qbench_a2_pair.csv

Model,Completeness,Precision,Relevance,Sum
InfiMM (Zephyr-7B),0.75,0.91,1.61,3.28
Emu2-Chat (LLaMA-33B),0.63,0.87,1.53,3.03
Fuyu-8B (Persimmon-8B),0.7,0.84,1.56,3.12
BakLLava (Mistral-7B),0.8,0.8,1.79,3.4
mPLUG-Owl2 (Q-Instruct),1.02,0.79,1.87,3.69
mPLUG-Owl2 (LLaMA-7B),0.94,0.92,1.63,3.5
LLaVA-v1.5 (Vicuna-v1.5-7B),0.86,0.82,1.53,3.22
LLaVA-v1.5 (Vicuna-v1.5-13B),0.89,0.92,1.63,3.44