Lv Yang

@loucif01 @tmbdev I'm sorry to bother you. have you get an answer? I met the same question when I try to adapt SegFormer for a binary semantic segmentation problem. When...

Should I increase the learning rate ? Now it's 5e-6. 

And should I modify the value of minimum version of mmcv in code?

@WZMIAOMIAO 感谢回复！

HI Have you solve the problem, I met the same trouble on 3090! DeepLabv3+（mobilenetV2）is faster than the data in paper. And I use the mmseg in the same way.

This is my test code and result! 

@bubbliiiing 网络结构输出结果中，我也没发现问题

 我在Segformer论文里看到它应该在15M左右，所以才有此疑惑，您认为我需要从那个地方排查一下呢？

  以上为报错信息

谢谢谢谢导师回复