The Containerization of Machine Learning: TensorFlow: Kubernetes, and Kubeflow
summary
Explains why Kubernetes fits ML workloads and shows a working Kubeflow stack deployed to GKE with TFJob CRD for distributed TensorFlow, Katib hyperparameter tuning, JupyterHub notebooks, and TFX components (Data Validation, Transform, Model Analysis) feeding TensorFlow Serving or NVIDIA TensorRT. Demo trains a taxi fare model on Chicago data with about 1,000 training and 5,000 evaluation examples, composes the pipeline into a container image and runs end-to-end successfully.
AI-written summary · about
chapters (8)
- 0:00Introduction to containerization
- 1:27Machine learning everywhere
- 4:18Why cloud-native ML
- 6:31Kubeflow overview
- 11:01Demo setup and data
- 13:10Deploying Kubeflow
- 15:28Building ML pipelines
- 21:16Summary and Q&A
related talks
31mKubernetes for Machine Learning: Managing and Scaling AI Workloads
46mKubeflow, TensorFlow on Kubernetes
42mKubernetes for Machine Learning: Managing and Scaling AI Workloads
46mDeploying scalable Machine Learning models with Kubernetes
29mKubeflow: TensorFlow on Kubernetes
36mKubeflow: Tensorflow on Kubernetes
39mDeploy your ML Workloads on Kubernetes with Kubeflow
46mAccelerating Machine Learning DevOps with Kubeflow