I have a dataset of around 150k stool images, and I’m trying to better understand the “right” way to use it for training a computer vision model. Right now, our process is pretty manual. We initially trained on about 5k images that were individually verified by a human. For every image, we checked/c
