SlideShare a Scribd company logo
1 of 22
© 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Amazon SageMaker
Develop, Train, Tune and Deploy Machine Learning Models
2019-04-24, D2SI Toulouse
Olivier Cruchant, ML Specialist SA, AWS
© 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Put machine learning in the hands of every developer
and data scientist
ML @ AWS: Our mission
© 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Customer Running ML on AWS Today
© 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
AWS ML Stack
P L A T F O R M S
A P P L I C A T I O N
S E R V I C E S
F R A M E W O R K S E T
I N F R A S T R U C T U R E
Amazon SageMaker
Vision
Rekognition Image
Rekognition Video
Amazon Textract
Speech
Amazon Polly
Amazon Transcribe
Language
Amazon Lex
Amazon Translate
Amazon Comprehend
Forecasting
Amazon Forecast
Personalization
Amazon
Personalize
DL AMI EC2 P3, C5, F1 GreengrassElastic Inference
© 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Build, Train and Deploy your Machine Learning Models
Amazon SageMaker
Models are abstracted into docker containers
Fast & accurate
data labeling
Built-in, high
performance
algorithms &
notebooks
BUILD
1
One-click
training and
tuning
TRAI N
Model tuning and
optimization
2
Fully managed
hosting with
auto-scaling and
elastic inference
One-click
Deployment of
model or Inference
Pipeline
DEPLOY
3
© 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Amazon SageMaker: TensorFlow workflow example
Develop on Notebook instance
with TensorFlow Jupyter kernel
BUILD
1
Train custom
code in
TensorFlow
container
TRAI N
Model tuning and
optimization
2
Deploy to auto-scaled
tensorflow-serving
container
DEPLOY
3
Stream training
data with
PipeModeDataset
Distributed training
with Horovod
Monitor with
TensorBoard
© 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Amazon SageMaker
Deployment & Inference
A. Deploy model to highly-available Amazon SageMaker HTTPS endpoint
B. Run Amazon SageMaker batch-transform job
C. On-device deployment (AWS Greengrass)
© 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Amazon SageMaker: algorithms
A. Train and deploy a built-in algorithm
B. Train and deploy your own code in a framework container
C. Train and deploy your own container
D. Bring a pre-trained model
E. Train and Deploy a Marketplace Algorithm
F. Deploy a Marketplace Model Package
© 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Amazon SageMaker: interface
A. Console
B. AWS Command Line Interface (CLI):
aws sagemaker create-endpoint
--endpoint-name <value>
--endpoint-config-name <value>
C. SageMaker Python SDK
model.deploy(initial_instance_count=1,
instance_type='ml.m4.xlarge')
D. Boto3 (python) SageMaker client
client.create_endpoint(EndpointName='string',
EndpointConfigName='string')
E. Other SDKs
© 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
In summary:
Python SDK: Amazon Algorithm
Python SDK: Deep Learning Estimator
© 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Amazon SageMaker
Built-in algorithms
Supervised
Linear Learner
Factorization Machines
XGBoost
Image Classification
Object Detection
Sequence to Sequence
DeepAR Forecasting
K-Nearest Neighbors
BlazingText (classifier)
Object2Vec
Semantic Segmentation
Unsupervised
K-Means
Principal Component Analysis
Latent Dirichlet Allocation
Neural Topic Model
Random Cut Forest
BlazingText (word2vec)
IP Insights
© 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Object Detection
A. Single-Shot multi-box Detector (SSD): learns to locate and classify at the same time
B. Extract features with a convolutional base (VGG-16 and ResNet-50 available)
C. Transfer learning from ImageNet available
D. Example: aerial imagery analysis
© 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
DeepAR
A. Forecasts scalar time series with RNN
B. Adapted to forecast families of similar time series (product sales, energy by household)
C. Learning set can include arbitrary categorical and time series features
D. Automatically derives features based on target frequency (eg day of month, day of year)
E. Automatically feeds lagged values from target time series
https://docs.aws.amazon.com/sagemaker/latest/dg/deepar_how-it-works.html
© 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
BlazingText Word2Vec
A. Goal: produce low-dimension, dense semantic word
embeddings
B. CUDA implementation of word2vec that can be
distributed over multiple GPUs
C. Byproduct of a text classification problem (CBOW or
Skip-gram)
1
0
…
…
…
…
…
0
n1
n300
0
0
…
…
…
1
0
0
…
Apple Juice
wn1, 1
Voc size:
40k
wn1, 2
wn1, 40k
wn300, 1
wn300, 2
wn300, 40k
Word e1 … e300
Apple wn1, 1 … wn300, 1
… … … …
Zebra … … …
⟹
Simplified architecture. See original papers for more details
https://papers.nips.cc/paper/5021-distributed-representations-of-words-and-phrases-
and-their-compositionality.pdf
https://arxiv.org/pdf/1301.3781.pdf
https://arxiv.org/pdf/1607.04606.pdf
© 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
BlazingText Classifier
A. Extends fastText to work on GPU
B. Learns word embeddings that are averaged then linearly fit
C. Features include n-grams
D. Features are hashed to limit dimensionality
https://aws.amazon.com/fr/blogs/machine-learning/enhanced-text-classification-and-word-
vectors-using-amazon-sagemaker-blazingtext/
© 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Object2Vec
A. Generate predictions or embeddings
B. Use-cases: recommendation, similarity,
multi-label classification
C. Supports 2 types of inputs:
• Discrete tokens
• Sequences of discrete tokens
© 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Random Cut Forest
Algo:
1. Partition input space into random sample
2. For each sample:
i. Draw bounding box
ii. Random cut on a random feature (sampling proportional
to variance)
iii. Points alone in their bounding boxes become tree leaves
3. Outlier score inversely proportional to leaf depth
… …
© 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
SageMaker Ground Truth
© 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
SageMaker Neo
Compiler Runtime
© 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
https://aws.amazon.com/sagemaker
http://d2l.ai/
Getting Started

