A configurable analog neural memory system for deep learning neural networks

摘要

A number of embodiments are disclosed for a configurable hardware system for use in an analog neural memory system for a deep learning neural network. Components within a configurable configurable hardware system may include vector matrix multiplication arrays, summer circuits, activation circuits, inputs, reference devices, neurons, and test circuits. These devices can be configured to provide vector matrix multiplication arrays of various layers or various sizes, allowing the same hardware to be used in analog neural memory systems with different requirements.