Fundamentals

Overview

A workspace is a comprehensive virtual environment tailored for managing machine learning (ML) projects across their lifecycles.

Key Features

• Isolation and Security: Ensures that each project operates in a secure and isolated environment, preventing conflicts and safeguarding sensitive data. User access to workspaces can be managed through Workspace Access. Additionally, for broader user and role management, refer to the User Governance.

• Resource Allocation: Manages and allocates computational resources such as CPU, GPU, memory to different projects within the workspace. Resource allocation can be administered by the tenant admin, providing them control over the distribution of resources. See Resource Allocation Documentation.

• Usage Monitoring: Tracks resource utilization within the workspace, providing insights into how computational resources are allocated and used across different experiments and tasks. This helps in optimizing resource allocation and identifying bottlenecks.

• Model Management and Deployment: Facilitates the management of ML models, from development and validation to deployment and monitoring in production environments.

• Job Management and Scheduling: facilitating the efficient execution and management of machine learning (ML) tasks and workflows. These features are essential for optimizing computational resources, ensuring timely project delivery, and maintaining a high level of system performance.

• Experiment Tracking: Enables tracking of various experiments, allowing data scientists to log results, compare different models, and keep track of changes over time.

• Dataset Management: Provides tools for storing, accessing, and versioning datasets, ensuring data integrity and consistency across projects.