Keynote: Containerizing Deep Learning Using Kubeflow on AWS

on YouTube ↗18m

" summary

Describes using Kubeflow on Amazon EKS to containerize and productionize deep learning workflows, addressing problems like framework heterogeneity, composability, portability, performance optimization, collaborative development, scalability, and infrastructure management. Includes a demo of single-node and distributed deep learning training on AWS EKS using Kubeflow.

AI-written summary · about

" chapters (14)

1. 0:00Intro
2. 0:18Why Machine Learning
3. 2:34The Machine Learning Workflow - Steps
4. 2:48The Machine Learning Workflow - Code
5. 3:13The Machine Learning SDLC
6. 3:54ML Ops
7. 5:58Challenges while doing MLOps
8. 7:59Containers to rescue
9. 8:56Machine Learning on Kubernetes - without Kubeflow
10. 9:50Kubeflow ecosystem for Machine Learning
11. 11:33How does it fit into the ML workflow?
12. 13:50Other options
13. 15:52AWS EKS - how does it work
14. 16:11A Data Scientist's workflow using Kubeflow on AWS

" related talks

• 51mMachine Learning using Kubeflow and KubernetesArun Gupta · Devoxx Belgium 2019
• 10mKeynote: Machine Learning using Kubeflow and KubernetesArun Gupta · Kubernetes Forum Delhi 2020
• 54mMachine learning at scale using Kubeflow with Amazon EKS & Amazon EFSSuman Debnath, Bimal Gajjar · AWS Summit Atlanta 2022
• 39mOperating Deep Learning Pipelines Anywhere Using KubeflowJörg Schad, Gilbert Song · KubeCon + CloudNativeCon China 2018
• 46mKubeflow, TensorFlow on KubernetesIsrael Olalla · Commit Conf 2018
• 15mKeynote: Machine Learning on Kubernetes Made Easy With KubeflowMasoud Mirmomeni, Jimmy Guerrero · KubeCon + CloudNativeCon North America 2021
• 30mKubeflow, Your ML Tools on KubernetesLaura Morillo, Iván Cañaveral · Commit Conf 2018
• 54mEconomics and Best Practices of Running AI/ML Workloads on KubernetesMaulin Patel · KubeCon + CloudNativeCon Europe 2019