meta.json
file containing Project Meta, dataset folders with the individual annotation files (and optionally the original data files) in them. This allows you to seamlessly cycle data between Supervisely and local storage with the use of Supervisely Format
import plugin, if you so require.project_name.tar
project name
meta.json
file obj_class_to_machine_color.json
file (optional, for image annotation projects)key_id_map.json
file (optional)dataset_name
, which contains:ann
folder, contains annotation files, each named source_media_file_name.json
for the corresponding fileimg
(video
or pointcloud
) optional folder, contains source mediamasks_human
optional folder for image annotation projects, contains .png files with annotations marked on themmasks_machine
optional folder for image annotation projects, contains .png files with machine annotations