Advantage over training with all data instead of samples

[leo2105]

Hi, I just wanna know what is the difference between training with all the 600 samples and training 100 samples first, 200, 300, .... What does active learning step does? it really select the best images or what? I didn't get clear for me.

Thanks in advance

[IgorSusmelj]

Hi @leo2105,

Active Learning and this tutorial are mostly written for situations where you don't have all data labeled. So you can't just train on all images. In that situation, you'd start with the already labeled data and train a model on it. You can then use the model predictions on the unlabeled data to figure out which batch to label next.

Active Learning is a research field where we try to find the next images to label that maximizes the gain in model accuracy.