We provide a collection of popular neural network models and compare their floating point and quantized performance.