Does this model still support visual VQA? Has its comprehension ability decreased compared to the original Qwen2-VL?
· Sign up or log in to comment
