Loss become in the beginning of the training phase.

[ztyxd]

Use s2aNet in configs to train the model on the DOTA dataset. At the beginning of training, the loss of the regression branch is NAN. Do you have any suggestions for this problem?

[nmxian]

我也是跑 s2anet.cfg文件时，loss在第3个epoch变得很大：

[bigFatCatTom]

我也是跑 s2anet.cfg文件时，loss在第3个epoch变得很大：

请问这个问题您解决了么

[soHardToHaveAName]

Use s2aNet in configs to train the model on the DOTA dataset. At the beginning of training, the loss of the regression branch is NAN. Do you have any suggestions for this problem?

请问这个问题您解决了吗？

[soHardToHaveAName]

我也是跑 s2anet.cfg文件时，loss在第3个epoch变得很大：

请问这个问题您解决了么

您也遇到这个问题了吗？ 是怎么解决的呢

[KaisHasan]

I faced the same problem, lowering the learning rate solved it.

[Crystal-Zhang97]

@bigFatCatTom @ztyxd I'm currently experiencing this issue as well, how did you solve it?