Anaconda上安装TensorFlow:
#预先安装anaconda,从https://www.continuum.io下载mkdir Anacondacd Anacondawget http://repo.continuum.io/archive/Anaconda3-2.5.0-Linux-x86_64.shchmod +x Anaconda3-2.5.0-Linux-x86_64.sh./Anaconda3-2.5.0-Linux-x86_64.sh#这里假定部署在conda环境tensor下。conda create -n tensor anacondasource activate tensor#安装这几个极其强大的支持库conda install -n tensor -y -c https://conda.anaconda.org/jjhelmus tensorflowconda install -n tensor -y -c https://conda.anaconda.org/anaconda-cluster sparkconda install -n tensor -y -c https://conda.anaconda.org/menpo opencv3conda install -n tensor -y -c https://conda.anaconda.org/menpo ffmpeg#安装一些辅助的库(可选)conda install -n tensor -y blazeconda install -n tensor -y daskconda install -n tensor -y libdyndconda install -n tensor -y ipyparallel#安装Jupyter,原来的IPython的notebook功能分拆出来的。#基于浏览器的python界面,支持notebook。conda install -n tensor -y jupyter#创建jupyter配置文件,原来的IPython的一些教程要退休了,按这个:jupyter notebook --generate-config#修改jupyter配置文件,在~/.jupyter/下。gedit ~/.jupyter/jupyter_notebook_config.py#搜到这一行:c.NotebookApp.ip = 'LOCALHOST'#将LocalHost改为自己的机器名、域名或者IP地址。#不知道IP地址的,用ifconfig查看(linux上)。#保存,退出gedit。#启动Jupyterjupyter notebook#打开浏览器,远程访问http://myhost_ip:8888#恭喜!可以在联机的其它系统上访问了。#注意:一定要在对应的conda环境下启动,否则指定的库找不到。#如果需要账号限制,可以修改配置文件的其它项。