Cvat: [Question] Distribute docker container with a built in task

[egrinstein]

Good morning,

We would like to distribute annotation tasks for users within our company in the following manner:

1. The annotator downloads the Cvat image
2. A mechanism (either copying files through docker-compose, or an Http call using the api)
   creates a task with a zipped dataset.
3. The annotator performs the task
4. We retrieve the annotations from the container using the Http api

My questions are:

1. Would this be possible?
2. Could I do this by creating task files and copying the images to the image, or do I have to do this using the ReST API?

Thank you,

Eric

[nmanovic]

@egrinstein , the usual way to annotate data using a team:

• install CVAT on a server inside your company
• create all necessary tasks. If you have only one dataset you can split it on jobs using segment size parameter
• Ask your users to annotate a specific task/job (send them a link).

[nmanovic]

@egrinstein , please ask questions on our gitter channel: https://gitter.im/opencv-cvat. I will close the issue for now.