More Related Content

What's hot

Optimize your machine learning workloads on AWS (March 2019)
Optimize your machine learning workloads on AWS (March 2019)Optimize your machine learning workloads on AWS (March 2019)
Optimize your machine learning workloads on AWS (March 2019)Julien SIMON
 
Frontend and Mobile with AWS Amplify | AWS Summit Tel Aviv 2019
Frontend and Mobile with AWS Amplify | AWS Summit Tel Aviv 2019Frontend and Mobile with AWS Amplify | AWS Summit Tel Aviv 2019
Frontend and Mobile with AWS Amplify | AWS Summit Tel Aviv 2019AWS Summits
 
[REPEAT] Get hands on with AWS DeepRacer & compete in the AWS DeepRacer Leagu...
[REPEAT] Get hands on with AWS DeepRacer & compete in the AWS DeepRacer Leagu...[REPEAT] Get hands on with AWS DeepRacer & compete in the AWS DeepRacer Leagu...
[REPEAT] Get hands on with AWS DeepRacer & compete in the AWS DeepRacer Leagu...Amazon Web Services
 
Increase the value of video using ML and AWS media services - SVC301 - Atlant...
Increase the value of video using ML and AWS media services - SVC301 - Atlant...Increase the value of video using ML and AWS media services - SVC301 - Atlant...
Increase the value of video using ML and AWS media services - SVC301 - Atlant...Amazon Web Services
 
Building Machine Learning inference pipelines at scale | AWS Summit Tel Aviv ...
Building Machine Learning inference pipelines at scale | AWS Summit Tel Aviv ...Building Machine Learning inference pipelines at scale | AWS Summit Tel Aviv ...
Building Machine Learning inference pipelines at scale | AWS Summit Tel Aviv ...AWS Summits
 
Unleash the Power of ML with AWS | AWS Summit Tel Aviv 2019
Unleash the Power of ML with AWS | AWS Summit Tel Aviv 2019Unleash the Power of ML with AWS | AWS Summit Tel Aviv 2019
Unleash the Power of ML with AWS | AWS Summit Tel Aviv 2019AWS Summits
 
Amazon SageMaker - ML for every developer & data scientist ft. Workday - AIM2...
Amazon SageMaker - ML for every developer & data scientist ft. Workday - AIM2...Amazon SageMaker - ML for every developer & data scientist ft. Workday - AIM2...
Amazon SageMaker - ML for every developer & data scientist ft. Workday - AIM2...Amazon Web Services
 
Artifical Intelligence and Machine Learning 201, AWS Federal Pop-Up Loft
Artifical Intelligence and Machine Learning 201, AWS Federal Pop-Up LoftArtifical Intelligence and Machine Learning 201, AWS Federal Pop-Up Loft
Artifical Intelligence and Machine Learning 201, AWS Federal Pop-Up LoftAmazon Web Services
 
Amazon Aurora, funzionalità e best practice per la migrazione di database su AWS
Amazon Aurora, funzionalità e best practice per la migrazione di database su AWSAmazon Aurora, funzionalità e best practice per la migrazione di database su AWS
Amazon Aurora, funzionalità e best practice per la migrazione di database su AWSAmazon Web Services
 
Sicurezza in AWS automazione e best practice
Sicurezza in AWS automazione e best practiceSicurezza in AWS automazione e best practice
Sicurezza in AWS automazione e best practiceAmazon Web Services
 
Deriving Value with Next Gen Analytics and ML Architectures
Deriving Value with Next Gen Analytics and ML ArchitecturesDeriving Value with Next Gen Analytics and ML Architectures
Deriving Value with Next Gen Analytics and ML ArchitecturesAmazon Web Services
 
Build, train and deploy Machine Learning models on Amazon SageMaker (May 2019)
Build, train and deploy Machine Learning models on Amazon SageMaker (May 2019)Build, train and deploy Machine Learning models on Amazon SageMaker (May 2019)
Build, train and deploy Machine Learning models on Amazon SageMaker (May 2019)Julien SIMON
 
Microservices on AWS: Architectural Patterns and Best Practices | AWS Summit ...
Microservices on AWS: Architectural Patterns and Best Practices | AWS Summit ...Microservices on AWS: Architectural Patterns and Best Practices | AWS Summit ...
Microservices on AWS: Architectural Patterns and Best Practices | AWS Summit ...AWS Summits
 
Machine learning for developers & data scientists with Amazon SageMaker - AIM...
Machine learning for developers & data scientists with Amazon SageMaker - AIM...Machine learning for developers & data scientists with Amazon SageMaker - AIM...
Machine learning for developers & data scientists with Amazon SageMaker - AIM...Amazon Web Services
 
Machine Learning on AWS (December 2018)
Machine Learning on AWS (December 2018)Machine Learning on AWS (December 2018)
Machine Learning on AWS (December 2018)Julien SIMON
 
Build, Train and Deploy Machine Learning Models at Scale (April 2019)
Build, Train and Deploy Machine Learning Models at Scale (April 2019)Build, Train and Deploy Machine Learning Models at Scale (April 2019)
Build, Train and Deploy Machine Learning Models at Scale (April 2019)Julien SIMON
 
