yangzishy

The test cannot calculate ap on its own dataset

12

### Prerequisite - [X] I have searched [Issues](https://github.com/open-mmlab/mmrotate/issues) and [Discussions](https://github.com/open-mmlab/mmrotate/discussions) but cannot get the expected help. - [X] I have read the [FAQ documentation](https://mmrotate.readthedocs.io/en/1.x/notes/4_faq.html) but cannot get the expected help....

module 'torch' has no attribute 'pi'

1

### Prerequisite - [X] I have searched [Issues](https://github.com/open-mmlab/mmrotate/issues) and [Discussions](https://github.com/open-mmlab/mmrotate/discussions) but cannot get the expected help. - [X] I have read the [FAQ documentation](https://mmrotate.readthedocs.io/en/1.x/notes/4_faq.html) but cannot get the expected help....

> 你这么说提醒了我，这个margin我一直使用默认值没有改动过。另外，一开始的学习率可以大一些。代码默认0.001是我后期使用的数值。你可以参考下作者在论文里使用过的数值。

1

> 你这么说提醒了我，这个margin我一直使用默认值没有改动过。另外，一开始的学习率可以大一些。代码默认0.001是我后期使用的数值。你可以参考下作者在论文里使用过的数值。 > > 如果还是不收敛，换到resnet分支试一下。我发现ResNet收敛比HRNet更快一点。 改了学习率也不收敛 而且准确率accuracy稳定在百分之二十多 不会提升了 这是什么情况 _Originally posted by @Yangzishy in https://github.com/yinguobing/arcface/issues/2#issuecomment-2259537511_

2024-07-25 08:59:42.137204: I T:\src\github\tensorflow\tensorflow\core\platform\cpu_feature_guard.cc:140] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 重载模型训练 epoch:1/150 Step: 1/725400, accuracy: 0.000000, center loss: 0.998900, cross loss: 8.594902,...