Deep learning edge computing system framework based on industrial Internet of Things platform

The invention discloses a deep learning edge computing system framework based on an industrial Internet of Things platform. The deep learning edge computing system framework comprises the following steps: S1, weight quantification of a deep learning network and optimization of a model network structure and parameters; s2, further refining convolution operation and acceleration training operation; s3, constructing a distributed deep neural network training model; s4, constructing a distributed deep neural network reasoning model; according to the invention, an edge computing system framework module comprising a model network structure, parameter optimization, parameter quantification, distributed training and distributed reasoning is creatively provided; data acquisition and data training of edge equipment are efficiently realized, and a low-delay calculation target is realized in a low-bandwidth and weak-calculation-power scene; and furthermore, real-time, effective, accurate and high-quality underlying suppo