20210317
遇到三个问题:
1
服务器上下载的pytorch是cpu版本的
解决:配置pytorch源
conda config --add channels https:
//mirrors
.tuna.tsinghua.edu.cn
/anaconda/cloud/pytorch/
conda
install
pytorch torchvision cudatoolkit=10.1
2
训练到(700/1000)停止,pycharm连接不上服务器的解释器:Couldn't refresh skeletons for remote interpreter
解决:根据网上的方法,
删掉pycharm安装文件夹下的skeletons文件夹,
没解决问题的话重新改远程连接的配置,有两个地方
(1)tools -> Deployment -> Configuration connection配置(2)File -> Setting Project Interpreter
3
训练(24/100)停止:
RuntimeError: main thread is not in main loop
Tcl_AsyncDelete: async handler deleted by the wrong thread
Aborted (core dumped)
解决:未解决。
本文原创自博客园文章
地址:https://www.cnblogs.com/oxygen-99/