TensorflowOnSpark-单机版

GetStarted_Standalone(单机版)

参考别人部署： http://mp.weixin.qq.com/s/sLyjwU-FZqoOtiGasJnTZw

https://github.com/yahoo/TensorFlowOnSpark/wiki/GetStarted_standalone

http://blog.csdn.net/u013041398/article/details/55190582

http://blog.csdn.net/fishseeker/article/details/61918138?utm_source=tuicool&utm_medium=referral

http://www.tuicool.com/articles/zqE7rib

使用到的软件说明：

虚拟机 ubuntu17.04

python 2.7.13(ubuntu自带)

Java7+ jdk-8u131-linux-x64.tar.gz

spark spark-1.6.0-bin-hadoop2.6.tgz

TensorFlowOnSpark git 官网

TensorFlow 0.12.1

参考： https://github.com/yahoo/TensorFlowOnSpark/wiki/GetStarted_standalone

1、直接在虚拟机操作 成功

2、SSH连接(未设置无密码模式) 失败

3、SSH连接(设置无密码模式) 成功

注：SSH 无密码模式设置，

参考：Linux / 在线虚拟机 / ssh访问Linux系统

Running TensorFlowOnSpark Locally

We illustrate how to apply TensorFlowOnSpark in a standalone Spark cluster, which is installed on your local machine.

1. Clone TensorFlowOnOnSpark code.

git clone --recurse-submodules https://github.com/yahoo/TensorFlowOnSpark.git
cd TensorFlowOnSpark
git submodule init
git submodule update --force
git submodule foreach --recursive git clean -dfx

cd TensorFlowOnSpark 
export TFoS_HOME=$(pwd)
pushd src
zip -r ../tfspark.zip *
popd

2. Install Spark

install Apache Spark 1.6.0 per instruction at  http://spark.apache.org/downloads.html .

先配置Java环境（参考：http://blog.csdn.net/wc781708249/article/details/78223371）

${TFoS_HOME}/scripts/local-setup-spark.sh  #自动下载及安装spark-1.6.0-bin-hadoop2.6.tar
# --------上面代码详解---------------------------
# vim ${TFoS_HOME}/scripts/local-setup-spark.sh 
wget http://archive.apache.org/dist/spark/spark-1.6.0/spark-1.6.0-bin-hadoop2.6.tgz
gunzip spark-1.6.0-bin-hadoop2.6.tgz
tar -xvf spark-1.6.0-bin-hadoop2.6.tar
# ------------------------------------
或
wget https://d3kbcqa49mib13.cloudfront.net/spark-1.6.0-bin-hadoop2.6.tgz
tar -zxvf spark-1.6.0-bin-hadoop2.6.tgz
rm spark-1.6.0-bin-hadoop2.6.tarexport SPARK_HOME=$(pwd)/spark-1.6.0-bin-hadoop2.6export PATH=${SPARK_HOME}/bin:${PATH}

3. Installing TensorFlow from Binary

Please build and install TensorFlow per  instruction .

On Mac OS, for example, you could install TensorFlow as below:

export TF_BINARY_URL=https://storage.googleapis.com/tensorflow/mac/cpu/tensorflow-0.12.1-py2-none-any.whl
sudo pip install --upgrade $TF_BINARY_URL

Linux 安装 0.12.1版本

sudo pip install --upgrade https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-0.12.1-cp27-none-linux_x86_64.whl

Test TensorFlow:

# download MNIST files, if not already done
mkdir ${TFoS_HOME}/mnist
pushd ${TFoS_HOME}/mnist
curl -O "http://yann.lecun.com/exdb/mnist/train-images-idx3-ubyte.gz"
curl -O "http://yann.lecun.com/exdb/mnist/train-labels-idx1-ubyte.gz"
curl -O "http://yann.lecun.com/exdb/mnist/t10k-images-idx3-ubyte.gz"
curl -O "http://yann.lecun.com/exdb/mnist/t10k-labels-idx1-ubyte.gz"
popd

python ${TFoS_HOME}/tensorflow/tensorflow/examples/tutorials/mnist/mnist_with_summaries.py --data_dir ${TFoS_HOME}/mnist

