AI Platform Training brings the power and flexibility of TensorFlow, scikit-learn, XGBoost, and custom containers to the cloud. You can use AI Platform Training to train your machine learning models using the resources of Google Cloud.
Getting started
-
Introduction to AI Platform
An overview of AI Platform products.
-
Training overview
An introduction to training machine learning models on AI Platform Training.
-
Development environment
Requirements for your local development environment.
-
Training with TensorFlow 2
Details about training with TensorFlow 2.
-
Getting started: training and prediction with TensorFlow Keras
Train a TensorFlow Keras model in AI Platform Training and deploy the model to AI Platform Prediction.
-
Getting started: training and prediction with TensorFlow Estimator
Train a TensorFlow Estimator model in AI Platform Training and deploy the model to AI Platform Prediction.
-
Getting started with scikit-learn and XGBoost
Train a scikit-learn or XGBoost model.
-
Getting started with custom containers
Train a PyTorch model by using a custom container.
Training workflow
-
Packaging a training application
Package your Python training code to make it compatible with AI Platform Training.
-
Running a training job
Run a training job.
-
Specifying machine types or scale tiers
Configure which types of virtual machines your training job uses.
-
Monitoring training jobs
Monitor the status of your training jobs with logs and resource utilization metrics.
Training at scale
Hyperparameter tuning
Accelerators
Custom containers
-
Overview of containers
An introduction to how you can customize your training jobs by providing your own Docker container.
-
Training with containers on AI Platform
Create a custom Docker container and use it to run a training job.
-
Distributed training with containers
Configure a custom container job to use multiple virtual machines.
Integrating with tools and services
Monitoring and security
-
Viewing audit logs
Monitor admin activity and data access with Cloud Audit Logs.
-
Access control with IAM
An overview of permissions required to perform various actions in the AI Platform Training and Prediction API, as well as IAM roles that provide these permissions.
-
Training with a custom service account
Use a custom service account for your training application
-
Using VPC Service Controls with Training
Mitigate the risk of data exfiltration by using a service perimeter.
-
Using customer-managed encryption keys (CMEK)
Encrypt training job data with customer-managed encryption keys.
-
Monitoring and debugging training with an interactive shell
Learn how to use an interactive shell to inspect your training container while it runs.
AI Platform Training resources
Tutorials
Runtime versions
AI Platform built-in algorithms
-
Introduction to built-in algorithms
An overview of built-in algorithms.
-
Preprocessing data for tabular built-in algorithms
Use automatic preprocessing to prepare your data for training.
-
Getting started with the linear learner algorithm
Train a model with the built-in TensorFlow linear learner algorithm.
-
Training using the built-in linear learner algorithm
Customize how you use the built-in linear learner algorithm for training.
-
Linear learner algorithm reference
Configuration options for the built-in linear learner algorithm.
-
Getting started with the wide and deep algorithm
Train a model with the built-in TensorFlow wide and deep algorithm.
-
Training using the built-in wide and deep algorithm
Customize how you use the built-in wide and deep algorithm for training.
-
Wide and deep algorithm reference
Configuration options for the built-in wide and deep algorithm.
-
Getting started with the XGBoost algorithm
Train a model with the built-in XGBoost algorithm.
-
Training using the built-in XGBoost algorithm
Customize how you use the built-in XGBoost algorithm for training.
-
Training using the built-in distributed XGBoost algorithm
Customize how you use the distributed version of the built-in XGBoost algorithm.
-
XGBoost algorithm reference
Configuration options for the built-in XGBoost algorithm.
-
Getting started with the image classification algorithm
Train a model with the built-in image classification algorithm.
-
Training using the built-in image classification algorithm
Customize how you use the built-in image classification algorithm for training.
-
Image classification algorithm reference
Configuration options for the built-in image classification algorithm.
-
Getting started with the image object detection algorithm
Train with the built-in image object detection algorithm.
-
Training using the built-in image object detection algorithm
Customize how you use the built-in image object detection algorithm for training.
-
Image object detection algorithm reference
Configuration options for the built-in image object detection algorithm.