Guides to bringing your code from various Machine Learning frameworks to Google Cloud Platform.
The goal is to present recipes and practices that will help you spend less time wrangling with the various interfaces and more time exploring your datasets, building your models, and in general solving the problems you really care about.
-
Genomic ancestry inference with deep learning - Ancestry inference on Google Cloud Platform using the 1000 Genomes dataset
-
Running TensorFlow inference workloads at scale with TensorRT 5 and NVIDIA T4 GPUs - Creating a demo of ML inference using Tesla T4, TensorFlow, TensorRT, Load balancing and Auto-scale.
- Estimators - A guide to the Estimator interface.
-
scikit-learn on GCE - Train a simple model with scikit-learn on a Google Compute Engine
-
Model serve - Serve model with Google App Engine and Cloud Endpoints.
-
Hyperparameter search - Hyperparameter search on a Google Kubernetes Engine cluster from a Jupyter notebook.
-
Compute Engine survival training - Introduces a framework for running resilient training jobs on Google Compute Engine.
-
Compute Engine burst training - A guide to using powerful VMs to quickly and cheaply perform computationally intensive training jobs. (The example training job in this guide uses xgboost as well as scikit-learn.)