List your company

Text-to-Image Retrieval on Retrieval VQA Visual Genome (test)

66.8Count AccuracyGoldiCLIP
57.23259.71662.264.684Mar 25, 2026

Evaluation results

MethodLinks
GoldiCLIPData size=30M, Zero-shot=true2026.0366.848.551.556.455.955.8
SigLIP 2Data size=10B, Zero-shot=true2026.0366.649.154.756.757.757
PE_coreData size=5B, Zero-shot=true2026.0365.167.448.952.457.552.9
SigLIPData size=10B, Zero-shot=true2026.0364.146.551.153.253.553.5
OpenCLIPData size=2B, Zero-shot=true2026.0362.14547.751.348.550.9
COSMOSData size=30M, Zero-shot=true2026.0360.742.446.151.648.949.9
FLAIRData size=30M, Zero-shot=true2026.0357.639.144.44947.747.6