본문 바로가기
Machine Learning

Google I/O Korea 2017

by swconsulting swconsulting 2017. 7. 3.


I think important key word.


TPU / Cloud TPU


TensorFlow 1.3 (every 6weeks update <- It is really quick update.)


Keras API (easy tensorflow ?)


XLA Compiler (TF support java ?)


TensorFlow Serving -> Model save/export used by protocol buffer


Android O =>TensorFlow Lite (It will include tensorflow library)


Google Lens -> Image matching...


Kotlin -> Android Studio 3.0 support -> Google officially support it

(Lambda,etc)


DEV OPS -> NoOPS


WAS  -> Normally 1 seconds 1500 connections

       -> Good development max 1 seconds 5000 connections

       -> Cloud no limit 


[Data Pipeline]

Machine Learning ( Deep Learning ) -> Data collect most important -> Easy to access massive data <- It is very important.


Machine Learning => IDEA + PLUS(Business Knowledge) 



[Firebase -> Very useful]



'Machine Learning' 카테고리의 다른 글

Download windows virtual machine Images.  (0) 2019.05.31
Google I/O Korea 2017  (0) 2017.07.03
Create a bot with the Bot Builder SDK for .NET  (0) 2017.07.01

댓글0