DSCI-272: Predicting with MLOps on Cloudera AI
Enterprise data science teams need collaborative access to business data, tools, and computing resources required to develop and deploy machine learning workflows. Cloudera Machine Learning (CML), part of the Cloudera Data Platform (CDP), provides the solution, giving data science teams the required resources.This course covers machine learning workflows and operations using CML. Participants explore, visualize, and analyze data. You will also train, evaluate, and deploy machine learning models. The course walks through an end-to-end data science and machine learning workflow based on realistic scenarios and datasets from a fictitious technology company. The demonstrations and exercises are conducted in Python (with PySpark) using CML.
28 hours
VirtualThrough lecture and hands-on exercises, you will learn how to:Utilize Cloudera SDX and other components of the Cloudera Data Platform to locate data for machine learning experimentsUse an Applied ML Prototype (AMP)Manage machine learning experimentsConnect to various data sources and explore dataUtilize Apache Spark and Spark MLDeploy an ML model as a REST APIManage and monitor deployed ML models
Introduction to CMLOverview.CML Versus CDSW.ML Workspaces.Workspace Roles.Projects and Teams.Settings.Runtimes/Legacy Engines.Introduction to AMPs and the WorkbenchEditors and IDE.Git.Embedded Web Applications.AMPs.Data Access and LineageSDX Overview.Data Catalog.Authorization.Lineage.Data Visualization in CMLeData Visualization Overview.CDP Data Visualization Concepts.Using Data Visualization in CML.ExperimentsExperiments in CML.An Introduction to the CML Native WorkbenchEntering Code.Getting Help.Accessing the Linux Command Line.Working With Python Packages.Formatting Session Output.Spark OverviewHow Spark Works.The Spark Stack.File Formats in Spark.Spark Interface Languages.Introduction to PySpark.How DataFrame Operations Become Spark Jobs.How Spark Executes a Job.Running a Spark ApplicationRunning a Spark Application.Reading data into a Spark SQL DataFrame.Examining the Schema of a DataFrame.Computing the Number of Rows and Columns of a DataFrame.Examining a Few Rows of a DataFrame.Stopping a Spark Application.Inspecting a Spark DataFrameInspecting a DataFrame.Inspecting a DataFrame Column.Transforming DataFramesSpark SQL DataFrames.Working with Columns.Working with Rows.Working with Missing Values.Transforming DataFrame ColumnsSpark SQL Data Types.Working with Numerical Columns.Working with String Columns.Working with Date and Timestamp Columns.Working with Boolean Columns.Complex TypesComplex Collection Data Types.Arrays.Maps.Structs.User-Defined FunctionsUser-Defined Functions.Example 1: Hour of Day.Example 2: Great-Circle Distance.Reading and Writing DataFramesWorking with Delimited Text Files.Working with Text Files.Working with Parquet Files.Working with Hive Tables.Working with Object Stores.Working with Pandas DataFrames.Combining and Splitting DataFramesCombining and Splitting DataFrames.Joining DataFrames.Splitting a DataFrame.Summarizing and Grouping DataFramesSummarizing Data with Aggregate Functions.Grouping Data.Pivoting Data.Window FunctionsWindow Functions.Example: Cumulative Count and Sum.Example: Compute Average Days Between Rides for Each Rider.Machine Learning OverviewIntroduction to Machine Learning.Machine Learning Tools.Apache Spark MLlibIntroduction to Apache Spark MLlib.Exploring and Visualizing DataFramesPossible Workflows for Big Data.Exploring a Single variable.Exploring a Pair of Variables.Monitoring, Tuning, and Configuring Spark ApplicationsMonitoring Spark Applications.Configuring the Spark Environment.Fitting and Evaluating Regression ModelsAssemble the Feature Vector.Fit the Linear Regression Model.Fitting and Evaluating Classification ModelsGenerate Label.Fit the Logistic Regression Model.Tuning Algorithm Hyperparameters Using Grid SearchRequirements for Hyperparameter Tuning.Tune the Hyperparameters Using Holdout Cross- Validation.Tune the Hyperparameters Using K-Fold Cross- Validation.Fitting and Evaluating Clustering ModelsPrint and Plot the Home Coordinates.Fit a Gaussian Mixture Model.Explore the Cluster Profiles.Processing Text: Fitting and Evaluating Topic ModelsFit a Topic Model Using Latent Dirichlet Allocation.Fitting and Evaluating Recommender ModelsRecommender Models.Generate Recommendations.Working with Machine Learning PipelinesFit the Pipeline Model.Inspect the Pipeline Model.Applying a Scikit-Learn Model to a Spark DataFrameBuild a Scikit-Learn Model.Apply the Model Using a Spark UDF.Deploying a Machine Learning Model as a REST API in CMLLoad the Serialized Model.Define a Wrapper Function to Generate a Prediction.Test the Function.Autoscaling, Performance, and GPU SettingsAutoscaling Workloads.Working with GPUs.Model Metrics and MonitoringWhy Monitor Models?.Common Models Metrics.Models Monitoring With Evidently.Continuous Model Monitoring.Appendix: Workspace ProvisioningWorkspace and Environment.
The course is designed for data scientists who need to understand how to utilize Cloudera Machine Learning and the Cloudera Data Platform to achieve faster model development and deliver production machine learning at scale. Data engineers, developers, and solution architects who collaborate with data scientists will also find this course valuable.
