CMSIS NN

CMSIS NN software library is a collection of efficient neural network kernels developed to maximize the performance and minimize the memory footprint of neural networks on Arm Cortex-M processors.

Supported Framework

The library follows the int8 and int16 quantization specification of TensorFlow Lite for Microcontrollers. This means CMSIS-NN is bit-exact with Tensorflow Lite reference kernels. In some cases TFL and TFLM reference kernels may not be bit-exact. In that case CMSIS-NN follows TFLM reference kernels. The unit test readme provides an overview.

Experimental Float API

CMSIS-NN also provides experimental float32 and float16 APIs. The float API intentionally follows the same CMSIS-NN integer style, which is itself shaped by TFLM integration patterns, to keep the public surface consistent across data types. This includes float16 even though TFLM does not define a float16 operator contract.

The float API is primarily intended for Cortex-M CPUs with Arm Helium Technology (MVE).…