Flyte School: A Practical Introduction to Machine Learning Orchestration скачать в хорошем качестве

Flyte School: A Practical Introduction to Machine Learning Orchestration

A practical, in-depth overview of Machine Learning orchestration using the open-source platform Flyte, one of the leading orchestration tools. This session will enhance your understanding of an orchestrator's role in building Machine Learning pipelines. We will discuss the key features of a robust orchestrator, illustrated through practical examples from Flyte. You'll gain insights into crucial concepts such as scheduling, versioning, parameterization, and data lineage, and understand how Directed Acyclic Graphs (DAGs) can provide clear visualizations of data dependencies Learning Objectives: Understand how tasks, workflows, and launchplans can be used to manage the flow of data and models through a sequence of computations. Learn how to spin up a local Flyte cluster on their own workstation and run workflows on it for local debugging. Programmatically run tasks and workflows through the CLI or in a Python runtime to quickly iterate on your code. Feel free to reach out to Samhita on Twitter:   / samhita_alla   Join the Flyte Slack community: https://flyte-org.slack.com/ Checkout Flyte, the open-source orchestrator that facilitates building production-grade data and ML pipelines: https://github.com/flyteorg/flyte