From the course: Azure Spark Databricks Essential Training
Unlock the full course today
Join today to access over 22,600 courses taught by industry experts or purchase this course individually.
Use distributed ML training
From the course: Azure Spark Databricks Essential Training
Use distributed ML training
- [Instructor] We've been looking at relatively common use cases on ML pipelines for Spark with manual pipelines and Spark ML pipelines, and I think it's fun to look at what for most of us will be the future, and this is around distributed TensorFlow, which is a deep neural network, the most popular library, open sourced by Google, with a new library called Horovod and this is open sourced from Uber. So, to understand the class of problems that this is designed to solve you really want to put your future hat on. Uber is using this to process all of their massive amounts of data from their worldwide distributed driver network and companies like Google and Facebook are using it as well. So, this is really at the highest volumes of data, the most complex compute, but as I've mentioned, I've been working in the world of genomics and I'm starting to applications from some of these companies be applicable in other domains, such as genomics. So, we're going to take a look. Now, one thing…
Practice while you learn with exercise files
Download the files the instructor uses to teach the course. Follow along and learn by watching, listening and practicing.
Contents
-
-
-
-
-
-
Use Databricks jobs and role-based control5m 37s
-
(Locked)
Use Databricks Runtime ML2m 52s
-
(Locked)
Understand ML Pipelines API4m 16s
-
(Locked)
Use ML Pipelines API8m 39s
-
(Locked)
Use distributed ML training9m 59s
-
(Locked)
Understand Databricks Delta3m 41s
-
(Locked)
Use Databricks Delta5m 10s
-
(Locked)
Use Azure Blob storage2m 41s
-
(Locked)
Understand MLflow7m 34s
-
-
-