batch-shipyard/recipes/Caffe-GPU

Caffe-GPU

This recipe shows how to run Caffe on GPUs using N-series Azure VM instances in an Azure Batch compute pool.

Configuration

Please see refer to this set of sample configuration files for this recipe.

Pool Configuration

The pool configuration should enable the following properties:

• vm_size must be a GPU enabled VM size. Because Caffe is a GPU-accelerated compute application, you should choose a GPU compute accelerated VM instance size.
• vm_configuration is the VM configuration. Please select an appropriate platform_image with GPU as supported by Batch Shipyard.

Global Configuration

The global configuration should set the following properties:

• docker_images array must have a reference to a valid Caffe GPU-enabled Docker image. Although you can use the official BVLC/caffe Docker images, for this recipe the alfpark/caffe:gpu contains all of the required files and scripts to run the MNIST convolutional example.

Jobs Configuration

The jobs configuration should set the following properties within the tasks array which should have a task definition containing:

• docker_image should be the name of the Docker image for this container invocation, e.g., alfpark/caffe:gpu
• command should contain the command to pass to the Docker run invocation. For the alfpark/caffe:gpu Docker image and to run the MNIST convolutional example on all available GPUs, we are using a run_mnist.sh helper script. Thus, the command would simply be: "/caffe/run_mnist.sh -gpu all"
• gpu can be set to true, however, it is implicitly enabled by Batch Shipyard when executing on a GPU-enabled compute pool.

Dockerfile and supplementary files

The Dockerfile for the Docker image can be found here.

You must agree to the Caffe License prior to use.