© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Hagay Lupesko, Amazon AI
AWS ML Day in San Francisco, CA
June 19 2018
Build, train and deploy ML
models using Amazon
SageMaker
© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
The Amazon Machine Learning Stack
PLATFORM SERVICES
APPLICATION SERVICES
FRAMEWORKS & INTERFACES
Caffe2 CNTK
Apache
MXNet
PyTorch TensorFlow Chainer Keras Gluon
AWS Deep Learning AMIs
Amazon SageMaker
Rekognition Transcribe Translate Polly Comprehend Lex
AWS
DeepLens
EDUCATION
Amazon EMRAmazon MTurk
© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Amazon SageMaker
A fully-managed platform
that provides a quick and easy way to
get models from idea to production.
https://aws.amazon.com/sagemaker/
Amazon SageMaker Workflow
Amazon’s fast, scalable algorithms
Distributed TensorFlow, Apache MXNet, Chainer, PyTorch
Bring your own algorithm
Hyperparameter Tuning
Building HostingTraining
Amazon SageMaker Workflow
Amazon’s fast, scalable algorithms
Distributed TensorFlow, Apache MXNet, Chainer, PyTorch
Bring your own algorithm
Hyperparameter Tuning
Building HostingTraining
Building Machine Learning Models
Create and
manage cloud-
based notebooks
Use SageMaker‘s
web interface to
get started
Or build models
locally, then upload to
SageMaker
Amazon SageMaker Workflow
Amazon’s fast, scalable algorithms
Distributed TensorFlow, Apache MXNet, Chainer, PyTorch
Bring your own algorithm
Hyperparameter Tuning
Building HostingTraining
Training Machine Learning Models
Zero setup Streaming datasets
+ distributed
compute
Deploy trained models
to Amazon
SageMaker, AWS
Greengrass or AWS
DeepLens
Integrate with your
Spark-based data
pipeline
Amazon SageMaker Workflow
Amazon’s fast, scalable algorithms
Distributed TensorFlow, Apache MXNet, Chainer, PyTorch
Bring your own algorithm
Hyperparameter Tuning
Building HostingTraining
Hosting
One-click
deployment
Low latency, high
throughput, and
high reliability
A/B testing Bring your own
model
Amazon SageMaker Workflow
Amazon’s fast, scalable algorithms
Distributed TensorFlow, Apache MXNet, Chainer, PyTorch
Bring your own algorithm
Hyperparameter Tuning
Building HostingTraining
SageMaker built-in algorithms
XGBoost, FM,
Linear, and
Forecasting
for supervised
learning
Kmeans, PCA,
and Word2Vec
for clustering
and pre-
processing
Image
classification
with
convolutional
neural networks
LDA and NTM
for topic
modeling,
seq2seq for
translation
Amazon SageMaker Workflow
Amazon’s fast, scalable algorithms
Distributed TensorFlow, Apache MXNet, Chainer, PyTorch
Bring your own algorithm
Hyperparameter Tuning
Building HostingTraining
TensorFlow and Apache MXNet Docker
Containers
… explore and
refine models in a
single Notebook
instance…
… deploy to
production
Sample your
data…
… use the same code
to train on the full
dataset in a cluster of
instances…
Amazon SageMaker components
Amazon’s fast, scalable algorithms
Distributed TensorFlow, Apache MXNet, Chainer, PyTorch
Bring your own algorithm
Hyperparameter Tuning
Building HostingTraining
Bring your own algorithm
... add algorithm
code to a Docker
container...
Pick your preferred
framework...
... publish your
Docker image to ECR
Hyperparameter Tuning
(Automated Model Tuning)
Run a large set of training
jobs with varying
hyperparameters...
... and search the
hyperparameter space for
improved accuracy.
Zero setup for data exploration
Resizable as you
need
Common tools
pre-installed
Easy access to
your data sources
No servers to
manage
Pay as you go
ML compute by the
second starting
at $0.0464/hr
ML storage by the
second
at $0.14
per GB-month
Data processed in
notebooks and hosting
at $0.016 per GB
Free trial to get started
quickly
© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Demo
Sentiment Analysis with
SageMaker and MXNet
https://bit.ly/2KEkBKt
(source: https://github.com/lupesko/sentiment-analysis-with-sagemaker-mxnet)
© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Thank You
To learn more, visit
https://aws.amazon.com/sagemaker/

