快速运行TensorFlow的6种方式

快速运行TensorFlow的6种方式

TensorFlow（http://tensorflow.org）是一个深度学习计算引擎，天然是能够直接安装运行的，并且能获得最佳的性能。可是，考虑到机器学习须要安装大量的软件，之间必然会带来软件管理和版本兼容性问题，并且在集群中运行更为复杂，所以不推荐这种方式。

• 在Linux上安装TensorFlow，https://www.tensorflow.org/install/install_linux
• 搭建分布式Tensorflow集群，http://www.javashuo.com/article/p-bobzteka-r.html
  • 分布式TensorFlow，中文，http://www.javashuo.com/article/p-rrneniir-gd.html
  • TensorBoard张量图可视化，https://my.oschina.net/repine/blog/918576
• TensorFlow on Kubernetes架构实践，https://my.oschina.net/jxcdwangtao/blog/1612667

这里将介绍几种能够在隔离环境中运行TensorFlow的方式，包括：Anaconda、Docker、Jupyter、Kubernetes POD、Kubeflow、Spark DL on MLlib，将更容易管理和扩展到集群中运行。

• 本文中代码在：https://github.com/openthings/kubernetes-tools/tree/master/tensorflow

一、Anaconda

经过Anaconda来运行Tensorflow的Python调用代码，优势是能够提供一个相对隔离的Python运行环境，避免与其它的python任务产生版本冲突。虽然也能够经过VirtualEnv完成相似的任务，但Anaconda自带包管理功能，用起来更为方便。

下载和安装Anaconda:

echo ""
echo "====================================================================="
echo "Downloading Anaconda3 5.1.0 to ~/openthings..."
echo ""

cd ~/openthings
wget -c https://repo.anaconda.com/archive/Anaconda3-5.1.0-Linux-x86_64.sh

echo "====================================================================="
echo "Finished.Latest version at https://repo.continuum.io/archive/"
echo "More practice at https://my.oschina.net/u/2306127/blog"
echo "====================================================================="
echo ""

下载和安装Miniconda：

echo ""
echo "====================================================================="
echo "Downloading Miniconda3 4.5.1 to ~/openthings..."
echo ""

cd ~/openthings
wget -c https://repo.continuum.io/miniconda/Miniconda3-4.5.1-Linux-x86_64.sh

echo "====================================================================="
echo "Finished.Latest version at https://repo.continuum.io/miniconda/"
echo "More practice at https://my.oschina.net/u/2306127/blog"
echo "====================================================================="
echo ""

安装TensorFlow on Anaconda：

echo ""
echo "================================================================="
echo "Create conda env for Tensorflow, python 3.6.5 ..."
conda create --yes -n tensorflow pip python=3.6.5
source activate tensorflow
echo ""

echo ""
echo "================================================================="
echo "Install Tensorflow 1.8.0 with GPU support."
pip install --ignore-installed --upgrade \
	https://storage.googleapis.com/tensorflow/linux/gpu/tensorflow_gpu-1.8.0-cp36-cp36m-linux_x86_64.whl
echo "================================================================="
echo "More practice at https://my.oschina.net/u/2306127/blog"
echo "More info,Visit https://www.tensorflow.org/install/install_linux "
echo "================================================================="
echo ""

而后运行：

source activate tensorflow

这样切换Python运行环境，就能够在Anaconda环境下经过Python执行Tensorflow调用代码了，固然也能够安装Jupyter Notebook之类的组件，而后在浏览器里编写和运行tensorflow计算任务。

二、Docker

将TensorFlow放到Docker容器中执行，安装部署都更为方便，不污染宿主机环境，快速实验多种版本。

echo ""
echo "================================================================="
echo "Running Tensorflow in Docker with bash shell."
echo "More practice at https://my.oschina.net/u/2306127/blog"
echo "Please Visit https://www.tensorflow.org/install/install_linux "
echo "================================================================="
echo ""

nvidia-docker run -it tensorflow/tensorflow:latest-gpu bash

三、Jupyter

在Docker中安装Anaconda和Jupyter。

echo ""
echo "================================================================="
echo "Running Tensorflow in Docker with jupyter notebook."
echo "More practice at https://my.oschina.net/u/2306127/blog"
echo "Please Visit https://www.tensorflow.org/install/install_linux "
echo "================================================================="
echo ""

nvidia-docker run -it -p 8888:8888 tensorflow/tensorflow:latest-gpu

而后按照提示的地址，到浏览器打开网址 http://localhost:8888，就可使用了。

四、Kubernetes POD

若是使用Kubernetes/Minikube或者OpenShift，能够将Tensorflow部署到K8S集群中，参见：

• Kubernetes集成TensorFlow服务，http://www.javashuo.com/article/p-mxgqswxy-mp.html

五、Kubeflow

经过Kubernetes创建TensorFlow机器学习集群的更佳方式是使用KubeFlow，目前该项目启动时间不长，还须要进一步完善。使用参考：

• Kubeflow使用指南，http://www.javashuo.com/article/p-nppedcay-hp.html

六、Spark DL on MLlib

因为TensorFlow只是完成计算功能，经过Spark ML创建深度学习流水线，从而能够充分利用Spark的分布式内存和数据处理、数据IO、交互分析等功能，而后进一步将其运行在Kubernetes集群之中统一调度。

• Apache Spark上的深度学习流水线 ，http://www.javashuo.com/article/p-wafhvhjg-cq.html