Is the ltrb calculation inconsistent with the paper?

[saojianjian]

I found that the calculation of ltrb in the code was inconsistent with that in the paper, line 102 and line 104 of the model/loss.py file What ltrb means in the paper is I think the correct calculation code should be

Please let me know if there is an error in my understanding. Thank you!

[VectXmy]

你可以试着按照你那样的计算方式算一下看看，b_off和t_off会变成负值

[saojianjian]

你可以试着按照你那样的计算方式算一下看看，b_off和t_off会变成负值

想请教一下，你在复现代码的时候有看论文作者在他的代码里是这么实现的吗？ 我看你DetectHead里的_coords2boxes是按照你GenTargets里_gen_level_targets函数的ltrb的实现方式解析回原图坐标的，逻辑上没什么问题。如果要改代码的话，得把这一处也给改了

[VectXmy]

不懂你想说啥