I have use a same prompt both in qwen2.5-vl-7b and this model ,but i found ,the vl-7b can easily follow the prompt to generate json format answer ,but QvQ can not and even far away from vl-7b .
· Sign up or log in to comment
