jetson安装torchvision的一个报错及解决办法

记一个报错。

别的帖子和Nvidia官网上安装torchvision都是用源码编译的方法，但我死活就是装不上去

 warning: ignoring #pragma omp parallel [-Wunknown-pragmas]
 #pragma omp parallel for if ((end - begin) >= grain_size)
 
In file included from /home/deeplearn/下载/vision-0.8.0/torchvision/csrc/vision.cpp:14:0:
/home/deeplearn/下载/vision-0.8.0/torchvision/csrc/ROIAlign.h: In function ‘at::Tensor ROIAlign_autocast(const at::Tensor&, const at::Tensor&, double, int64_t, int64_t, int64_t, bool)’:
/home/deeplearn/下载/vision-0.8.0/torchvision/csrc/ROIAlign.h:52:28: error: ‘cached_cast’ is not a member of ‘at::autocast’
              at::autocast::cached_cast(at::kFloat, input),
                            ^~~~~~~~~~~
/home/deeplearn/下载/vision-0.8.0/torchvision/csrc/ROIAlign.h:53:28: error: ‘cached_cast’ is not a member of ‘at::autocast’
              at::autocast::cached_cast(at::kFloat, rois),
                            ^~~~~~~~~~~
In file included from /home/deeplearn/下载/vision-0.8.0/torchvision/csrc/vision.cpp:17:0:
/home/deeplearn/下载/vision-0.8.0/torchvision/csrc/nms.h: In function ‘at::Tensor nms_autocast(const at::Tensor&, const at::Tensor&, double)’:
/home/deeplearn/下载/vision-0.8.0/torchvision/csrc/nms.h:31:21: error: ‘cached_cast’ is not a member of ‘at::autocast’
       at::autocast::cached_cast(at::kFloat, dets),
                     ^~~~~~~~~~~
/home/deeplearn/下载/vision-0.8.0/torchvision/csrc/nms.h:32:21: error: ‘cached_cast’ is not a member of ‘at::autocast’
       at::autocast::cached_cast(at::kFloat, scores),
                     ^~~~~~~~~~~
error: command 'aarch64-linux-gnu-gcc' failed with exit status 1

NVIDIA论坛上有一个帖子问了同样的问题，但没有解决方案。

后来在https://qengineering.eu/install-pytorch-on-jetson-nano.html看到torchvision也可以用.whl文件安装，

 

于是下载.whl文件，很顺利地就安装好了。

困扰一个星期的问题解决了。