Amazon SageMaker sviluppa, addestra e distribuisci modelli di Machine Learnin...
Amazon SageMaker sviluppa, addestra e distribuisci modelli di Machine Learnin...Amazon SageMaker sviluppa, addestra e distribuisci modelli di Machine Learnin...
Amazon SageMaker sviluppa, addestra e distribuisci modelli di Machine Learnin...Amazon Web Services
 
Building machine learning inference pipelines at scale (March 2019)
Building machine learning inference pipelines at scale (March 2019)Building machine learning inference pipelines at scale (March 2019)
Building machine learning inference pipelines at scale (March 2019)Julien SIMON
 
Build a VR experience in 60 minutes - SVC222 - New York AWS Summit
Build a VR experience in 60 minutes - SVC222 - New York AWS SummitBuild a VR experience in 60 minutes - SVC222 - New York AWS Summit
Build a VR experience in 60 minutes - SVC222 - New York AWS SummitAmazon Web Services
 
Need for Speed – Intro To Real-Time Data Streaming Analytics on AWS | AWS Sum...
Need for Speed – Intro To Real-Time Data Streaming Analytics on AWS | AWS Sum...Need for Speed – Intro To Real-Time Data Streaming Analytics on AWS | AWS Sum...
Need for Speed – Intro To Real-Time Data Streaming Analytics on AWS | AWS Sum...AWS Summits
 

What's hot (20)

Optimize your machine learning workloads on AWS (March 2019)
Optimize your machine learning workloads on AWS (March 2019)Optimize your machine learning workloads on AWS (March 2019)
Optimize your machine learning workloads on AWS (March 2019)
 
Frontend and Mobile with AWS Amplify | AWS Summit Tel Aviv 2019
Frontend and Mobile with AWS Amplify | AWS Summit Tel Aviv 2019Frontend and Mobile with AWS Amplify | AWS Summit Tel Aviv 2019
Frontend and Mobile with AWS Amplify | AWS Summit Tel Aviv 2019
 
[REPEAT] Get hands on with AWS DeepRacer & compete in the AWS DeepRacer Leagu...
[REPEAT] Get hands on with AWS DeepRacer & compete in the AWS DeepRacer Leagu...[REPEAT] Get hands on with AWS DeepRacer & compete in the AWS DeepRacer Leagu...
[REPEAT] Get hands on with AWS DeepRacer & compete in the AWS DeepRacer Leagu...
 
Increase the value of video using ML and AWS media services - SVC301 - Atlant...
Increase the value of video using ML and AWS media services - SVC301 - Atlant...Increase the value of video using ML and AWS media services - SVC301 - Atlant...
Increase the value of video using ML and AWS media services - SVC301 - Atlant...
 
Building Machine Learning inference pipelines at scale | AWS Summit Tel Aviv ...
Building Machine Learning inference pipelines at scale | AWS Summit Tel Aviv ...Building Machine Learning inference pipelines at scale | AWS Summit Tel Aviv ...
Building Machine Learning inference pipelines at scale | AWS Summit Tel Aviv ...
 
Unleash the Power of ML with AWS | AWS Summit Tel Aviv 2019
Unleash the Power of ML with AWS | AWS Summit Tel Aviv 2019Unleash the Power of ML with AWS | AWS Summit Tel Aviv 2019
Unleash the Power of ML with AWS | AWS Summit Tel Aviv 2019
 
Amazon SageMaker - ML for every developer & data scientist ft. Workday - AIM2...
Amazon SageMaker - ML for every developer & data scientist ft. Workday - AIM2...Amazon SageMaker - ML for every developer & data scientist ft. Workday - AIM2...
Amazon SageMaker - ML for every developer & data scientist ft. Workday - AIM2...
 
Artifical Intelligence and Machine Learning 201, AWS Federal Pop-Up Loft
Artifical Intelligence and Machine Learning 201, AWS Federal Pop-Up LoftArtifical Intelligence and Machine Learning 201, AWS Federal Pop-Up Loft
Artifical Intelligence and Machine Learning 201, AWS Federal Pop-Up Loft
 
Amazon Aurora, funzionalità e best practice per la migrazione di database su AWS
Amazon Aurora, funzionalità e best practice per la migrazione di database su AWSAmazon Aurora, funzionalità e best practice per la migrazione di database su AWS
Amazon Aurora, funzionalità e best practice per la migrazione di database su AWS
 
Sicurezza in AWS automazione e best practice
Sicurezza in AWS automazione e best practiceSicurezza in AWS automazione e best practice
Sicurezza in AWS automazione e best practice
 
Deriving Value with Next Gen Analytics and ML Architectures
Deriving Value with Next Gen Analytics and ML ArchitecturesDeriving Value with Next Gen Analytics and ML Architectures
Deriving Value with Next Gen Analytics and ML Architectures
 
Build, train and deploy Machine Learning models on Amazon SageMaker (May 2019)
Build, train and deploy Machine Learning models on Amazon SageMaker (May 2019)Build, train and deploy Machine Learning models on Amazon SageMaker (May 2019)
Build, train and deploy Machine Learning models on Amazon SageMaker (May 2019)
 
Microservices on AWS: Architectural Patterns and Best Practices | AWS Summit ...
Microservices on AWS: Architectural Patterns and Best Practices | AWS Summit ...Microservices on AWS: Architectural Patterns and Best Practices | AWS Summit ...
Microservices on AWS: Architectural Patterns and Best Practices | AWS Summit ...
 
Machine learning for developers & data scientists with Amazon SageMaker - AIM...
Machine learning for developers & data scientists with Amazon SageMaker - AIM...Machine learning for developers & data scientists with Amazon SageMaker - AIM...
Machine learning for developers & data scientists with Amazon SageMaker - AIM...
 
