SparkNet

Distributed Neural Networks for Spark. Details are available in the paper. Ask questions on the sparknet-users mailing list!

Quick Start

Start a Spark cluster using our AMI

1. Create an AWS secret key and access key. Instructions here.

2. Run export AWS_SECRET_ACCESS_KEY= and export AWS_ACCESS_KEY_ID= with the relevant values.

3. Clone our repository locally.

4. Start a 5-worker Spark cluster on EC2 by running

    SparkNet/ec2/spark-ec2 --key-pair=key \
                           --identity-file=key.pem \
                           --region=eu-west-1 \
                           --zone=eu-west-1c \
                           --instance-type=g2.8xlarge \
                           --ami=ami-d0833da3 \
                           --copy-aws-credentials \
                           --spark-version=1.5.0 \
                           --spot-price=1.5 \
                           --no-ganglia \
                           --user-data SparkNet/ec2/cloud-config.txt \
                           --slaves=5 \
                           launch sparknet
   

You will probably have to change several fields in this command. For example, the flags --key-pair and --identity-file specify the key pair you will use to connect to the cluster. The flag --slaves specifies the number of Spark workers.

Train Cifar using SparkNet

1. SSH to the Spark master as root.

2. Run bash /root/SparkNet/data/cifar10/get_cifar10.sh to get the Cifar data

3. Train Cifar on 5 workers using

    /root/spark/bin/spark-submit --class apps.CifarApp /root/SparkNet/target/scala-2.10/sparknet-assembly-0.1-SNAPSHOT.jar 5
   

4. That's all! Information is logged on the master in /root/SparkNet/training_log*.txt.

Train ImageNet using SparkNet

1. Obtain the ImageNet data by following the instructions here with

   wget http://.../ILSVRC2012_img_train.tar
   wget http://.../ILSVRC2012_img_val.tar
   

   This involves creating an account and submitting a request.

2. On the Spark master, create ~/.aws/credentials with the following content:

   [default]
   aws_access_key_id=
   aws_secret_access_key=
   

   and fill in the two fields.

3. Copy this to the workers with ~/spark-ec2/copy-dir ~/.aws (copy this command exactly because it is somewhat sensitive to the trailing backslashes and that kind of thing).

4. Create an Amazon S3 bucket with name S3_BUCKET.

5. Upload the ImageNet data in the appropriate format to S3 with the command

   python $SPARKNET_HOME/scripts/put_imagenet_on_s3.py $S3_BUCKET \
       --train_tar_file=/path/to/ILSVRC2012_img_train.tar \
       --val_tar_file=/path/to/ILSVRC2012_img_val.tar \
       --new_width=256 \
       --new_height=256
   

   This command resizes the images to 256x256, shuffles the training data, and tars the validation files into chunks.

6. Train ImageNet on 5 workers using

   /root/spark/bin/spark-submit --class apps.ImageNetApp /root/SparkNet/target/scala-2.10/sparknet-assembly-0.1-SNAPSHOT.jar 5 $S3_BUCKET
   

Installing SparkNet on an existing Spark cluster

The specific instructions might depend on your cluster configurations, if you run into problems, make sure to share your experience on the mailing list.

1. If you are going to use GPUs, make sure that CUDA-7.0 is installed on all the nodes.

2. Depending on your configuration, you might have to add the following to your ~/.bashrc, and run source ~/.bashrc.

   export LD_LIBRARY_PATH=/usr/local/cuda-7.0/targets/x86_64-linux/lib/
   export _JAVA_OPTIONS=-Xmx8g
   export SPARKNET_HOME=/root/SparkNet/
   

   Keep in mind to substitute in the right directories (the first one should contain the file libcudart.so.7.0).

3. Clone the SparkNet repository git clone https://github.com/amplab/SparkNet.git in your home directory.

4. Copy the SparkNet directory on all the nodes using

   ~/spark-ec2/copy-dir ~/SparkNet
   

5. Build SparkNet with

   cd ~/SparkNet
   git pull
   sbt assembly
   

6. Now you can for example run the CIFAR App as shown above.

Building your own AMI

1. Start an EC2 instance with Ubuntu 14.04 and a GPU instance type (e.g., g2.8xlarge). Suppose it has IP address xxx.xx.xx.xxx.

2. Connect to the node as ubuntu:

   ssh -i ~/.ssh/key.pem ubuntu@xxx.xx.xx.xxx
   

3. Install an editor

   sudo apt-get update
   sudo apt-get install emacs
   

4. Open the file

   sudo emacs /root/.ssh/authorized_keys
   

   and delete everything before ssh-rsa ... so that you can connect to the node as root.

5. Close the connection with exit.

6. Connect to the node as root:

   ssh -i ~/.ssh/key.pem root@xxx.xx.xx.xxx
   

7. Install CUDA-7.0.

   wget http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1404/x86_64/cuda-repo-ubuntu1404_7.0-28_amd64.deb
   dpkg -i cuda-repo-ubuntu1404_7.0-28_amd64.deb
   apt-get update
   apt-get upgrade -y
   apt-get install -y linux-image-extra-`uname -r` linux-headers-`uname -r` linux-image-`uname -r`
   apt-get install cuda-7-0 -y
   

8. Install sbt. Instructions here.

9. apt-get update

10. apt-get install awscli s3cmd

11. Install Java apt-get install openjdk-7-jdk.

12. Clone the SparkNet repository git clone https://github.com/amplab/SparkNet.git in your home directory.

13. Add the following to your ~/.bashrc, and run source ~/.bashrc.

    export LD_LIBRARY_PATH=/usr/local/cuda-7.0/targets/x86_64-linux/lib/
    export _JAVA_OPTIONS=-Xmx8g
    export SPARKNET_HOME=/root/SparkNet/
    

    Some of these paths may need to be adapted, but the LD_LIBRARY_PATH directory should contain libcudart.so.7.0 (this file can be found with locate libcudart.so.7.0 after running updatedb).

14. Build SparkNet with

    cd ~/SparkNet
    git pull
    sbt assembly
    

15. Create the file ~/.bash_profile and add the following:

    if [ "$BASH" ]; then
      if [ -f ~/.bashrc ]; then
        . ~/.bashrc
      fi
    fi
    export JAVA_HOME=/usr/lib/jvm/java-7-openjdk-amd64
    

    Spark expects JAVA_HOME to be set in your ~/.bash_profile and the launch script SparkNet/ec2/spark-ec2 will give an error if it isn't there.

16. Clear your bash history cat /dev/null > ~/.bash_history && history -c && exit.

17. Now you can create an image of your instance, and you're all set! This is the procedure that we used to create our AMI.

JavaCPP Binaries

We have built the JavaCPP binaries for a couple platforms. They are stored at the following locations:

1. Ubuntu with GPUs: http://www.eecs.berkeley.edu/~rkn/snapshot-2016-03-05/
2. Ubuntu with CPUs: http://www.eecs.berkeley.edu/~rkn/snapshot-2016-03-16-CPU/
3. CentOS 6 with CPUs: http://www.eecs.berkeley.edu/~rkn/snapshot-2016-03-23-CENTOS6-CPU/