Frameworks supported by Arm

Arm supports a wide array of machine learning frameworks. If you have requirements that are not yet supported by our fully optimized stack, you can choose from a range of frameworks and libraries.

TensorFlow

Type: Framework

Google framework with optimized 32-bit and 8-bit NEON routines for Arm CPUs.
Includes an OpenCL backend supporting Arm Mali GPU via SYCL.

TensorFlow Arm setup guide

Learn more

PyTorch

Type: Framework

Facebook framework with optimized routines for Arm CPUs with Neon.



Learn more

MxNet

Type: Framework

Apache framework with optimized routines for Arm CPUs.
Open AI version optimized with the Arm Compute Library for CPUs and GPUs.

MxNet GitHub

Learn more

Microsoft Cognitive Toolkit

Type: Framework

Microsoft framework with optimized routines for Arm CPUs with Neon.



Learn more

Caffe

Type: Framework

Berkley Vision and Learning Centre framework with several options for running on Arm.
CaffeOnACL is a version optimized with the Arm Compute Library for CPUs and GPUs.

CaffeOnACL GitHub

Learn more

Keras

Type: API

A high-level neural networks API, written in Python and capable of running on top of several framework backends including TensorFlow and Microsoft Cognitive Toolkit (CNTK).

Learn more

PlaidML

Type: Keras backend

Vertex.ai OpenCL framework compatible with Keras API with support for Arm Mali GPU.

Learn more

Arm SDKs

Arm software products maximize the performance of your machine learning applications.

Keil MDK

Arm Keil MDK is the most comprehensive software development solution for Arm-based microcontrollers and includes all components that you need to create, build, and debug embedded applications.

Learn more

Application development software

Arm NN

Arm NN is the Arm inference engine. Arm NN is designed to optimally run networks that are trained on popular frameworks, like TensorFlow and Caffe, on Arm IP.

Learn more

Arm Compute Library

Arm Compute Library optimizes low-level functions for computer vision and machine learning. Arm Compute Library focuses on Convolutional Neural Networks for 32-bit float and 9-bit integers across an array of Arm CPUs and GPUs.

Learn more

Embedded development software

CMSIS-NN

CMSIS-NN is the Arm library of efficient neural network kernels for Arm Cortex-M CPUs.

Learn more

DSP extensions for Arm

DSP extensions for Cortex-M

Cortex-M processors with DSP provide a high level of signal processing and integer performance, while maintaining the energy-efficiency and ease-of-use hallmarks of the Cortex-M family.

Learn more

DSP extensions for Cortex-R

This instruction set for Cortex-R processors includes enhanced DSP instructions that improve execution performance for arithmetic operations.

Learn more

Get Support


Community Forums

Not answered Tensorflow/Pytorch with GPU on ARM64
  • AArch64
  • TensorFlow
  • gpu
  • python
0 votes 66 views 0 replies Started 2 days ago by Maciek Answer this
Suggested answer Example cpp running onnx on arm NN 0 votes 1358 views 1 replies Latest 2 days ago by Jermey Answer this
Suggested answer NPU requirements 0 votes 208 views 1 replies Latest 3 days ago by Ben Clark Answer this
Answered How to make Ethos-U NPU work on an ARM Cortex-A + Cortex-M processor? 0 votes 1909 views 13 replies Latest 9 days ago by Kristofer Jonsson Answer this
Not answered Tensorflow/Pytorch with GPU on ARM64 Started 2 days ago by Maciek 0 replies 66 views
Suggested answer Example cpp running onnx on arm NN Latest 2 days ago by Jermey 1 replies 1358 views
Suggested answer NPU requirements Latest 3 days ago by Ben Clark 1 replies 208 views
Answered How to make Ethos-U NPU work on an ARM Cortex-A + Cortex-M processor? Latest 9 days ago by Kristofer Jonsson 13 replies 1909 views