen2-Audio, Adapter=LoRA2026.06 | 56.37 | 37.16 | |
| Audio Flamingo 3 + Titans+LoRABackbone=Audio Flamingo 3, Adapter=Titans+LoRA2026.06 | 56.07 | 35.82 | |
| Audio Flamingo 3 + LoRABackbone=Audio Flamingo 3, Adapter=LoRA2026.06 | 55.11 | 34.78 | |
| Ultravox-v0.4 + Titans+LoRABackbone=Ultravox-v0.4, Adapter=Titans+LoRA2026.06 | 55.04 | 33.32 | |
| Ultravox-v0.4 + LoRABackbone=Ultravox-v0.4, Adapter=LoRA2026.06 | 54.85 | 30.74 | |
| Ultravox-v0.4Backbone=Ultravox-v0.4, Adapter=None2026.06 | 34.24 | 19.78 | |
| Audio Flamingo 3Backbone=Audio Flamingo 3, Adapter=None2026.06 | 31.64 | 21.08 | |
| Qwen2-AudioBackbone=Qwen2-Audio, Adapter=None2026.06 | 28.27 | 17.65 |