Projects
Last updated
Last updated
Project is a primary organizational unit designed to house and manage datasets, along with metadata such as classes and tags. Think of it as a "superfolder" where all datasets within the project share common metadata configurations, including predefined classes and tags set at the project-level.
The Projects page displays all available projects in the currently selected Workspace. Each project appears as a card or line, showing key details such as:
Thumbnail Preview: Provides a quick visual overview of the project's content (e.g. an image or video frame).
Project Name: The title of the project, such as "Wood Defect Detection".
Data Count: Shows the total items within the project, broken down by data type (e.g., "888 images in 3 datasets").
Project ID and Creation Date: Helps track and organize projects, especially in collaborative environments.
Projects are displayed either in Card View (default) or Table View. You can switch between these views by using the icon at the top-right corner of the page.
At the Projects page you can view all projects you have in the current workspace.
Please note the "three dots" (⋮) icon in the bottom right corner of a project. This is what we call the "context menu". From here, you can perform many important activities related to the project, such as cloning the project, running an app for the project, or deleting the project.
Wait, there is no Create button on this page. So how do I create one?
You can take a sample project from the ecosystem.
Upload images, videos or other files from your computer at the Import page. You will be asked to provide a name for a Project — and after successful import, you will have one.
If you want to automate the process of adding new data, it's a way to go! We have a powerful API and SDK that lets you start in no time.
At the moment we support:
Images
Videos
3D Point Clouds
DICOM
You can see a project type in the top left corner of a project card.
At the Project page you can change how to list projects and datasets: as cards or as a table. You can switch the view by clicking the icon at the most right side of the page.
Show only specific projects and datasets by multiple parameters, including project type, labeling job, author and so on.
Learn more about using conditional filters and building custom queries on your training data in our comprehensive blog post.
You can also find the list of your projects across all of your teams and workspaces by navigating to the Data Commander. There, you can move or copy both projects, datasets, images and even project metadata.
You can remove one or multiple projects and datasets by selecting them using checkboxes and clicking the Move to Trash button. This won’t delete your data immediately — you can learn more in the Storage Cleanup section.
Datasets
Dataset is the second most important concept in Supervisely.
Classes
Classes are pre-defined types of your annotations.
Tags
Sometime you need more than a bunch of marked pixels on an image.
Statistics
Having entered the project, in the top menu we see the statistics section.
