A Taste of Open Source Neural Network Frameworks on Cell Phones

Let's Read the Source Code

IB503

08/11 15:30 - 16:00

臺語 / Taiwanese

Skilled / 中階

Android/iOS developers who want to put "AI" to edge devices.

AI or more precisely neural network-based machine learning is getting popular on edge devices. In this talk I'll review important open source NN frameworks, namely TensorFlow Lite + Android NN and Caffe2. Then I'll walk the audience through important source code of TenorFlow Lite and Caffe 2. For TensorFlow Lite, the interpreter and sample command line label_image will be covered. The talk will be a heavily modified one of [1] and [2], a much more code-oriented one.

[1] https://www.slideshare.net/kstan2/introduction-to-tensorflow-lite
[2] https://www.slideshare.net/kstan2/caffe2-on-android


Collaborative notes on HackMD

"freedom" Koan-Sin Tan

Koan-Sin Tan is a veteran open source user who learned to use Unix on VAX-11/780 running 4.3BSD. He was interested in making NN models working cell phone recently. He wrote the label_image for TFLite, the first non-Google C/C++ command line example for TensorFlow Lite.

Titanium Sponsor

Co-Host Sponsor

Diamond Sponsor

Gold Sponsor

Silver Sponsor

Bronze Sponsor

Partner

Co-Organizer

Special Thanks