Machine Learning on AWS (December 2018)
Machine Learning on AWS (December 2018)Machine Learning on AWS (December 2018)
Machine Learning on AWS (December 2018)
 
Build, Train and Deploy Machine Learning Models at Scale (April 2019)
Build, Train and Deploy Machine Learning Models at Scale (April 2019)Build, Train and Deploy Machine Learning Models at Scale (April 2019)
Build, Train and Deploy Machine Learning Models at Scale (April 2019)
 
Amazon SageMaker sviluppa, addestra e distribuisci modelli di Machine Learnin...
Amazon SageMaker sviluppa, addestra e distribuisci modelli di Machine Learnin...Amazon SageMaker sviluppa, addestra e distribuisci modelli di Machine Learnin...
Amazon SageMaker sviluppa, addestra e distribuisci modelli di Machine Learnin...
 
Building machine learning inference pipelines at scale (March 2019)
Building machine learning inference pipelines at scale (March 2019)Building machine learning inference pipelines at scale (March 2019)
Building machine learning inference pipelines at scale (March 2019)
 
Build a VR experience in 60 minutes - SVC222 - New York AWS Summit
Build a VR experience in 60 minutes - SVC222 - New York AWS SummitBuild a VR experience in 60 minutes - SVC222 - New York AWS Summit
Build a VR experience in 60 minutes - SVC222 - New York AWS Summit
 
Need for Speed – Intro To Real-Time Data Streaming Analytics on AWS | AWS Sum...
Need for Speed – Intro To Real-Time Data Streaming Analytics on AWS | AWS Sum...Need for Speed – Intro To Real-Time Data Streaming Analytics on AWS | AWS Sum...
Need for Speed – Intro To Real-Time Data Streaming Analytics on AWS | AWS Sum...
 

Similar to Amazon SageMaker Deep Dive - Meetup AWS Toulouse at D2SI

MXNet Paris Workshop - Intro To MXNet
MXNet Paris Workshop - Intro To MXNetMXNet Paris Workshop - Intro To MXNet
MXNet Paris Workshop - Intro To MXNetApache MXNet
 
Build, train and deploy ML models with SageMaker (October 2019)
Build, train and deploy ML models with SageMaker (October 2019)Build, train and deploy ML models with SageMaker (October 2019)
Build, train and deploy ML models with SageMaker (October 2019)Julien SIMON
 
[NEW LAUNCH] Introducing AWS Deep Learning Containers
[NEW LAUNCH] Introducing AWS Deep Learning Containers[NEW LAUNCH] Introducing AWS Deep Learning Containers
[NEW LAUNCH] Introducing AWS Deep Learning ContainersAmazon Web Services
 
Integrate Your Favourite Microsoft DevOps Tools with AWS - AWS Summit Sydney
Integrate Your Favourite Microsoft DevOps Tools with AWS - AWS Summit SydneyIntegrate Your Favourite Microsoft DevOps Tools with AWS - AWS Summit Sydney
Integrate Your Favourite Microsoft DevOps Tools with AWS - AWS Summit SydneyAmazon Web Services
 
AWS Summit Stockholm - Fargate: deploy containers, not infrastructure
AWS Summit Stockholm - Fargate: deploy containers, not infrastructureAWS Summit Stockholm - Fargate: deploy containers, not infrastructure
AWS Summit Stockholm - Fargate: deploy containers, not infrastructureMassimo Ferre'
 
Progetta, crea e gestisci Modern Application per web e mobile su AWS
Progetta, crea e gestisci Modern Application per web e mobile su AWSProgetta, crea e gestisci Modern Application per web e mobile su AWS
Progetta, crea e gestisci Modern Application per web e mobile su AWSAmazon Web Services
 
Optimize your Machine Learning workloads | AWS Summit Tel Aviv 2019
Optimize your Machine Learning workloads  | AWS Summit Tel Aviv 2019Optimize your Machine Learning workloads  | AWS Summit Tel Aviv 2019
Optimize your Machine Learning workloads | AWS Summit Tel Aviv 2019Amazon Web Services
 
Optimize your Machine Learning workloads | AWS Summit Tel Aviv 2019
Optimize your Machine Learning workloads  | AWS Summit Tel Aviv 2019Optimize your Machine Learning workloads  | AWS Summit Tel Aviv 2019
Optimize your Machine Learning workloads | AWS Summit Tel Aviv 2019AWS Summits
 
Breaking the Monolith Using AWS Container Services
Breaking the Monolith Using AWS Container ServicesBreaking the Monolith Using AWS Container Services
Breaking the Monolith Using AWS Container ServicesAmazon Web Services
 
Continuous Integration and Continuous Delivery for your serverless apps - Seb...
Continuous Integration and Continuous Delivery for your serverless apps - Seb...Continuous Integration and Continuous Delivery for your serverless apps - Seb...
Continuous Integration and Continuous Delivery for your serverless apps - Seb...Shift Conference
 
Modern-Application-Design-with-Amazon-ECS
Modern-Application-Design-with-Amazon-ECSModern-Application-Design-with-Amazon-ECS
Modern-Application-Design-with-Amazon-ECSAmazon Web Services
 
AWS DevDay Cologne - Automating building blocks choices you will face with co...
AWS DevDay Cologne - Automating building blocks choices you will face with co...AWS DevDay Cologne - Automating building blocks choices you will face with co...
AWS DevDay Cologne - Automating building blocks choices you will face with co...Cobus Bernard
 
Building a Critical Communications Platform Using Serverless Technologies
Building a Critical Communications Platform Using Serverless TechnologiesBuilding a Critical Communications Platform Using Serverless Technologies
Building a Critical Communications Platform Using Serverless TechnologiesAmazon Web Services
 
