Learn how to design machine learning solutions with Google Cloud Platform. Review services such as AutoML, CloudML Engine, and the GCP machine learning APIs.
Overview
Syllabus
Introduction
- Build complete solutions with machine learning services
- What you should know
- About using cloud services
- Business scenarios for machine learning
- Which algorithm should you use?
- GCP AI servers vs. platforms
- Enable GCP ML APIs
- Data preparation with Cloud Dataflow and Cloud Dataprep
- An ML notebook in action: Colaboratory
- An ML notebook in action: Set up Cloud Datalab
- An ML notebook in action: Use Cloud Datalab
- Overview of GCP ML APIs
- Predict via the Cloud Vision API for images
- Predict via the Cloud Video Intelligence API for video
- Predict via the Natural Language API for NLP
- Predict via the Text-to-Speech API
- Predict via the Speech-to-Text API
- Predict via the Cloud Translation API
- Predict via BigQuery ML
- Understand Cloud AutoML services
- Understand AutoML Vision
- Prepare data and labels for AutoML Vision
- Train model for AutoML Vision
- Evaluate model with AutoML Vision
- Predict using a trained AutoML Vision model
- Why build custom ML models?
- Using containers to host ML models
- Use Cloud ML Engine
- Evaluate Cloud ML Engine output
- Scale custom ML models
- Understanding deep learning
- Work with TensorBoard
- Work with Keras for TensorFlow
- GPUs and TPUs for TensorFlow
- TensorFlow for JavaScript and mobile
- Chatbot with ML
- Image search with Cloud Vision and Cloud ML
- GCP serverless machine learning architecture
- GCP machine learning with structured data
- GCP ML service for IoT apps
- Next steps
Taught by
Lynn Langit