Introduction

This user manual describes the CMSIS NN software library, a collection of efficient neural network kernels developed to maximize the performance and minimize the memory footprint of neural networks on Cortex-M processor cores.

The library is divided into a number of functions each covering a specific category:

Neural Network Convolution Functions
Neural Network Activation Functions
Fully-connected Layer Functions
Neural Network Pooling Functions
Softmax Functions
Neural Network Support Functions

The library has separate functions for operating on different weight and activation data types including 8-bit integers (q7_t) and 16-bit integers (q15_t). The descrition of the kernels are included in the function description. The implementation details are also described in this paper [1].

Block Diagram

Examples

The library ships with a number of examples which demonstrate how to use the library functions.

Pre-processor Macros

Each library project have differant pre-processor macros.

ARM_MATH_DSP:

Define macro ARM_MATH_DSP, If the silicon supports DSP instructions.

ARM_MATH_BIG_ENDIAN:

Define macro ARM_MATH_BIG_ENDIAN to build the library for big endian targets. By default library builds for little endian targets.

ARM_NN_TRUNCATE:

Define macro ARM_NN_TRUNCATE to use floor instead of round-to-the-nearest-int for the computation.

Copyright Notice

[1] CMSIS-NN: Efficient Neural Network Kernels for Arm Cortex-M CPUs https://arxiv.org/abs/1801.06601