benchmark,samples,scored_samples,metric,score_percent vstar,191,191,accuracy,83.25 zoombench,845,844,valid_subset_accuracy,58.29 hrbench-4k,800,800,accuracy,74.62 hrbench-8k,800,800,accuracy,72.25 mmvp,300,300,accuracy,78.00 cv-bench,2638,2638,mean_2D_3D_accuracy,74.77 mmstar,1500,1500,accuracy,62.20 pope,9000,9000,accuracy,89.03 mmvp,150,150,paired_accuracy,57.33 pope,9000,9000,f1,88.78