Simplifying Microsoft Architectures with AWS Services
Simplifying Microsoft Architectures with AWS ServicesSimplifying Microsoft Architectures with AWS Services
Simplifying Microsoft Architectures with AWS ServicesAWS Summits
 
AWS CloudFormation Deep Dive and Recent Enhancements
AWS CloudFormation Deep Dive and Recent EnhancementsAWS CloudFormation Deep Dive and Recent Enhancements
AWS CloudFormation Deep Dive and Recent EnhancementsAmazon Web Services
 
Deep Learning on Amazon SageMaker | AWS Floor28
Deep Learning on Amazon SageMaker | AWS Floor28Deep Learning on Amazon SageMaker | AWS Floor28
Deep Learning on Amazon SageMaker | AWS Floor28Amazon Web Services
 
AWS ReInvent 2023 Recap: AWS User GroupKolkata
AWS ReInvent 2023 Recap: AWS User GroupKolkataAWS ReInvent 2023 Recap: AWS User GroupKolkata
AWS ReInvent 2023 Recap: AWS User GroupKolkataAritra Nag
 
AWS reInvent 2023 re:Cap services Slide deck
AWS reInvent 2023 re:Cap services Slide deckAWS reInvent 2023 re:Cap services Slide deck
AWS reInvent 2023 re:Cap services Slide deckSammy Cheung
 
Machine Learning using Kubernetes - AI Conclave 2019
Machine Learning using Kubernetes - AI Conclave 2019Machine Learning using Kubernetes - AI Conclave 2019
Machine Learning using Kubernetes - AI Conclave 2019Arun Gupta
 
Breaking the Monolith road to containers.pdf
Breaking the Monolith road to containers.pdfBreaking the Monolith road to containers.pdf
Breaking the Monolith road to containers.pdfAmazon Web Services
 

Similar to Amazon SageMaker Deep Dive - Meetup AWS Toulouse at D2SI (20)

MXNet Paris Workshop - Intro To MXNet
MXNet Paris Workshop - Intro To MXNetMXNet Paris Workshop - Intro To MXNet
MXNet Paris Workshop - Intro To MXNet
 
Build, train and deploy ML models with SageMaker (October 2019)
Build, train and deploy ML models with SageMaker (October 2019)Build, train and deploy ML models with SageMaker (October 2019)
Build, train and deploy ML models with SageMaker (October 2019)
 
[NEW LAUNCH] Introducing AWS Deep Learning Containers
[NEW LAUNCH] Introducing AWS Deep Learning Containers[NEW LAUNCH] Introducing AWS Deep Learning Containers
[NEW LAUNCH] Introducing AWS Deep Learning Containers
 
Integrate Your Favourite Microsoft DevOps Tools with AWS - AWS Summit Sydney
Integrate Your Favourite Microsoft DevOps Tools with AWS - AWS Summit SydneyIntegrate Your Favourite Microsoft DevOps Tools with AWS - AWS Summit Sydney
Integrate Your Favourite Microsoft DevOps Tools with AWS - AWS Summit Sydney
 
AWS Summit Stockholm - Fargate: deploy containers, not infrastructure
AWS Summit Stockholm - Fargate: deploy containers, not infrastructureAWS Summit Stockholm - Fargate: deploy containers, not infrastructure
AWS Summit Stockholm - Fargate: deploy containers, not infrastructure
 
Progetta, crea e gestisci Modern Application per web e mobile su AWS
Progetta, crea e gestisci Modern Application per web e mobile su AWSProgetta, crea e gestisci Modern Application per web e mobile su AWS
Progetta, crea e gestisci Modern Application per web e mobile su AWS
 
Optimize your Machine Learning workloads | AWS Summit Tel Aviv 2019
Optimize your Machine Learning workloads  | AWS Summit Tel Aviv 2019Optimize your Machine Learning workloads  | AWS Summit Tel Aviv 2019
Optimize your Machine Learning workloads | AWS Summit Tel Aviv 2019
 
Optimize your Machine Learning workloads | AWS Summit Tel Aviv 2019
Optimize your Machine Learning workloads  | AWS Summit Tel Aviv 2019Optimize your Machine Learning workloads  | AWS Summit Tel Aviv 2019
Optimize your Machine Learning workloads | AWS Summit Tel Aviv 2019
 
Breaking the Monolith Using AWS Container Services
Breaking the Monolith Using AWS Container ServicesBreaking the Monolith Using AWS Container Services
Breaking the Monolith Using AWS Container Services
 
Continuous Integration and Continuous Delivery for your serverless apps - Seb...
Continuous Integration and Continuous Delivery for your serverless apps - Seb...Continuous Integration and Continuous Delivery for your serverless apps - Seb...
Continuous Integration and Continuous Delivery for your serverless apps - Seb...
 
Modern-Application-Design-with-Amazon-ECS
Modern-Application-Design-with-Amazon-ECSModern-Application-Design-with-Amazon-ECS
Modern-Application-Design-with-Amazon-ECS
 
AWS DevDay Cologne - Automating building blocks choices you will face with co...
AWS DevDay Cologne - Automating building blocks choices you will face with co...AWS DevDay Cologne - Automating building blocks choices you will face with co...
AWS DevDay Cologne - Automating building blocks choices you will face with co...
 
