Deploying Machine Learning Solutions
This course covers the important conceptual reasons why models underperform post-deployment, the actual implementation of model deployment using Python Flask, using serverless, cloud-based compute options and using platform-specific machine learning frameworks.
What you'll learn
Machine Learning is exploding in popularity, but serious early warning signs are emerging around the performance of ML models in production.
In this course, Deploying Machine Learning Solutions you will gain the ability to identify reasons why models might be under-performing in production after doing just fine in training and testing, and ways to mitigate this worrying phenomenon.
First, you will learn how training-serving skew, concept drift, and overfitting are different causes of model underperformance, and how they can be mitigated by post-deployment monitoring.
Next, you will discover how ML models can be deployed, that is made available on HTTP endpoints, using Flask, the popular Python web-serving framework. You will also see how you can deploy models to serverless environments such as Google Cloud Functions
Finally, you will work with platform-specific machine learning services such as Google AI Platform and Amazon SageMaker for model deployment.
When you’re finished with this course, you will have the skills and knowledge to identify issues with models that have been deployed but are not performing to expectations, as well as how to implement deployment using both on-prem and cloud infrastructure.
Table of contents
- Module Overview 1m
- Serializing Model Parameters 4m
- Demo: Serializing and Deserializing Models Using JSON 7m
- Demo: Using Pickle and Joblib to Serialize and Deserialize Models 5m
- Demo: Checkpointing Models and Resuming Training from a Checkpoint 6m
- Demo: Serializing Pre-processors and Models 6m
- Demo: Serializing Pipelines 2m
- Using Flask for Model Deployment 2m
- Demo: Deploying a Model for Prediction Using Flask 7m
- Module Summary 1m
- Module Overview 1m
- Introducing the Google AI Platform 6m
- Demo: Getting Started with Cloud AI Platform 2m
- Demo: Creating a Model and a Version 5m
- Demo: Scheduling an Evaluation Job to Sample Prediction Instances 5m
- Demo: Testing the Deployed Model Using the Web Console 3m
- Demo: Model Predictions Using the gcloud Command Line Utility 3m
- Demo: Invoking the Predictions API Using cURL 4m
- Demo: Monitoring Deployed Models Using Stackdriver 7m
- Module Summary 1m
- Module Overview 1m
- Introducing Amazon SageMaker 2m
- Training a Model on SageMaker 3m
- Deploying a Model on SageMaker 3m
- Demo: Creating a SageMaker Notebook Instance 5m
- Demo: Getting Started with SageMaker for Distributed Training 3m
- Demo: Tensor Flow Script for Distributed Training 6m
- Demo: Distributed Training Using the SageMaker Tensor Flow Estimator 6m
- Demo: Deploying the Model for Predictions 5m
- Demo: Auditing and Compliance Using Cloud Trail 4m
- Summary and Further Study 2m
About the author
Janani has a Masters degree from Stanford and worked for 7+ years at Google. She was one of the original engineers on Google Docs and holds 4 patents for its real-time collaborative editing framework. After spending years working in tech in the Bay Area, New York, and Singapore at companies such as Microsoft, Google, and Flipkart, Janani finally decided to combine her love for technology with her passion for teaching. She is now the co-founder of Loonycorn, a content studio focused on providing ... more