Thank you for open-sourcing this great work. While reading the paper, I found out a small typo in Table 3.
In table 1, Qwen2.5-VL-7B shows 31.6 in the Orientation task, but in table 3, first row, it says 21.6 in Orientation task. Comparing with the number below (Full RL@think), it say 31.5(-0.1), which clearly indicates that the correct number should be 31.6
Just kindly pointing out a very small typo that might requires your attention and correction.
Thank you for open-sourcing this great work. While reading the paper, I found out a small typo in Table 3.
In table 1, Qwen2.5-VL-7B shows 31.6 in the Orientation task, but in table 3, first row, it says 21.6 in Orientation task. Comparing with the number below (Full RL@think), it say 31.5(-0.1), which clearly indicates that the correct number should be 31.6
Just kindly pointing out a very small typo that might requires your attention and correction.