Build, Train and Deploy ML Models using Amazon SageMaker

  • 1.
    © 2018, AmazonWeb Services, Inc. or its Affiliates. All rights reserved. Hagay Lupesko, Amazon AI AWS ML Day in San Francisco, CA June 19 2018 Build, train and deploy ML models using Amazon SageMaker
  • 2.
    © 2018, AmazonWeb Services, Inc. or its Affiliates. All rights reserved. The Amazon Machine Learning Stack PLATFORM SERVICES APPLICATION SERVICES FRAMEWORKS & INTERFACES Caffe2 CNTK Apache MXNet PyTorch TensorFlow Chainer Keras Gluon AWS Deep Learning AMIs Amazon SageMaker Rekognition Transcribe Translate Polly Comprehend Lex AWS DeepLens EDUCATION Amazon EMRAmazon MTurk
  • 3.
    © 2018, AmazonWeb Services, Inc. or its Affiliates. All rights reserved. Amazon SageMaker A fully-managed platform that provides a quick and easy way to get models from idea to production. https://aws.amazon.com/sagemaker/
  • 4.
    Amazon SageMaker Workflow Amazon’sfast, scalable algorithms Distributed TensorFlow, Apache MXNet, Chainer, PyTorch Bring your own algorithm Hyperparameter Tuning Building HostingTraining
  • 5.
    Amazon SageMaker Workflow Amazon’sfast, scalable algorithms Distributed TensorFlow, Apache MXNet, Chainer, PyTorch Bring your own algorithm Hyperparameter Tuning Building HostingTraining
  • 6.
    Building Machine LearningModels Create and manage cloud- based notebooks Use SageMaker‘s web interface to get started Or build models locally, then upload to SageMaker
  • 7.
    Amazon SageMaker Workflow Amazon’sfast, scalable algorithms Distributed TensorFlow, Apache MXNet, Chainer, PyTorch Bring your own algorithm Hyperparameter Tuning Building HostingTraining
  • 8.
    Training Machine LearningModels Zero setup Streaming datasets + distributed compute Deploy trained models to Amazon SageMaker, AWS Greengrass or AWS DeepLens Integrate with your Spark-based data pipeline
  • 9.
    Amazon SageMaker Workflow Amazon’sfast, scalable algorithms Distributed TensorFlow, Apache MXNet, Chainer, PyTorch Bring your own algorithm Hyperparameter Tuning Building HostingTraining
  • 10.
    Hosting One-click deployment Low latency, high throughput,and high reliability A/B testing Bring your own model
  • 11.
    Amazon SageMaker Workflow Amazon’sfast, scalable algorithms Distributed TensorFlow, Apache MXNet, Chainer, PyTorch Bring your own algorithm Hyperparameter Tuning Building HostingTraining
  • 12.
    SageMaker built-in algorithms XGBoost,FM, Linear, and Forecasting for supervised learning Kmeans, PCA, and Word2Vec for clustering and pre- processing Image classification with convolutional neural networks LDA and NTM for topic modeling, seq2seq for translation
  • 13.
    Amazon SageMaker Workflow Amazon’sfast, scalable algorithms Distributed TensorFlow, Apache MXNet, Chainer, PyTorch Bring your own algorithm Hyperparameter Tuning Building HostingTraining
  • 14.
    TensorFlow and ApacheMXNet Docker Containers … explore and refine models in a single Notebook instance… … deploy to production Sample your data… … use the same code to train on the full dataset in a cluster of instances…
  • 15.
    Amazon SageMaker components Amazon’sfast, scalable algorithms Distributed TensorFlow, Apache MXNet, Chainer, PyTorch Bring your own algorithm Hyperparameter Tuning Building HostingTraining
  • 16.
    Bring your ownalgorithm ... add algorithm code to a Docker container... Pick your preferred framework... ... publish your Docker image to ECR
  • 17.
    Hyperparameter Tuning (Automated ModelTuning) Run a large set of training jobs with varying hyperparameters... ... and search the hyperparameter space for improved accuracy.
  • 18.
    Zero setup fordata exploration Resizable as you need Common tools pre-installed Easy access to your data sources No servers to manage
  • 19.
    Pay as yougo ML compute by the second starting at $0.0464/hr ML storage by the second at $0.14 per GB-month Data processed in notebooks and hosting at $0.016 per GB Free trial to get started quickly
  • 20.
    © 2018, AmazonWeb Services, Inc. or its Affiliates. All rights reserved.© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Demo Sentiment Analysis with SageMaker and MXNet https://bit.ly/2KEkBKt (source: https://github.com/lupesko/sentiment-analysis-with-sagemaker-mxnet)
  • 21.
    © 2018, AmazonWeb Services, Inc. or its Affiliates. All rights reserved. Thank You To learn more, visit https://aws.amazon.com/sagemaker/

Editor's Notes

  • #6 Let’s start with Building
  • #22 So let's summarize what we went through: - We introduced SageMaker - We understood how you can use SM to build and traing your models - We learned how we can use SM to host your models - We went through a simple end to end example I will do another talk later today about model serving for deep learning for a more in-depth look into serving models. Hope you enjoyed!