Managed JupyterLab notebook instances
AI Platform Notebooks is a managed service that offers an integrated JupyterLab environment in which machine learning developers and data scientists can create instances running JupyterLab that come pre-installed with the latest data science and machine learning frameworks in a single click. Notebooks is integrated with BigQuery, Cloud Dataproc, and Cloud Dataflow, making it easy to go from data ingestion to preprocessing and exploration, and eventually model training and deployment.
Get up and running fast
You can deploy new JupyterLab instances with one click and start analyzing your data immediately. Each instance comes pre-configured with optimized versions of the most popular data science and machine learning libraries; no worrying about creating and managing VMs.
Use popular open source frameworks
There’s no learning curve — Notebooks uses the industry-standard JupyterLab interface and comes pre-installed with optimized versions of popular libraries like TensorFlow, PyTorch, scikit-learn, pandas, NumPy, SciPy, and Matplotlib.
Scale on demand
You can start small and scale up by adding CPUs, RAM, and GPUs. When your data gets too big for one machine, seamlessly switch to distributed services like BigQuery, Cloud Dataproc, Cloud Dataflow, and AI Platform Training and Prediction.
You’ll go from data to a deployed machine learning model without leaving Notebooks. Pull data from BigQuery, use Cloud Dataproc to transform it, and leverage AI Platform services or Kubeflow for distributed training and online prediction.
Managed JupyterLab experience
Notebooks is built on the industry standard JupyterLab. So you can use it with the RPython and R data science community and customize your environment by installing JupyterLab plugins.
AI Hub support
Notebooks is tightly integrated with AI Hub, a rich catalog of plug-and-play AI components. You can discover and run readily deployable and interactive notebooks by AI researchers and customer engineers at Google and our partners and foster collaboration by sharing your own notebooks within your GCP organization to help others build on your work.
Support for data science frameworks
We provide a pre-configured environment that supports the most popular data science libraries, including R, pandas, NumPy, SciPy, scikit-learn, and Matplotlib, and ML frameworks like TensorFlow, Keras, XGBoost, and PyTorch.
Optimized for machine learning
Notebooks' optimized versions of TensorFlow and PyTorch enable you to get the most out of GCP hardware and seamlessly add and remove GPUs from your instance.
Notebooks comes pre-installed with GCP client libraries that make it easy to analyze your data and build, train, and deploy ML models.
It’s easy to pull and push notebooks from your Git repository, making it also easy to share your notebooks with colleagues.
You can select any virtual private cloud (VPC) for their Notebook instances, provided that your VPC has access either through Google Private Access or the internet to Google cloud storage. You can also turn off public IP address and access your instance via proxy.
Single user mode and service account access
You can create a Notebook instance that is accessible by a single user. You can also specify a service account that has access to the Notebook.
Bring your own container
You can run a Notebook instance on a container of your choice. This provides you the flexibility to install specific libraries mandated by your organization or preconfigure the environment running JupyterLab to your preference.
There are no minimum fees or up-front commitments. There’s no charge for using Notebooks. You pay only for the cloud resources you use with the Notebooks instance: AI Platform Training, AI Platform Predictions, Compute Engine, BigQuery, and Cloud Storage. Our pricing calculator can help you estimate the costs of running your workloads.
Take the next step
This product is in beta. For more information on our product launch stages, see here.