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
Kubeflowではじめる ML on Kubernetes #2 Kubeflow Not...
Search
Sponsored
·
Ship Features Fearlessly
Turn features on and off without deploys. Used by thousands of Ruby developers.
→
Keita Watanabe
October 06, 2022
Programming
870
0
Share
Embed
Copy iframe code
Copy JS code
Copy link
Start on current slide
Kubeflowではじめる ML on Kubernetes #2 Kubeflow Notebooks
https://k8sjp.connpass.com/event/259350/
発表資料
Keita Watanabe
October 06, 2022
More Decks by Keita Watanabe
See All by Keita Watanabe
[re:Invent 2025] Distributed inference on AWS: Deep dive into inference optimizations (AIM353)
keitaw
1
180
Scalable Infrastructure for Large-Scale AI Training with AWS Sagemaker Hyperpod @ Singapore AI Hour
keitaw
0
130
[AWS Summit Japan 2025] Optimizing Foundation Model Development with Amazon SageMaker HyperPod: Insights from Training the Amazon Nova Model
keitaw
0
69
Building foundation models on AWS
keitaw
1
590
[re:Invent2024 Chalktalk] Cost-effectively deploy PyTorch LLMs on AWS Inferentia using Amazon EKS
keitaw
0
150
AWS Summit New York 2024: CMP 301 Demystifying the ML software stack on Amazon EC2 accelerated instances
keitaw
0
520
re:Invent 2023 CMP319 Deploy LLMs with AWS Inferentia & Ray to optimize performance and cost
keitaw
0
75
re:Invent 2023: CMP332 De-mystifying ML software stack on Amazon EC2 accelerated instances
keitaw
1
260
AWS における LLM・GenAI 大規模学習への取り組み / Large scale GenAI・LLM training on AWS
keitaw
1
610
Other Decks in Programming
See All in Programming
不変条件と整合性境界—ビジネスが決める設計判断と実現パターン / Invariants and Consistency Boundaries
nrslib
14
5.8k
スマートグラスで並列バイブコーディング
hyshu
0
260
脅威をエンジニアリングの糧にして――現場編 / Turning Threats into Engineering Fuel — Field Edition
nrslib
0
300
Semantic Version 単位で戦略を柔軟に変えて、パッケージアップデートを自動化する
daitasu
1
300
代数的データ型って何が嬉しいの? #frontend_phpcon_do
kajitack
8
3.8k
エンジニア向け会社紹介/Findy Company Profile
findyinc
6
350k
「なぜそう決めたのか」を残し続ける仕組み ― Notion AI カスタムエージェント × Slack連携による設計判断の自動記録 - NIKKEI Tech Talk #47
niftycorp
PRO
0
230
LaravelLive Japan の裏方のすべて — 第188回 PHP勉強会@東京 (2026-06-24)
suguruooki
2
110
LLMによるContent Moderationの本番運用の裏側と品質担保への挑戦
suikabar
3
740
依存関係から依存物へ―Dependencyという言葉の歴史をひも解く
j_lee
0
130
気づいたらRubyで100作品 ー クリエイティブコーディングが生活の一部になるまで / 100 Ruby Sketches Later: How Creative Coding Became Part of My Life
chobishiba
3
610
Spring Security 実践 ─ GraphQL APIで実務に役立つ 認証・認可 を学ぶ
wagyu
0
260
Featured
See All Featured
Odyssey Design
rkendrick25
PRO
2
700
Automating Front-end Workflow
addyosmani
1370
210k
End of SEO as We Know It (SMX Advanced Version)
ipullrank
3
4.2k