4. Launch  standalone Spark cluster

Start master:

${SPARK_HOME}/sbin/start-master.sh

Start one or more workers and connect them to the master via master-spark-URL. Go to MasterWebUI , make sure that you have the exact number of workers launched.

export MASTER=spark://$(hostname):7077
export SPARK_WORKER_INSTANCES=2
export CORES_PER_WORKER=1 
export TOTAL_CORES=$((${CORES_PER_WORKER}*${SPARK_WORKER_INSTANCES})) 
${SPARK_HOME}/sbin/start-slave.sh -c $CORES_PER_WORKER -m 3G ${MASTER}

5. Convert the MNIST zip files

cd ${TFoS_HOME}
rm -rf examples/mnist/csv
${SPARK_HOME}/bin/spark-submit \
--master ${MASTER} \
${TFoS_HOME}/examples/mnist/mnist_data_setup.py \
--output examples/mnist/csv \
--format csv
ls -lR examples/mnist/csv

6. Run distributed MNIST training (using feed_dict)

# rm -rf mnist_model
${SPARK_HOME}/bin/spark-submit \
--master ${MASTER} \
--py-files ${TFoS_HOME}/tfspark.zip,${TFoS_HOME}/examples/mnist/spark/mnist_dist.py \
--conf spark.cores.max=${TOTAL_CORES} \
--conf spark.task.cpus=${CORES_PER_WORKER} \
--conf spark.executorEnv.JAVA_HOME="$JAVA_HOME" \
${TFoS_HOME}/examples/mnist/spark/mnist_spark.py \
--cluster_size ${SPARK_WORKER_INSTANCES} \
--images examples/mnist/csv/train/images \
--labels examples/mnist/csv/train/labels \
--format csv \
--mode train \
--model mnist_model

ls -l mnist_model

7. Run distributed MNIST inference (using feed_dict)

# rm -rf predictions
${SPARK_HOME}/bin/spark-submit \
--master ${MASTER} \
--py-files ${TFoS_HOME}/tfspark.zip,${TFoS_HOME}/examples/mnist/spark/mnist_dist.py \
--conf spark.cores.max=${TOTAL_CORES} \
--conf spark.task.cpus=${CORES_PER_WORKER} \
--conf spark.executorEnv.JAVA_HOME="$JAVA_HOME" \
${TFoS_HOME}/examples/mnist/spark/mnist_spark.py \
--cluster_size ${SPARK_WORKER_INSTANCES} \
--images examples/mnist/csv/test/images \
--labels examples/mnist/csv/test/labels \
--mode inference \
--format csv \
--model mnist_model \
--output predictions

less predictions/part-00000

7. Interactive Learning with Jupyter Notebook

Install additional software required by Jupyter Notebooks.

sudo pip install jupyter jupyter[notebook]

Launch IPython notebook on Master node.

pushd ${TFoS_HOME}/examples/mnist
PYSPARK_DRIVER_PYTHON="jupyter" \
PYSPARK_DRIVER_PYTHON_OPTS="notebook --no-browser --ip=`hostname`" \
pyspark  --master ${MASTER} \
--conf spark.cores.max=${TOTAL_CORES} \
--conf spark.task.cpus=${CORES_PER_WORKER} \
--py-files ${TFoS_HOME}/tfspark.zip,${TFoS_HOME}/examples/mnist/spark/mnist_dist.py \
--conf spark.executorEnv.JAVA_HOME="$JAVA_HOME"

Launch a browser to access  notebook

8. Shutdown Spark cluster

${SPARK_HOME}/sbin/stop-slave.sh
${SPARK_HOME}/sbin/stop-master.sh