Building a Critical Communications Platform Using Serverless Technologies
Building a Critical Communications Platform Using Serverless TechnologiesBuilding a Critical Communications Platform Using Serverless Technologies
Building a Critical Communications Platform Using Serverless Technologies
 
Simplifying Microsoft Architectures with AWS Services
Simplifying Microsoft Architectures with AWS ServicesSimplifying Microsoft Architectures with AWS Services
Simplifying Microsoft Architectures with AWS Services
 
AWS CloudFormation Deep Dive and Recent Enhancements
AWS CloudFormation Deep Dive and Recent EnhancementsAWS CloudFormation Deep Dive and Recent Enhancements
AWS CloudFormation Deep Dive and Recent Enhancements
 
Deep Learning on Amazon SageMaker | AWS Floor28
Deep Learning on Amazon SageMaker | AWS Floor28Deep Learning on Amazon SageMaker | AWS Floor28
Deep Learning on Amazon SageMaker | AWS Floor28
 
AWS ReInvent 2023 Recap: AWS User GroupKolkata
AWS ReInvent 2023 Recap: AWS User GroupKolkataAWS ReInvent 2023 Recap: AWS User GroupKolkata
AWS ReInvent 2023 Recap: AWS User GroupKolkata
 
AWS reInvent 2023 re:Cap services Slide deck
AWS reInvent 2023 re:Cap services Slide deckAWS reInvent 2023 re:Cap services Slide deck
AWS reInvent 2023 re:Cap services Slide deck
 
Machine Learning using Kubernetes - AI Conclave 2019
Machine Learning using Kubernetes - AI Conclave 2019Machine Learning using Kubernetes - AI Conclave 2019
Machine Learning using Kubernetes - AI Conclave 2019
 
Breaking the Monolith road to containers.pdf
Breaking the Monolith road to containers.pdfBreaking the Monolith road to containers.pdf
Breaking the Monolith road to containers.pdf
 

More from Amazon Web Services

Come costruire servizi di Forecasting sfruttando algoritmi di ML e deep learn...
Come costruire servizi di Forecasting sfruttando algoritmi di ML e deep learn...Come costruire servizi di Forecasting sfruttando algoritmi di ML e deep learn...
Come costruire servizi di Forecasting sfruttando algoritmi di ML e deep learn...Amazon Web Services
 
Big Data per le Startup: come creare applicazioni Big Data in modalità Server...
Big Data per le Startup: come creare applicazioni Big Data in modalità Server...Big Data per le Startup: come creare applicazioni Big Data in modalità Server...
Big Data per le Startup: come creare applicazioni Big Data in modalità Server...Amazon Web Services
 
Esegui pod serverless con Amazon EKS e AWS Fargate
Esegui pod serverless con Amazon EKS e AWS FargateEsegui pod serverless con Amazon EKS e AWS Fargate
Esegui pod serverless con Amazon EKS e AWS FargateAmazon Web Services
 
Costruire Applicazioni Moderne con AWS
Costruire Applicazioni Moderne con AWSCostruire Applicazioni Moderne con AWS
Costruire Applicazioni Moderne con AWSAmazon Web Services
 
Come spendere fino al 90% in meno con i container e le istanze spot
Come spendere fino al 90% in meno con i container e le istanze spot Come spendere fino al 90% in meno con i container e le istanze spot
Come spendere fino al 90% in meno con i container e le istanze spot Amazon Web Services
 
Rendi unica l’offerta della tua startup sul mercato con i servizi Machine Lea...
Rendi unica l’offerta della tua startup sul mercato con i servizi Machine Lea...Rendi unica l’offerta della tua startup sul mercato con i servizi Machine Lea...
Rendi unica l’offerta della tua startup sul mercato con i servizi Machine Lea...Amazon Web Services
 
OpsWorks Configuration Management: automatizza la gestione e i deployment del...
OpsWorks Configuration Management: automatizza la gestione e i deployment del...OpsWorks Configuration Management: automatizza la gestione e i deployment del...
OpsWorks Configuration Management: automatizza la gestione e i deployment del...Amazon Web Services
 
Microsoft Active Directory su AWS per supportare i tuoi Windows Workloads
Microsoft Active Directory su AWS per supportare i tuoi Windows WorkloadsMicrosoft Active Directory su AWS per supportare i tuoi Windows Workloads
Microsoft Active Directory su AWS per supportare i tuoi Windows WorkloadsAmazon Web Services
 
Database Oracle e VMware Cloud on AWS i miti da sfatare
Database Oracle e VMware Cloud on AWS i miti da sfatareDatabase Oracle e VMware Cloud on AWS i miti da sfatare
Database Oracle e VMware Cloud on AWS i miti da sfatareAmazon Web Services
 
Crea la tua prima serverless ledger-based app con QLDB e NodeJS
Crea la tua prima serverless ledger-based app con QLDB e NodeJSCrea la tua prima serverless ledger-based app con QLDB e NodeJS
Crea la tua prima serverless ledger-based app con QLDB e NodeJSAmazon Web Services
 
API moderne real-time per applicazioni mobili e web
API moderne real-time per applicazioni mobili e webAPI moderne real-time per applicazioni mobili e web
API moderne real-time per applicazioni mobili e webAmazon Web Services
 
Database Oracle e VMware Cloud™ on AWS: i miti da sfatare
Database Oracle e VMware Cloud™ on AWS: i miti da sfatareDatabase Oracle e VMware Cloud™ on AWS: i miti da sfatare
Database Oracle e VMware Cloud™ on AWS: i miti da sfatareAmazon Web Services
 
Tools for building your MVP on AWS
Tools for building your MVP on AWSTools for building your MVP on AWS
Tools for building your MVP on AWSAmazon Web Services
 
How to Build a Winning Pitch Deck
How to Build a Winning Pitch DeckHow to Build a Winning Pitch Deck
How to Build a Winning Pitch DeckAmazon Web Services
 
