Just wanted to chime in to say that the Tensorflow 2.0 Dataset Library is nice for this sort of problem. It supports batching, caching, remote fetching, multipl

Just wanted to chime in to say that the Tensorflow 2.0 Dataset Library is nice for this sort of problem. It supports batching, caching, remote fetching, multiple cpu core splitting...etc.