Hacker News Reader

so how many images does the current SOTA take to train a classifier? Trying to gauge how much of an improvement Deepmind has made here.

The results that should be used to compare their results to others involves training on just under 1.3 million images across the 1,000 classes.

Their best results involve 'pretraining' on a dataset of 300 million examples, before 'tuning' it on the actual ImageNet training dataset as above.