Скачать или смотреть Continuous Delivery and Automation Pipelines in Machine Learning // Part 1 // MLOps Coffee Sessions

In this video, we talk about Continuous Integration, Continuous Delivery, and Continuous Training in your Machine Learning Development lifecycle. Some may have seen this called CI/CD/CT

In this last episode, we covered how Google is thinking about MLOps and how automation plays a key part in their view of MLOps. We started to talk about CI, CD, and the role they play in a pipeline setup for Continuous training in machine learning. In the next episode, we'll pick up where we left off, starting our discussion of CT and some of the reasons you’d want to set up a pipeline with continuous training in the first place.

-------------- ✌️Connect With Us ✌️ ------------
Join our slack community: https://go.mlops.community/slack
Follow us on Twitter: @mlopscommunity
Sign up for the next meetup: https://go.mlops.community/register
Catch all episodes, Feature Store, Machine Learning Monitoring and Blogs: https://mlops.community/

Connect with Demetrios on LinkedIn:   / dpbrinkm  
Connect with David on LinkedIn:   / aponteanalytics  

Timestamps:
[00:00] Introduction to the discussion
[04:10] Why is there a need to automate
[07:22] Google's culture
[10:34] Definition of MLOps
[12:33] Google's idea of pipelines
[16:57] Reliable and repeatable training
[18:36] Parameters for models
[19:32] Overview of its architecture
[23:11] Pipelining with Kubeflow and others
[26:00] Not much difference to traditional software
[30:12] Lineage tracking
[31:37] Operations side
[33:25] How is it different with CML
[38:12] Over-automation or over-engineering
[43:36] Ownership of responsibilities in pipelines
[45:34] Continuous integration
[48:52] Continuous delivery
[49:52] Continuous training
[52:16] Automation monitoring
[57:16] Wrap up