We are excited to announce the availability of the CloudML package, which provides an R interface to the Google Cloud Machine Learning Engine. CloudML provides a number of services including on-demand access to training on GPUs and hyperparameter tuning to optimize key features of the model architecture.
Overview
We are excited to announce the availability of cloudml package, which provides an R interface. Google Cloud Machine Learning Engine. CloudML provides a number of services including:
-
Scalable training of models built with Kira, tfestimatorsAnd Tensor flow R packages.
-
On-demand access to training on GPUs, including new ones Tesla P100 GPUs From NVIDIA®.
-
Hyperparameter tuning to optimize key attributes of model architectures to maximize prediction accuracy.
-
Deploying models trained on Google's global prediction platform can support thousands of users and TBs of data.
Training with CloudML
Once you've configured your system to publish to CloudML, training the model is as straightforward as cloudml_train() Function:
library(cloudml)
cloudml_train("train.R")CloudML provides a variety of GPU configurations, which can be easily selected when calling. cloudml_train(). For example, the following will train the same model as above but with a Tesla K80 GPU:
cloudml_train("train.R", master_type = "standard_gpu")To train using a Tesla P100 GPU You will explain "standard_p100":
cloudml_train("train.R", master_type = "standard_p100")When the training is complete, the work is collected and the training run report is displayed:
Learn more
check CloudML package documentation To get started with training and deploying models on CloudML.
You can learn more about the various capabilities of CloudML in these articles:
-
Training with CloudML The training goes into additional depth on managing jobs and their production.
-
Hyperparameter tuning It explores how you can improve the performance of your models by running several trials with different hyperparameters (eg number and size of layers) to determine the optimal values.
-
Google Cloud Storage Provides information on copying data between your local machine and Google Storage and also explains how to use data in Google Storage during training.
-
Deployment of models Describes how to deploy trained models and generate predictions from them.
Reuse
Text and figures are licensed under a Creative Commons Attribution. CC BY 4.0. Figures reused from other sources are not covered by this license and may be identified by a note in their caption: “Figure from …”
Reference
For attribution, please cite this work.
Allaire (2018, Jan. 10). Posit AI Blog: R Interface to Google CloudML. Retrieved from
BibTeX reference
@misc{allaire2018r,
author = {Allaire, J.J.},
title = {Posit AI Blog: R Interface to Google CloudML},
url = {},
year = {2018}
}