Machine learning pipeline Kubeflow, let's start
summary
Demonstrates building runnable Kubeflow Pipelines on Kubernetes using Kubeflow Pipelines and Argo Workflows, with artifacts stored in Minio/S3. Shows two concrete containerized pipelines: a build pipeline and a conditional ML training pipeline that compares classifier precision to a 90% threshold to choose between two models. Includes a one-command local setup (Jupyter + Kubeflow Pipelines) that launches pods and GPU/CPU containers so the demos execute end-to-end.
AI-written summary · about
chapters (11)
- 0:00Kubeflow ML pipeline
- 3:06ML growth demand
- 7:26Kubeflow components
- 8:49Pipeline key features
- 11:43Standard demo
- 13:50Conditional demo
- 17:11Defining pipeline
- 21:26Isolation & reproducibility
- 23:13GPU scalability
- 26:05Quick test
- 29:46Q&A
related talks
54mScaling Machine Learning Pipelines in Cloud
51mMachine Learning using Kubeflow and Kubernetes
29mML Engineering Kubeflow
05mLightning Talk: Build a Reproducible ML Workflow with Kubeflow Pipelines
05mCI/CD Pipelines and Machine Learning
35mOrganizing data engineering around Kubeflow
39mOperating Deep Learning Pipelines Anywhere Using Kubeflow
31mKubeflow - Data Science on Steroids