Glossary

A list of terms and definitions related to NimbleBox.ai.

Access

Access refers to the ability of a member to access and utilize the resources within the workspace. It is typically controlled by roles and permissions within the workspace.

App Bar

An app bar is a vertical toolbar located on the left side of the instance home. The app bar provides a convenient way to access the different IDEs of the build instance.

Auto-scaling

The autoscaling feature allows you to scale the number of instances in your application based on its demand. This means that if the demand for the application increases, the number of instances can be scaled in order to handle the additional workload. Conversely, if the demand for the application decreases, the number of instances can be decreased to save costs. Learn more about the benefits of this auto-scaling here.

Auto-shutdown

Auto-shutdown is a feature that helps you to automatically stop or terminate your instances at a specific time. Auto-shutdown helps you to save costs by automatically stopping instances that have been running for an extended period of time. Learn how to set up an auto-shutdown timer here.

Backup

A backup is a copy of the instance that can be used to restore the original instance in the event of failure or accidental deletion.

Boot up script

A boot-up script is a script that is automatically executed when a build instance is started. It automates the initialization process and saves time.

Cores

Cores refer to the individual processing units within a CPU. A CPU with multiple cores can perform multiple tasks at the same time, which can improve the performance and speed of the instance.

Custom ports

Custom ports help you to specify port numbers that are used to host your application. It should be different from the default port numbers. Get a closer look of custom ports here.

Dedicated Instances

Dedicated instances come with the ability to run your workloads without any interruption. Dedicated instances can be a useful option when you need a high level of performance and to ensure that your training will not be interrupted.

Development environment

Development environment is a space inside the build instance that allows you to build, test and debug your code.

Git provider

Git provider is a platform that allows users to manage your git repositories in a centralized location. NimbleBox allows you to integrate with git providers such as GitHub, GitLab, and Bitbucket.

Hardware configuration

The hardware configuration refers to the specific hardware components that can be customized to meet your computational workloads in the build instance. The hardware configuration typically includes CPUs, GPUs, storage, and the region of the instance.

Image

An Image is a file that contains the instructions and dependencies needed to create a container. Image can be shared and reused across different environments.

Instance

An instance is a virtual machine that runs on our cloud platform. Instances can be easily created, configured, and managed according to your computing needs. An instance consists of code, libraries, dependencies, and any other necessary files that you can have to train a model. NimbleBox provides flexibility and customizable hardware configuration, as the user can easily allocate and adjust the resources for each instance as needed. It is also referred to as compute instance or build instance. Learn more about creating an instance here.

Instance Disk

Instance disk is a type of instance storage that contains files, folders, and the environment of your build instance. Learn more about managing storage here.

Instance Home

Instance Home is the layout inside the build instance where you can code and build machine learning models using various integrated development environments (IDEs)

Instance Snapshot

An instance snapshot is a point-in-time copy of a build instance, which includes the instance's data and the environment. Instance snapshots can be used to create new build instances based on the snapshot, which can be useful for versioning, testing, or development purposes. To find out more about managing versions here.

Permissions

Permissions refer to the specific actions or operations that are allowed or restricted for the member of a workspace.

Platform

NimbleBox platform consists of a comprehensive set of tools and components that form a foundation for developing and deploying machine learning models. The platform is organized into a number of different sections, each of which is designed to help you manage different aspects of your MLOps workflow.

Ports

Ports are a unique number that allows running multiple applications on a network.

Pre-emptive Instances

Pre-emptive instances are similar to regular compute instances, but they are typically offered at a lower cost and can be terminated at any point in time when there is an increase in demand for resources. It is also referred to as spot instances. Pre-emptive instances can be a useful option for users who want to save money on their cloud computing costs, but they should be used with caution and only for tasks that can tolerate potential interruptions. Discover more about pre-emptive instances here.

Project Visibility

Refers to the ability of users to access and view a project within the workspace. It determines who can access and view a project, and what actions are permitted to take within it.

Region

Region refers to the geographic location where the instances are hosted. It is spread across India, Taiwan, and the USA. You can choose the regions based on the availability of hardware, performance, cost, and compliance requirements of the data. Read more about the available regions here.

RAM

In an instance, random access memory stores data and instructions that are currently being used or processed by the CPU cores.

Roles

Roles refer to the specific tasks, responsibilities, and permissions that are associated with a member of a workspace.

VRAM

Data and instructions are stored in Video Random Access Memory (VRAM), which is used by GPUs to process data. VRAM is used to store the data and instructions that are needed by the GPU in order to perform the complex calculations and matrix operations that are required for machine learning. VRAM typically has a larger capacity than regular RAM, which can allow users to train larger and more complex models.

Workspace

A workspace is a designated area where team members can work on projects, collaborate with others, and access the tools provided to build your MLOps pipeline. Get started with workspace here.

Workspace Menu

The workspace menu is a navigation menu within the NimbleBox platform that allows you to access, and manage the different workspaces. The workspace menu provides the ability to create new workspaces, join an existing workspace, view the list of workspaces, and switch between them. For more information on workspaces, click here.