solidexu
solidexu
Thank you ~ I just found the C++ script from XingangPan. Using slope to determine the existence of lane is definitely a genius idea. But I am not sure if...
I don't know why, the issue is solved by commenting out the relu in RESAReducer. It's too STRANGE for me. 
I use torch 1.10.2. And I have tested with/without mixed precision, same issue.
In current master branch, I download your new branch three days ago in fact. Commenting out the relu also occur another error during training. I think I can try torch...
try to add a 1*1 conv at the top layer of RESA, it may be helpful
我也遇到这个问题,通过这种方式解决了,应该是代码优化问题 https://blog.csdn.net/u012858839/article/details/123850031