3. Learning system requires HPC architecture.
Inference system requires Web architecture.
My messages (Deep learning web services)
NVIDIA’s deep learning strategy intends for large-scale data center.
If you want to build the services on a cloud such as AWS, you can have other way.
長尾の個人的見解ですが・・
6. Learning
Neural Network
Structure
Inference
Learning: HPC Architecture Inference: Web Architecture
BIGDATA Labels
Deploy a trained
Network
apple Orange strawberryBanana
Deep Learning web services require two system.
Require real time processRequire Many Core, BIGDATA analysis
API Micro
Service
• マイクロサービスを提供する
• ~250μs response time
• リクエストを常時待ちうけ。
• 数日間、CPU and/or GPUの負荷がほぼ100%で動作する。マル
チノード、マルチGPUに対応
• ジョブを走らせているときのみ計算ノードを利用
7. Learning
Neural Network
Structure
Inference
Learning: HPC Architecture Inference: Web Architecture
BIGDATA Labels
Deploy a trained
Network
apple Orange strawberryBanana
Deep Learning web services require two system.
Require real time processRequire Many Core, BIGDATA analysis
API Micro
Service
Learningに適したGPU 推論に適したGPU
8. Learning
Neural Network
Structure
Inference
Learning: HPC Architecture Inference: Web Architecture
BIGDATA Labels
Deploy a trained
Network
apple Orange strawberryBanana
Require real time processRequire Many Core, BIGDATA analysis
API Micro
Service
nvidia-docker
• There are many Deep Learning framework and version.
• Need to deploy the trained network to Inference server.
9. Neural Network
Structure
Learning: HPC Architecture Inference: Web Architecture
BIGDATA Labels
Deploy a trained
Network
apple Orange strawberryBanana
Require real time processRequire Many Core, BIGDATA analysis
API
Learning Inference
nvidia-docker
• There are many Deep Learning framework and version.
• Need to deploy the trained network to Inference server.
Docker
Registry
PUSH PULL
Micro
Service
10.
11. Learning
Neural Network
Structure
Inference
Learning: HPC Architecture Inference: Web Architecture
BIGDATA Labels
Deploy a trained
Network
apple Orange strawberryBanana
Require real time processRequire Many Core, BIGDATA analysis
API
GPU Rest Engine is template written by go-lang to launch the micro-service.
This template launch webserver with the port number that is set by admin.
GPU Rest EngineDocker
Registry
https://github.com/NVIDIA/gpu-rest-engine
14. Neural Network
Structure
Learning: HPC Architecture Inference: Web Architecture
BIGDATA Labels
Deploy a trained
Network
apple Orange strawberryBanana
Require real time processRequire Many Core, BIGDATA analysis
API
ついに HPC でも Docker が! How deploy your apps?Mesos can give an abstract of datacenter with both HPC and Webserver
Submit Job Daemon
Learning Inference
GPU Rest EngineDocker
Registry
16. Neural Network
Structure
Learning: HPC Architecture Inference: Web Architecture
BIGDATA
Labels
Deploy a
trained
Network
apple Orange strawberryBanana
Require real time processRequire Many Core, BIGDATA analysis
API
ついに HPC でも Docker が! How deploy your apps?
Learning Inference
AWS Elastic
Beanstalk
bucket
cfncluster
Amazon
DynamoDB
Amazon
DynamoDB
bucket
Meta Data
Amazon API
Gateway
Submit Job Daemon
HPC on AWS 用ミドルウェア
• HPC Clusterの動的作成・削除・管理
• スケジューリング機能
17. Neural Network
Structure
Learning: HPC Architecture Inference: Web Architecture
BIGDATA
Labels
Deploy a
trained
Network
apple Orange strawberryBanana
Require real time processRequire Many Core, BIGDATA analysis
API
ついに HPC でも Docker が! How deploy your apps?
Learning Inference
AWS Elastic
Beanstalk
bucket
cfncluster
Amazon
DynamoDB
Amazon
DynamoDB
bucket
Meta Data
Amazon API
Gateway
学習もAPI起動するなら、
ジョブ管理ソフトをキッ
クするよう構成
Submit Job Daemon