一、(base) C:\Users\chenchao>activate tensorflow
Could not find conda environment: tensorflow
You can list all discoverable environments with `conda info --envs`.
解决方法:在anaconda prompt中输入conda create -n tensorflow python=3.6
二、2018-07-02 21:27:15.879531: 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
提升cpu性能的,若安装的是GPU版本 不用理会
三、distributed 1.21.8 requires msgpack, which is not installed.
缺少了这个包, 在anaconda prompt中输入pip install msgpack安装即可
四、cuda cudnn下载后安装和环境配置
cuda默认安装在 C:\Program Files\NVIDIA GPU Computing Toolkit,环境不用配置
cudnn下载后可以解压在任意目录下,解压后需要把cudnn目录里各个文件夹里面的文件分别对应复制到cuda文件夹里,紧接着在控制面板里搜索环境变量,在环境配置中找到path,点击编辑-——新建,把你自己的cudnn里的bin路径写进去(例如我的是:D:\cuda\bin)