An Exploration of Oracle Cloud Infrastructure Data Science and AI Services

December 22, 2023

Oracle Cloud Infrastructure (OCI) Data Science Services are revolutionizing the field of data science, providing a powerful suite of tools for data scientists to develop and deploy machine learning models with ease. In this article, we will explore the features and capabilities of OCI Data Science and AI services and discover how they can empower data scientists to tackle complex data science tasks.

Notebook Examples

The Accelerated Data Science (ADS) SDK is a data scientist-friendly library that streamlines common data science tasks and integrates seamlessly with other OCI services. To showcase the capabilities of ADS, a collection of JupyterLab notebook examples is provided. These notebooks offer tutorials on various aspects of ADS, including how to store secrets in the OCI Vault service.

Conda Environment Notebooks

OCI Data Science services utilize conda environments to manage libraries available to notebooks. Several pre-configured conda environments are provided for different data science tasks. Conda environment notebooks demonstrate how to perform various data science tasks using these environments, enabling data scientists to quickly get started on their projects.

Labs

For those looking for a comprehensive end-to-end data science experience, the Labs section provides examples of how to train machine learning models and deploy them on the OCI Data Science service. These labs serve as a practical guide, walking users through the entire process step by step.

Model Catalog Examples

The Model Catalog offers a managed and centralized storage space for machine learning models. ADS facilitates the creation of the artifacts required to utilize this service. This section provides examples of how to create the necessary score.py and runtime.yaml files for different machine learning models and configurations.

Jobs

OCI Data Science Jobs provide a powerful tool for running machine learning tasks on a fully managed infrastructure. With the ability to define and schedule jobs, businesses can automate their data processing workflows, reducing manual intervention and ensuring efficient and error-free execution. This section explores the capabilities of OCI Data Science Jobs and highlights the benefits of on-demand jobs and batch processing.

Distributed Training

Distributed training support with Jobs allows for faster and more efficient model training on large datasets. This section covers the support for distributed training with frameworks such as Dask, Horovod, TensorFlow Distributed, and PyTorch Distributed. Data scientists can leverage distributed training when dealing with complex models and large workloads.

Pipelines

Pipelines play a crucial role in streamlining and automating the model building and deployment process. By automating the machine learning process, data scientists can achieve faster and more consistent results. OCI Data Science Pipelines simplify the creation and deployment of machine learning pipelines, enabling data scientists to build, train, and deploy complex models with ease.

Data Labeling Examples

Data labeling is a fundamental task in machine learning, enabling the identification and annotation of properties in documents, text, and images. This section provides Python and Java scripts for bulk annotation of records in the OCI Data Labeling Service. Data scientists can leverage these examples to streamline their labeling workflows.

Notebook Lifecycle Script Examples

OCI Data Science services offer managed notebook sessions, and notebook lifecycle scripts allow users to execute custom scripts during different stages of the notebook’s lifecycle. This section provides examples of ready-to-use scripts for various lifecycle events, showcasing the flexibility and customization options available to data scientists.

In conclusion, Oracle Cloud Infrastructure Data Science and AI services empower data scientists with a comprehensive suite of tools and capabilities. From notebook examples and conda environments to labs, jobs, distributed training, pipelines, and data labeling, OCI offers a complete ecosystem for data science. By leveraging these services, data scientists can accelerate their workflow, scale their machine learning capabilities, and drive innovative solutions. For further information and to contribute to the project, refer to the provided resources.

I hope you found this article insightful and informative. If you have any questions or suggestions, please feel free to ask.

References:

Looking to contribute? Visit the OCI Data Science and AI services GitHub repository and review the contribution guide.

Group Sum