Wang Liu

This is my inplemetation of mobilenetv3 modified from this [repository.](https://github.com/xiaolai-sqlai/mobilenetv3.git). And the pytorch pretrained model can be download [here](https://drive.google.com/drive/folders/1-8uUpSiqHVgVNx25UD7OE-WQZu-7_EHQ?usp=sharing) import torch import torch.nn as nn import torch.nn.functional as F from...

> This is my inplemetation of mobilenetv3 modified from this [repository.](https://github.com/xiaolai-sqlai/mobilenetv3.git). And the pytorch pretrained model can be download [here](https://drive.google.com/drive/folders/1-8uUpSiqHVgVNx25UD7OE-WQZu-7_EHQ?usp=sharing) > > import torch import torch.nn as nn import torch.nn.functional...

数据集图片文件的后缀要与SplitOnlyImage_multi_process.splitbase类的参数‘ext’保持一致。 

这一步不能省略，不然后面使用mmdet就会出问题：ImportError: cannot import name 'deform_conv_cuda' from 'mmdet.ops.dcn' 。 解决办法：按照install.md中的软件版本进行下载，比如pytorch版本降到1.1，mmcv降到0.2.16，gcc版本降到5.5。 我用的cuda10.1，torch版本比较低的时候cuda10.1和10.0没什么区别。

> 请问，谁有python版的mmdet呢？谢谢 什么意思，不是直接pip install mmdet吗。况且AerialDetection项目里自带了mmdet，不用另外安装吧？

等他跑完就好了~

+1

> How to make sure the best parameter of the network on cityscapes. And the the best result we can achieve with this network on cityscapes. Is there someone can...