Building a web application without servers
Building a web application without serversBuilding a web application without servers
Building a web application without serversAmazon Web Services
 
AWS_HK_StartupDay_Building Interactive websites while automating for efficien...
AWS_HK_StartupDay_Building Interactive websites while automating for efficien...AWS_HK_StartupDay_Building Interactive websites while automating for efficien...
AWS_HK_StartupDay_Building Interactive websites while automating for efficien...Amazon Web Services
 
Introduzione a Amazon Elastic Container Service
Introduzione a Amazon Elastic Container ServiceIntroduzione a Amazon Elastic Container Service
Introduzione a Amazon Elastic Container ServiceAmazon Web Services
 

More from Amazon Web Services (20)

Come costruire servizi di Forecasting sfruttando algoritmi di ML e deep learn...
Come costruire servizi di Forecasting sfruttando algoritmi di ML e deep learn...Come costruire servizi di Forecasting sfruttando algoritmi di ML e deep learn...
Come costruire servizi di Forecasting sfruttando algoritmi di ML e deep learn...
 
Big Data per le Startup: come creare applicazioni Big Data in modalità Server...
Big Data per le Startup: come creare applicazioni Big Data in modalità Server...Big Data per le Startup: come creare applicazioni Big Data in modalità Server...
Big Data per le Startup: come creare applicazioni Big Data in modalità Server...
 
Esegui pod serverless con Amazon EKS e AWS Fargate
Esegui pod serverless con Amazon EKS e AWS FargateEsegui pod serverless con Amazon EKS e AWS Fargate
Esegui pod serverless con Amazon EKS e AWS Fargate
 
Costruire Applicazioni Moderne con AWS
Costruire Applicazioni Moderne con AWSCostruire Applicazioni Moderne con AWS
Costruire Applicazioni Moderne con AWS
 
Come spendere fino al 90% in meno con i container e le istanze spot
Come spendere fino al 90% in meno con i container e le istanze spot Come spendere fino al 90% in meno con i container e le istanze spot
Come spendere fino al 90% in meno con i container e le istanze spot
 
Open banking as a service
Open banking as a serviceOpen banking as a service
Open banking as a service
 
Rendi unica l’offerta della tua startup sul mercato con i servizi Machine Lea...
Rendi unica l’offerta della tua startup sul mercato con i servizi Machine Lea...Rendi unica l’offerta della tua startup sul mercato con i servizi Machine Lea...
Rendi unica l’offerta della tua startup sul mercato con i servizi Machine Lea...
 
OpsWorks Configuration Management: automatizza la gestione e i deployment del...
OpsWorks Configuration Management: automatizza la gestione e i deployment del...OpsWorks Configuration Management: automatizza la gestione e i deployment del...
OpsWorks Configuration Management: automatizza la gestione e i deployment del...
 
Microsoft Active Directory su AWS per supportare i tuoi Windows Workloads
Microsoft Active Directory su AWS per supportare i tuoi Windows WorkloadsMicrosoft Active Directory su AWS per supportare i tuoi Windows Workloads
Microsoft Active Directory su AWS per supportare i tuoi Windows Workloads
 
Computer Vision con AWS
Computer Vision con AWSComputer Vision con AWS
Computer Vision con AWS
 
Database Oracle e VMware Cloud on AWS i miti da sfatare
Database Oracle e VMware Cloud on AWS i miti da sfatareDatabase Oracle e VMware Cloud on AWS i miti da sfatare
Database Oracle e VMware Cloud on AWS i miti da sfatare
 
Crea la tua prima serverless ledger-based app con QLDB e NodeJS
Crea la tua prima serverless ledger-based app con QLDB e NodeJSCrea la tua prima serverless ledger-based app con QLDB e NodeJS
Crea la tua prima serverless ledger-based app con QLDB e NodeJS
 
API moderne real-time per applicazioni mobili e web
API moderne real-time per applicazioni mobili e webAPI moderne real-time per applicazioni mobili e web
API moderne real-time per applicazioni mobili e web
 
Database Oracle e VMware Cloud™ on AWS: i miti da sfatare
Database Oracle e VMware Cloud™ on AWS: i miti da sfatareDatabase Oracle e VMware Cloud™ on AWS: i miti da sfatare
Database Oracle e VMware Cloud™ on AWS: i miti da sfatare
 
Tools for building your MVP on AWS
Tools for building your MVP on AWSTools for building your MVP on AWS
Tools for building your MVP on AWS
 
How to Build a Winning Pitch Deck
How to Build a Winning Pitch DeckHow to Build a Winning Pitch Deck
How to Build a Winning Pitch Deck
 
Building a web application without servers
Building a web application without serversBuilding a web application without servers
Building a web application without servers
 
Fundraising Essentials
Fundraising EssentialsFundraising Essentials
Fundraising Essentials
 
AWS_HK_StartupDay_Building Interactive websites while automating for efficien...
AWS_HK_StartupDay_Building Interactive websites while automating for efficien...AWS_HK_StartupDay_Building Interactive websites while automating for efficien...
AWS_HK_StartupDay_Building Interactive websites while automating for efficien...
 
