Floating Point Operations Per Second, a measure of computer performance useful for scientific computations that require floating-point calculations, particularly in deep learning.
A deep learning model with 10 billion FLOPS can process large datasets much faster than one with only 1 billion FLOPS.