SageMaker#

RAPIDS can be used in a few ways with AWS SageMaker.

SageMaker Notebooks#

To get started head to the SageMaker console and create a new SageMaker Notebook Instance.

Choose Notebook > Notebook Instances > Create notebook instance.

Select your instance#

If a field is not mentioned below, leave the default values:

NOTEBOOK_INSTANCE_NAME = Name of the notebook instance
NOTEBOOK_INSTANCE_TYPE = Type of notebook instance. Select a RAPIDS-compatible GPU (see the RAPIDS docs) as the SageMaker Notebook instance type (e.g., ml.p3.2xlarge).
PLATFORM_IDENTIFIER = ‘Amazon Linux 2, Jupyter Lab 3’
IAM_ROLE = Create a new role > Create role

Screenshot of the create new notebook screen with a ml.p3.2xlarge selected

Create a RAPIDS lifecycle configuration#

SageMaker Notebook Instances can be augmented with a RAPIDS conda environment.

We can add a RAPIDS conda environment to the set of Jupyter ipython kernels available in our SageMaker notebook instance by installing in a lifecycle configuration script.

Create a new lifecycle configuration (via the ‘Additional Options’ dropdown).

Screenshot of the create lifecycle configuration screen

Give your configuration a name like rapids and paste the following script into the “start notebook” script.

#!/bin/bash

set -e

sudo -u ec2-user -i <<'EOF'

mamba create -y -n rapids -c rapidsai -c conda-forge -c nvidia rapids=24.10 python=3.12 cuda-version=12.5 \
    boto3 \
    ipykernel \
    sagemaker

conda activate rapids

python -m ipykernel install --user --name rapids
echo "kernel install completed"
EOF

Set the volume size to at least 15GB, to accommodate the conda environment.

Then launch the instance.

Select the RAPIDS environment#

Once your Notebook Instance is InService select “Open JupyterLab”

Note

If you see Pending to the right of the notebook instance in the Status column, your notebook is still being created. The status will change to InService when the notebook is ready for use.

Then in Jupyter select the rapids kernel when working with a new notebook.

Screenshot of Jupyter with the rapids kernel highlighted

Run the Example Notebook#

Once inside JupyterLab you should be able to upload the Running RAPIDS hyperparameter experiments at scale example notebook and continue following those instructions.

SageMaker Estimators#

RAPIDS can also be used in SageMaker Estimators. Estimators allow you to launch training jobs on ephemeral VMs which SageMaker manages for you. With this option, your Notebook Isntance doesn’t need to have a GPU… you are only charged for GPU instances for the time that your training job is running.

All you’ll need to do is bring in your RAPIDS training script and libraries as a Docker container image and ask Amazon SageMaker to run copies of it in parallel on a specified number of GPU instances.

Let’s take a closer look at how this works through a step-by-step approach:

Training script should accept hyperparameters as command line arguments. Starting with the base RAPIDS container (pulled from Docker Hub), use a Dockerfile to augment it by copying your training code and set WORKDIR path to the code.
Install sagemaker-training toolkit to make the container compatible with Sagemaker. Add other packages as needed for your workflow needs e.g. python, flask (model serving), dask-ml etc.
Push the image to a container registry (ECR).
Having built our container and custom logic, we can now assemble all components into an Estimator. We can now test the Estimator and run parallel hyperparameter optimization tuning jobs.

Estimators follow an API roughly like this:

# set up configuration for the estimator
estimator = sagemaker.estimator.Estimator(
    image_uri,
    role,
    instance_type,
    instance_count,
    input_mode,
    output_path,
    use_spot_instances,
    max_run=86400,
    sagemaker_session,
)

# launch a single remote training job
estimator.fit(inputs=s3_data_input, job_name=job_name)

# set up configuration for HyperparameterTuner
hpo = sagemaker.tuner.HyperparameterTuner(
    estimator,
    metric_definitions,
    objective_metric_name,
    objective_type="Maximize",
    hyperparameter_ranges,
    strategy,
    max_jobs,
    max_parallel_jobs,
)

# launch multiple training jobs (one per combination of hyperparameters)
hpo.fit(inputs=s3_data_input, job_name=tuning_job_name, wait=True, logs="All")

For a hands-on demo of this, try [“Deep Dive into running Hyper Parameter Optimization on AWS SageMaker”]/examples/rapids-sagemaker-higgs/notebook).