Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
TensorFlow for Mobile Developers
Search
Enrique López Mañas
February 22, 2018
Programming
0
140
TensorFlow for Mobile Developers
TensorFlow for Mobile Developers @ DroidKaigi
Enrique López Mañas
February 22, 2018
Tweet
Share
More Decks by Enrique López Mañas
See All by Enrique López Mañas
KMP_for_Swift.pdf
kikoso
0
38
Android Benchmarking and other stories
kikoso
0
120
KMP for Mobile Developers
kikoso
0
60
Diving into Coroutines
kikoso
1
61
K/N for mobile developers (including libraries' snippets)
kikoso
0
65
Kotlin/Native for Multiplatform development
kikoso
0
60
TensorFlow for Mobile Developers
kikoso
0
21
Kotlin/Native for Multiplatform development
kikoso
1
180
Digging into authentications and permissions in Firebase
kikoso
2
560
Other Decks in Programming
See All in Programming
Асинхронность неизбежна: как мы проектировали сервис уведомлений
lamodatech
0
1.3k
見えないメモリを観測する: PHP 8.4 `pg_result_memory_size()` とSQL結果のメモリ管理
kentaroutakeda
0
930
Beyond ORM
77web
11
1.6k
カンファレンス動画鑑賞会のススメ / Osaka.swift #1
hironytic
0
170
テストコード書いてみませんか?
onopon
2
340
ASP.NET Core の OpenAPIサポート
h455h1
0
110
DevFest - Serverless 101 with Google Cloud Functions
tunmise
0
140
functionalなアプローチで動的要素を排除する
ryopeko
1
190
ecspresso, ecschedule, lambroll を PipeCDプラグインとして動かしてみた (プロトタイプ) / Running ecspresso, ecschedule, and lambroll as PipeCD Plugins (prototype)
tkikuc
2
1.8k
BEエンジニアがFEの業務をできるようになるまでにやったこと
yoshida_ryushin
0
190
PHPカンファレンス 2024|共創を加速するための若手の技術挑戦
weddingpark
0
140
「とりあえず動く」コードはよい、「読みやすい」コードはもっとよい / Code that 'just works' is good, but code that is 'readable' is even better.
mkmk884
6
1.4k
Featured
See All Featured
Writing Fast Ruby
sferik
628
61k
Fantastic passwords and where to find them - at NoRuKo
philnash
50
2.9k
StorybookのUI Testing Handbookを読んだ
zakiyama
28
5.4k
The MySQL Ecosystem @ GitHub 2015
samlambert
250
12k
Cheating the UX When There Is Nothing More to Optimize - PixelPioneers
stephaniewalter
280
13k
How to train your dragon (web standard)
notwaldorf
89
5.8k
Helping Users Find Their Own Way: Creating Modern Search Experiences
danielanewman
29
2.4k
Automating Front-end Workflow
addyosmani
1366
200k
Docker and Python
trallard
43
3.2k
ピンチをチャンスに:未来をつくるプロダクトロードマップ #pmconf2020
aki_iinuma
113
50k
Why You Should Never Use an ORM
jnunemaker
PRO
54
9.1k
Creating an realtime collaboration tool: Agile Flush - .NET Oxford
marcduiker
26
1.9k
Transcript
TensorFlow for Mobile Developers Enrique López Mañas Google Developer Expert
Ego Slide • Freelance Dev • Google Developer Expert •
@eenriquelopez
Machine Learning / AI
Machine Learning / AI
Machine Learning / AI
Buzzwords classification Artificial intelligence Machine Learning The science of making
machines smarts Building machines that can learn Neural Networks One of the many different algorithms in Machine Learning
Buzzwords classification
Neural Network
Machine Learning
Machine Learning
Machine Learning 12% of all responses on mobile
Machine Learning
Machine Learning
Machine Learning
Cloud vs. Mobile Less Traffic & Faster Response Motion Sensors
TensorFlow OpenSource library for Machine Learning tensorflow.org Today most popular
ML framework
TensorFlow You can train: - Mac / Windows - GPU
Server - GPU/TPU on Cloud Prediction: - Android - iOS - Raspberry
TensorFlow CIFAR10
TensorFlow ecosystem
From Training to App (@yufengg)
Data Gathering?
From Training to App (@yufengg)
From Training to App (@yufengg)
Convolutional Neural Networks (@yufengg)
Convolutional Neural Networks (@yufengg)
Convolutional Neural Networks (@yufengg)
Convolutional Neural Networks (@yufengg)
Convolutional Neural Networks (@yufengg)
Convolutional Neural Networks (@yufengg)
Convolutional Neural Networks (@yufengg)
Convolutional Neural Networks (@yufengg)
From Training to App (@yufengg)
Optimizing model for Mobile (@yufengg)
From Training to App (@yufengg) Other inception versions (inception v1
quantised is 7 MB)
From Training to App (@yufengg) TensorFlow increases APK in 12
MB
From Training to App (@yufengg)
From Training to App (@yufengg)
TensorFlow community 1000+ contributors 22.000+ commits 18.000+ repositories with name
“TensorFlow”
TensorFlow community
Building with Android Studio Uses Bazel to build Set Bazel
binary location in /tensorflow/examples/ android/build.gradle Add project in tensorflor/examples/android folder to Android Studio
Building with Android Studio Installing Bazel first: Download from https://github.com/bazelbuild/bazel/releases
Install instructions in https://docs.bazel.build/versions/master/install.html
Building with Android Studio Download Android SDK (23) Download Android
NDK (12b) (recommend it with Android SDK Manager)
Building with Android Studio WORKSPACE FILE
Building with Android Studio Run Bazel bazel build -c opt
//tensorflow/examples/ android:tensorflow_demo
Building with Android Studio Install APK adb install -r bazel-bin/tensorflow/examples/android/
tensorflow_demo.apk
Building with Android Studio OR… TensorFlow AAR from JCenter YUV
-> RGB less efficient Object tracking not available
Android Samples TF Classify TF Detect TF Stylize
TF Classify Uses Google Inception (v3) to label images Model
easy to swap No “person label” Volume button up for statistics
TF Classify
TF Classify
TF Detect
TF Detect - Draw bounding boxes around people - Useful
to count objects - No training yet
TF Style
TF Style - Real time style transfer algorithm - Pick/mix
different styles - Can train Magenta models
TF Style
TensorFlow Written in C++ Android uses Kotlin/Java? HowTo? Android
inference Library https://github.com/tensorflow/tensorflow/tree/master/tensorflow/contrib/android
Building for iOS Install requirements: - Xcode 8 - Command
line tools (xcode-select install) - brew install automake - brew install lib tool - tensorflow/contrib/makefile_build_all_ios.sh (takes 20 minutes)
iOS Examples: simple
iOS Examples: camera
iOS Examples: camera - Run inception each frame - Models
can be replaced
iOS Examples: benchmark
Raspberry Pi
Resources TensorFlow: https://www.tensorflow.org/ Magenta: https://magenta.tensorflow.org/welcome-to-magenta TensorFlow for Poets: https://codelabs.developers.google.com/codelabs/ tensorflow-for-poets/index.html#0
Shameless self-promotion www.kotlinweekly.net
Feedback! http://bit.ly/droidkaigi
None