Introduzione a Amazon Elastic Container Service
Introduzione a Amazon Elastic Container ServiceIntroduzione a Amazon Elastic Container Service
Introduzione a Amazon Elastic Container Service
 

Amazon SageMaker Deep Dive - Meetup AWS Toulouse at D2SI

  • 1. © 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Amazon SageMaker Develop, Train, Tune and Deploy Machine Learning Models 2019-04-24, D2SI Toulouse Olivier Cruchant, ML Specialist SA, AWS
  • 2. © 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Put machine learning in the hands of every developer and data scientist ML @ AWS: Our mission
  • 3. © 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Customer Running ML on AWS Today
  • 4. © 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved. AWS ML Stack P L A T F O R M S A P P L I C A T I O N S E R V I C E S F R A M E W O R K S E T I N F R A S T R U C T U R E Amazon SageMaker Vision Rekognition Image Rekognition Video Amazon Textract Speech Amazon Polly Amazon Transcribe Language Amazon Lex Amazon Translate Amazon Comprehend Forecasting Amazon Forecast Personalization Amazon Personalize DL AMI EC2 P3, C5, F1 GreengrassElastic Inference
  • 5. © 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Build, Train and Deploy your Machine Learning Models Amazon SageMaker Models are abstracted into docker containers Fast & accurate data labeling Built-in, high performance algorithms & notebooks BUILD 1 One-click training and tuning TRAI N Model tuning and optimization 2 Fully managed hosting with auto-scaling and elastic inference One-click Deployment of model or Inference Pipeline DEPLOY 3
  • 6. © 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Amazon SageMaker: TensorFlow workflow example Develop on Notebook instance with TensorFlow Jupyter kernel BUILD 1 Train custom code in TensorFlow container TRAI N Model tuning and optimization 2 Deploy to auto-scaled tensorflow-serving container DEPLOY 3 Stream training data with PipeModeDataset Distributed training with Horovod Monitor with TensorBoard
  • 7. © 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Amazon SageMaker Deployment & Inference A. Deploy model to highly-available Amazon SageMaker HTTPS endpoint B. Run Amazon SageMaker batch-transform job C. On-device deployment (AWS Greengrass)
  • 8. © 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Amazon SageMaker: algorithms A. Train and deploy a built-in algorithm B. Train and deploy your own code in a framework container C. Train and deploy your own container D. Bring a pre-trained model E. Train and Deploy a Marketplace Algorithm F. Deploy a Marketplace Model Package
  • 9. © 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Amazon SageMaker: interface A. Console B. AWS Command Line Interface (CLI): aws sagemaker create-endpoint --endpoint-name <value> --endpoint-config-name <value> C. SageMaker Python SDK model.deploy(initial_instance_count=1, instance_type='ml.m4.xlarge') D. Boto3 (python) SageMaker client client.create_endpoint(EndpointName='string', EndpointConfigName='string') E. Other SDKs
  • 10. © 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved. In summary:
  • 11. Python SDK: Amazon Algorithm
  • 12. Python SDK: Deep Learning Estimator
  • 13. © 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Amazon SageMaker Built-in algorithms Supervised Linear Learner Factorization Machines XGBoost Image Classification Object Detection Sequence to Sequence DeepAR Forecasting K-Nearest Neighbors BlazingText (classifier) Object2Vec Semantic Segmentation Unsupervised K-Means Principal Component Analysis Latent Dirichlet Allocation Neural Topic Model Random Cut Forest BlazingText (word2vec) IP Insights
  • 14. © 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Object Detection A. Single-Shot multi-box Detector (SSD): learns to locate and classify at the same time B. Extract features with a convolutional base (VGG-16 and ResNet-50 available) C. Transfer learning from ImageNet available D. Example: aerial imagery analysis
  • 15. © 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved. DeepAR A. Forecasts scalar time series with RNN B. Adapted to forecast families of similar time series (product sales, energy by household) C. Learning set can include arbitrary categorical and time series features D. Automatically derives features based on target frequency (eg day of month, day of year) E. Automatically feeds lagged values from target time series https://docs.aws.amazon.com/sagemaker/latest/dg/deepar_how-it-works.html
  • 16. © 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved. BlazingText Word2Vec A. Goal: produce low-dimension, dense semantic word embeddings B. CUDA implementation of word2vec that can be distributed over multiple GPUs C. Byproduct of a text classification problem (CBOW or Skip-gram) 1 0 … … … … … 0 n1 n300 0 0 … … … 1 0 0 … Apple Juice wn1, 1 Voc size: 40k wn1, 2 wn1, 40k wn300, 1 wn300, 2 wn300, 40k Word e1 … e300 Apple wn1, 1 … wn300, 1 … … … … Zebra … … … ⟹ Simplified architecture. See original papers for more details https://papers.nips.cc/paper/5021-distributed-representations-of-words-and-phrases- and-their-compositionality.pdf https://arxiv.org/pdf/1301.3781.pdf https://arxiv.org/pdf/1607.04606.pdf
  • 17. © 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved. BlazingText Classifier A. Extends fastText to work on GPU B. Learns word embeddings that are averaged then linearly fit C. Features include n-grams D. Features are hashed to limit dimensionality https://aws.amazon.com/fr/blogs/machine-learning/enhanced-text-classification-and-word- vectors-using-amazon-sagemaker-blazingtext/
  • 18. © 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Object2Vec A. Generate predictions or embeddings B. Use-cases: recommendation, similarity, multi-label classification C. Supports 2 types of inputs: • Discrete tokens • Sequences of discrete tokens
  • 19. © 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Random Cut Forest Algo: 1. Partition input space into random sample 2. For each sample: i. Draw bounding box ii. Random cut on a random feature (sampling proportional to variance) iii. Points alone in their bounding boxes become tree leaves 3. Outlier score inversely proportional to leaf depth … …
  • 20. © 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved. SageMaker Ground Truth
  • 21. © 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved. SageMaker Neo Compiler Runtime
  • 22. © 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved. https://aws.amazon.com/sagemaker http://d2l.ai/ Getting Started