Blog‎ > ‎

Machine learning frameworks that have been used on edge devices

posted Nov 15, 2020, 1:53 AM by MUHAMMAD MUN`IM AHMAD ZABIDI   [ updated Nov 15, 2020, 1:55 AM ]
Table 3 from M. G. S. Murshed, C. Murphy, D. Hou, N. Khan, G. Ananthanarayanan, and F. Hussain, “Machine Learning at the Network Edge: A Survey,” arXiv Prepr. arXiv1908.00080, pp. 1–28, 2019.

Framework

Core language

Interface

Part running on the edge

Example applications

TensorFlow Lite (Google)

C++

Java

C/C++

Java

TensorFlow Lite NN API

computer vision [109],

speech recognition [42, 1]

Caffe2

Caffe2Go (Facebook)

C++

Android iOs

NNPack

image analysis, video analysis [53]

Apache MXNet

C++

Python R

Linux

MacOS Windows

Full Model

object detection, recognition [78]

Core ML2 (Apple)

Python

iOS

CoreML

image analysis [16]

NLP [105]

ML Kit (Google)

C++

Java

Android iOs

Full Model

image recognition,

text recognition, bar-code scaning [26]

AI2GO

C, Python Java, Swift

Linux macOs

Full Model

object detection, classification [5]

DeepThings

C/C++

Linux

Full Model

object detection [119]

DeepIoT

Python

Ubilinux

Full Model

human activity

recognition,

user identification [116]

DeepCham

C++

Java

Linux Android

Full Model

object recognition [62]

SparseSep

-

Linux Android

Full Model

mobile object

recognition, audio classification [15]

Edgent

-

Ubuntu

Major part of the DNN

image recognition [63]


17

Comments