Kubeflow: Data Scientist ready?

on YouTube ↗35m

" summary

Kubeflow can be installed quickly using a provided script that deploys a working cluster in minutes or manually with a more error‑prone process. It runs notebooks via JupyterHub where users spawn Jupyter images with CPU/GPU limits, trains with TFJob YAMLs that require persistentVolumeClaims to export models, and serves models by generating TFServing plus an HTTP proxy via ks/ksonnet—though many serving parameters are undocumented, versions often break, and ad hoc TensorFlow patches and large Docker images are frequently needed.

AI-written summary · about

" chapters (13)

1. 0:00Kubeflow for data scientists
2. 3:15Kubernetes & Kustomize
3. 5:25Kubeflow tools
4. 8:53Installing Kubeflow
5. 11:00JupyterHub exploration
6. 13:34TensorFlow training
7. 17:26Persistent Volumes
8. 20:33Distributed TensorFlow
9. 24:25TF Serving deployment
10. 25:50Ksonnet automation
11. 30:23Kubeflow docs & roadmap
12. 32:08Other Kubeflow tools
13. 34:36Conclusion

" related talks

• 29mKubeflow: TensorFlow on KubernetesLaurent Grangeau · DevFest Nantes 2018
• 31mKubeflow - Data Science on SteroidsSascha Grunert · ContainerDay 2019
• 21mIntroducing Kubeflow (w. Special Guests Tensorflow and Apache Spark)Holden Karau, Trevor Grant · Open Source Summit + Embedded Linux Conference North America 2019
• 36mKubeflow: Tensorflow on KubernetesPar Laurent Grangeau, Par Sylvain Lequeux · DataXDay 2019
• 36mIntroducing Kubeflow (w. Special Guests Tensorflow and Apache Spark)Trevor Grant · KubeCon + CloudNativeCon + Open Source Summit China 2019
• 46mKubeflow, TensorFlow on KubernetesIsrael Olalla · Commit Conf 2018
• 14mKubeflow: Where ML Meets KubernetesSuresh Peiris · KCD Sri Lanka 2023
• 36mIntroducing Kubeflow (w. Special Guests Tensorflow and Apache Spark)Trevor Grant · KubeCon + CloudNativeCon China 2019