A dataset represents a collection of files. A file contains raw data (like an image), and encoded meaning (like box or polygons, otherwise known as instances.).

A few example uses:

Separate data into two or more models. For the high/low class imbalance pattern.
Organize work. e.g. prior to export, use the power of the filters, only showing completed, and and then export only completed
Create training splits in order to test different theories

When creating a new batch of work, exporting, etc. you can choose which directory to pull from. This is Active Data in the sense that as the data changes, any new activity that uses the directory will also change.

Changing current directory

Select a directory from the drop down menu

Files will refresh automatically

New directory

Click the plus button
Choose a name and click create

Update / Rename / Archive

Please contact us if you need to restore an archived directory.