1、创建空间：

conda create --name tf-nogpu

2、打开tf-nogpu空间：

source activate tf-nogpu

3、退出空间：

source deactivate

4、删除这个空间：

conda remove --name tf-nogpu也可以去anaconda文件夹下面####5、删除envs下面的文件夹tf-nogpu

6、创建特定版本Python库的镜像空间

conda create --name tf-nogpu python=3.5

退出： Ctrl+D
安装文件：pip install tensorflow
pip install scikit-image用来处理图片的库
pip install keras类似tf的库
安装anconda时候不能用sudo命令，否则tensflow可能会调用不了
~/Downloads$ ls
~/Downloads$ chmod +x Anaconda3-5.0.1-Linux-x86_64.sh ~/Downloads$ ./Anaconda3-5.0.1-Linux-x86_64.sh
下载X86版本的 ，卸载anaconda3：
sudo rm -rf anaconda3

添加环境变量：
source ~/.bashrc
export PATH=/home/dongdong/anaconda3/bin:$PATH
安装pycharm,压缩后直接拖到特定的文件夹
~/programs/pycharm-community-2017.3.2/bin$ ./pycharm.sh
sudo apt-get install libcupti-dev

Fix the error ” make: execvp: Permission denied “

” make: execvp: <YOUR FILE NAME>: Permission denied … ”
You got this error message? In my case, I got this message when trying to compile a software in my Linux box. To fix this issue, just simply change file access permissions using chmod command
$ chmod 755 /path/to/yourfile
If you want to change file access permissions for a folder and all files in that folder, you can add the -R option to the chmod command，the commond needs to be operated at sudo
$ chmod 755 -R /path/to/yourfolder

kill the computer processes shown in nvidia-smi?
sudo kill -9 6154 6154 is the PID number`