Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

EMOTIC

Benchmarks

Task NameDataset NameSOTA ResultTrend
In-context Human Emotion RecognitionEMOTIC 1.0 (test)
mAP39.13
10
Emotion RecognitionEMOTIC NarraCap captions (val)
Suffering χ²3.28
6
Emotion ClassificationEmotic
Accuracy91.95
6
Emotion Recognition in ContextEMOTIC (test)
mAP35.48
3
Multi-label emotion recognitionEmotic (test)
Hamming Loss11.5
2
Sentiment ClassificationEMOTIC
Accuracy73.8
1
Group-level Multimodal Emotion RecognitionEMOTIC subset
Accuracy Improvement (vs Avg)14
1
Showing 7 of 7 rows