1 paper
Penghao Yin, Haomin Wang, Qihong Tang +3
Although visual reasoning is crucial for solving complex geometry tasks, existing vision-language models rely heavily on text-only reasoning. Some recent methods introduce intermed…