This dataset contains images and labels for the 2 example videos from the FACEMAP dataset. These images have been rescaled to fit the 224x224 pixel input to the ViT, but no further adjustments are made.