

Collaborators

Why do you need collaborators?

Deep learning requires millions of images as input. Task of labeling millions of images requires many individuals. These individuals might be freelancers, and they might be located in distant locations. A data-science team needs to engage freelance workforce to complete their tasks.

The Collaborator feature allows data-science team to invite individuals to help with a labeling project. If multiple users are collaborating on an annotating project, TrainingData.io application distributes all of the data attached to a project to the collaborators with access. Collaborators will only get unique data, meaning there will not be multiple annotators annotating the same data.

Roles

Annotator The Annotator role can annotate queued data and view their own work from the Activity Table. An annotator cannot view annotators performance or view/modify any aspect of the project configuration or data sources.

Reviewer The Reviewer role extends the annotator role by being able to see and review all annotated data in the project.

Admin The Admin role can add and edit projects, data and Annotators. By default, the user that signs up for an account has the role of Admin.

Project Based Roles When a user is given a project based role, they don't have access to projects by default. However, the user is granted access to each project with a specific role and the role of each project can be changed. The benefit of this is that you can limit the access and the role of a user depending on the project at your discretion.