FocusCLIP

Reported on 2 benchmarks across 2 tasks · 1 paper · 2 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Robots1 result

Activity RecognitiononStanford40
Top-3 Accuracy (%)· 2024-03-11
10.47
SOTA
Human Pose Descriptions and Subject-Focused Attention for Improved Zero-Shot Transfer in Human-Centric Classification Tasks arXiv:2403.06904

Audio1 result

Emotion RecognitiononEMOTIC
Top-3 Accuracy (%)· 2024-03-11
13.73
best: 14.73 (CAGE)
SOTA
Human Pose Descriptions and Subject-Focused Attention for Improved Zero-Shot Transfer in Human-Centric Classification Tasks arXiv:2403.06904