List your company

Emotion Recognition in Conversation on MELD (Per-Class Metrics)

67.29F1 ScoreVL2E
34.717243.173651.6360.0864Oct 24, 2024

Evaluation results

MethodLinks
VL2EModality=T+V2024.1067.2980.3161.0314.7146.2865.3328.5755.07——
TelME*Modality=T+V, Note=Reproduced version2024.1066.8379.8559.6131.1743.3565.2923.5355——
M2FNetModality=T+V+A2024.1066.71—————————
VL2E - w/o neutral normalizationModality=T+V2024.1066.6—————————
FacialMMTModality=T+V+A2024.1066.5880.1359.6319.1841.9964.8818.1856——
VL2E - w/o face extractionModality=T+V2024.1065.95—————————
VL2E - w/o visionModality=T2024.1065.56—————————
UniMSEModality=T+V+A2024.1065.51—————————
EmoCapsModality=T+V+A2024.106477.1263.193.0342.5257.57.6957.54——
DAG-ERCModality=T2024.1063.65—————————
MM-DFNModality=T+V+A2024.1059.4677.7650.69—22.9454.78—47.82——
ConGCNModality=T+A2024.1059.476.750.38.728.553.110.646.8——
GA2MIFModality=T+V+A2024.1058.9476.9249.08—27.1851.87—48.52——
MMGCNModality=T+V+A2024.1058.65—————————
DialogueRNNModality=T+V+A2024.1057.0373.549.41.223.850.71.741.5——
VL2E - w/o textModality=V2024.1035.97—————————
Audio Flamingo 3Size=7B2026.03—————————58.5
Audio-ReasonerSize=8.4B2026.03—————————63.2
BC-LSTM2022.03—73.847.75.425.151.35.238.455.9—
DAG-ERC2022.03————————63.65—
DialogueCRN2022.03————————58.39—
DialogueGCN2022.03————————58.23—
DialogueRNN2022.03—73.549.41.223.850.71.741.557.03—
DialogXL2022.03————————62.41—
EmoCaps2022.03—77.1263.193.0342.5257.57.6957.5464—
Gemini 2.5 FlashSize=-2026.03—————————61.5
GPT-4o AudioSize=-2026.03—————————62.5
MMGCN2022.03————————58.65—
Qwen2.5-omniSize=7B2026.03—————————60.3
Qwen2.5-omni + Only STASize=7B2026.03—————————62.8
Qwen2.5-omni + Reasoning SFTSize=7B2026.03—————————61.8
SALMONNSize=7B2026.03—————————37.2
TGSRSize=7B2026.03—————————64.6