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
Turi Create and Concept Behind Image Classifica...
Search
Merocode
February 14, 2018
Programming
1
140
Turi Create and Concept Behind Image Classification
CocoaHead Berlin
Merocode
February 14, 2018
Tweet
Share
More Decks by Merocode
See All by Merocode
Core ML Overview
merocode
0
230
Other Decks in Programming
See All in Programming
Design Foundational Data Engineering Observability
sucitw
3
200
複雑なドメインに挑む.pdf
yukisakai1225
5
1.2k
アプリの "かわいい" を支えるアニメーションツールRiveについて
uetyo
0
270
Updates on MLS on Ruby (and maybe more)
sylph01
1
180
rage against annotate_predecessor
junk0612
0
170
🔨 小さなビルドシステムを作る
momeemt
4
680
CJK and Unicode From a PHP Committer
youkidearitai
PRO
0
110
そのAPI、誰のため? Androidライブラリ設計における利用者目線の実践テクニック
mkeeda
2
310
はじめてのMaterial3 Expressive
ym223
2
740
速いWebフレームワークを作る
yusukebe
5
1.7k
ProxyによるWindow間RPC機構の構築
syumai
3
1.2k
奥深くて厄介な「改行」と仲良くなる20分
oguemon
1
540
Featured
See All Featured
GraphQLとの向き合い方2022年版
quramy
49
14k
The Language of Interfaces
destraynor
161
25k
What’s in a name? Adding method to the madness
productmarketing
PRO
23
3.7k
[RailsConf 2023] Rails as a piece of cake
palkan
57
5.8k
Optimizing for Happiness
mojombo
379
70k
Helping Users Find Their Own Way: Creating Modern Search Experiences
danielanewman
29
2.9k
Designing for humans not robots
tammielis
253
25k
What's in a price? How to price your products and services
michaelherold
246
12k
Dealing with People You Can't Stand - Big Design 2015
cassininazir
367
27k
10 Git Anti Patterns You Should be Aware of
lemiorhan
PRO
656
61k
Art, The Web, and Tiny UX
lynnandtonic
303
21k
How to train your dragon (web standard)
notwaldorf
96
6.2k
Transcript
TURI CREATE CONCEPT BEHIND IMAGE CLASSIFICATION
WHAT WE WANT TO ACHIEVE?
None
None
CORE ML
INFERENCE
None
FROM WHERE WE GET THE CORE ML FORMAT MODELS?
> Apple ready-to-use Core ML Models.
> Apple tool (Python Package) to convert known model formats
to Core ML model format. That is covered in that blog post
import coremltools caffe_model = ('EmotiW_VGG_S.caffemodel', 'deploy.prototxt') labels = 'labels.txt' coreml_model
= coremltools.converters.caffe.convert(caffe_model, class_labels=labels, image_input_names='data') coreml_model.save('EmotiW_VGG_S.mlmodel')
WHAT IS A MODEL?
TRAINING
TURI CREATE
> Under the hood of Turi Create is Apache’s MXNet
> Python 2.7, 3.5, 3.6 conda create -n turi python=3.6 source activate turi pip install -U turicreate import turicreate fails on macOS 10.12.6 conda create -n turi python=2.7 pip install -Iv turicreate==4.0
CONVOLUTIONAL NEURAL NETWORKS CNN MODEL Feature extraction part + Classification
part
None
None
None
None
None
None
None
VISUALIZING FEATURE / ACTIVATION MAPS
TRANSFER LEARNING
None
CREATE AND TRAIN A MODEL USING TURI CREATE
SQUEEZENET_V1.1
None
None
None
None
None
15 min.
44 min.
ResNet50
9 min.
17 min.
66 min.
IMAGE AUGMENTATION Invariant Representation Scale Rotation Translation
None
RESOURCES CODE Turi Create Sample Code and Demo Apps
RESOURCES TOOLS Turi Create Anaconda Miniconda
RESOURCES APPLE Turi Create User Guide Turi Create API Documentation
Sample Code: Classifying Images with Vision and Core ML Apple Machine Learning
RESOURCES ARTICLES Convolutional Neural Networks Image Augmentation in Keras
RESOURCES PHOTOS Photo by Gift Habeshaw on Unsplash Elon Musk
SpaceX launch footage
THANK YOU ❤ @_MeroCode_ LET’S LEARN ABOUT MACHINE LEARNING.