Beyond borders and beyond the search box: How to win the global "messy middle" with AI-driven SEO
davidcarrasco
3
170
The #1 spot is gone: here's how to win anyway
tamaranovitovic
2
1.1k
AI Search: Implications for SEO and How to Move Forward - #ShenzhenSEOConference
aleyda
1
1.3k
Money Talks: Using Revenue to Get Sh*t Done
nikkihalliwell
0
260
Writing Fast Ruby
sferik
630
63k
The untapped power of vector embeddings
frankvandijk
2
1.8k
How To Speak Unicorn (iThemes Webinar)
marktimemedia
1
490
The Anti-SEO Checklist Checklist. Pubcon Cyber Week
ryanjones
0
170
B2B Lead Gen: Tactics, Traps & Triumph
marketingsoph
0
160
Transcript
Keita Watanabe Kubernetes Meetup Tokyo # 53 LT KubeflowͰ͡ΊΔ ML
on Kubernetes #2 Kube fl ow Notebooks
Keita Watanabe Machine Learning Solutions Architect ▶ AWS JapanͰSelf-managedͳMachine Learning
ͷҊ݅Λ୲͍ͯ͠ΔSolutions Architect ▶ લ৬ͰDatascientist/ML Researcherͱͯ͠ ECαΠτ্ͷը૾ݕࡧػೳͷ։ൃʹैࣄ ▶ Twitter: keitaw09 ▶ Linkedin: keitawatanabe αϯσΟΤΰͰग़ձͬͨτϦ झຯ
Kubeflow overview Central Dashboard Data Ingestion Data Analysis Data Transform
Data Validation Training AutoML Model eval. and valid. Serving Logging KServe Seldon Kube fl ow Pipelines Kube fl ow Notebooks Katib Training Operators MDLCͷ֤εςοϓ ରԠ͢ΔTool લճͷ༰: Kube fl owͷηοτΞοϓ Slide: https://speakerdeck.com/keitaw/kube fl owdehazimeruml-on-kubernetes-number-1-kube fl owfalsegai-yao-tosetutoatupu
ຊͷ༰: Kubeflow Notebooks
Kubeflow Notebooksͱ • Webϕʔεͷ։ൃڥΛKubernetes্Ͱఏڙ͢Δπʔϧ • Կ͕خ͍͔͠ • ཧऀɿඞཁͳύοέʔδ͕Πϯετʔϧ͞Εͨ։ൃऀ༻ͷΠϝʔδΛ खܰʹఏڙͰ͖Δ +
RBACʹΑΔnotebookͷΞΫηε੍ޚ • ϢʔβʔɿϩʔΧϧͷܭࢉࢿݯʹཔΔ͜ͱͳ͘ɺ Ϋϥελ্ͰNotebookڥΛར༻Ͱ͖Δ JupyterLab, RStudio, Visual Studio Code (code-server)Λ ωΠςΟϒͰαϙʔτ͍ͯ͠Δ https://www.kube fl ow.org/docs/components/notebooks/overview/
σϑΥϧτͰαϙʔτ͞Ε͍ͯΔΠϝʔδ https://github.com/kube fl ow/kube fl ow/tree/master/components/example-notebook-servers#custom-images • Base: ϕʔεΠϝʔδʢUbuntuʣ •
PyTorch/TensorFlow Full: PyTorch/TensorFlow + Pandas ScipyɺXGBoostͷΑ͘ΘΕΔ ϥΠϒϥϦ • Jupyter imageʹGPU൛͋Δ • Custom image: BaseΛجʹ࡞ ʢ࡞ํ๏ԼهͷURLΛࢀরʣ https://www.kube fl ow.org/docs/components/notebooks/container-images/
Notebookͷ৽ن࡞
Notebookͷ৽ن࡞
Notebookͷ৽ن࡞ ❶ ❷ ❸ ❹
Kubeflow NotebooksͰJupyterLab σʔλͷղੳϞσϧͷ։ൃɺ͞ΒʹKube fl ow PipelinesʢৄࡉࠓޙͷLTͰʣ Λ༻͍ͨύΠϓϥΠϯͷߏஙͳͲ͕ߦ͑Δ
Kubeflow NotebooksͰVSCode
࣍ճ… Kubeflow Pipelines
ࢀߟࢿྉ • https://www.kube fl ow.org/docs/components/notebooks/overview/ • https://github.com/kube fl ow/kube fl
ow/tree/master/components/example- notebook-servers#custom-images • https://github.com/kube fl ow/pipelines/blob/sdk/release-1.8/samples/core/ lightweight_component/lightweight_component.ipynb • https://github.com/aws-samples/aws-do-kube fl ow