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

Compositional Evaluation on SugarCrepe (test)

95.52Replace (Object)

DataComp (xlarge)

55.355265.782676.2186.6374May 14, 2024Jun 12, 2024Jul 11, 2024Aug 9, 2024Sep 7, 2024Oct 6, 2024Nov 4, 2024
Updated 1mo ago

Evaluation Results

MethodLinks
2024.11
95.52---84.5269.9965.0466.8291.0384.97---79.7
2024.11
92.68---79.8263.9456.157.6684.3478.61---73.31
2024.05
86.86---73.4760.2459.3563.3673.3366.7673.1868.4260.24-
2024.05
86.62---75.563.2252.4463.0671.6566.7671.5768.3963.22-
2024.05
86.13---75.1364.6566.6763.3674.9271.2475.9169.9164.65-
2024.05
86.07---7562.2360.5759.1674.6868.6472.8263.4762.23-
2024.05
85.77---79.1864.5161.7858.7174.2468.3573.7168.7564.51-
2024.05
84.5---76.963.0962.261.8671.7765.4673.6667.663.09-
2024.05
84.07---75.886660.1661.5671.6763.1571.9766.8666-
2024.11
83.66---81.2279.0264.4963.6673.6775.43---74.45
2024.11
82.99---78.6875.7561.6365.4770.0876.01---72.94
2024.11
77.84---69.2963.2366.5362.3167.1769.65---68
2024.11
77---69.5457.6857.7257.0666.7364.88---64.37
2024.11
75.06---65.4858.6853.4757.6667.6566.76---63.54
2024.11
69.92---69.0364.7256.3357.9662.6163.87---63.49
2024.11
64.77---66.1265.9355.5155.4159.6564.45---61.69
2024.11
63.44---55.9650.7150.648.5756.9851.16---53.92
2024.11
62.71---58.1254.4856.3351.256.2661.13---57.18
2024.11
59.44---53.1751.4254.6949.2555.2955.35---54.09
2024.11
56.9---56.8551.9950.815053.9360.55---54.43
2024.04
-636257----------
2024.04
-646459----------
2024.04
-626364----------
2024.04
-666662----------
2025.02
-738062----------
2025.02
-878575----------
2025.02
-818274----------
2025.02
-838068----------