We augment the dataset to create more variation in the existing dataset (simulating images taken in varying conditions), and to increase the size of our training set itself.
Once all images have been labelled, navigate to Dataset in the left bar and ensure that the split is as close to 70/20/10 as possible (a few images of variance is fine). Then, navigate to Generate. Here, you will create a "Version" of the dataset.
Currently, here are the settings we use:
Generate an augmented version:
Leave default
Leave default
None
Flip, Hue, Saturation, Brightness:
2022-2023 used the following:
Flip: Horizontal, Vertical
Hue: -27° and +27°
Saturation: -75% and +75%
Brightness: -25% and +0%
3x (but if a higher multiplier is available, use it)
Click Generate (it takes a few minutes for the version to be generated)
To download a dataset version, see Dataset Download and Usage.