days
-3
-6
hours
-2
-3
minutes
-3
0
seconds
-2
-8
search
TensorFlow Enterprise and TensorBoard.dev

Google launches TensorFlow Enterprise for AI in the cloud

Maika Möbus
tensorflow
© Shutterstock / issaro prakalung

Google introduced TensorFlow Enterprise, a new collection of machine learning services and products. The beta version includes managed services, and some versions will receive long-term version support for up to three years. At the same time, Google unveiled the new website TensorBoard.dev.

Google has provided its machine learning platform TensorFlow with two new software additions. The first is TensorFlow Enterprise for ML in the cloud, aimed at business customers. The second, TensorBoard.dev, was developed for sharing ML results. Let’s take a closer look at their features.

TensorFlow Enterprise

TensorFlow Enterprise supports services like AI Platform and Kubernetes Engine for deploying and developing ML applications in the Google Cloud. It is a combination of services and products that are specifically targeted to the demands of enterprise customers.

TensorFlow Enterprise was shown to outperform TensorFlow 1.14 in terms of speed. Some versions will receive long-term support, as security patches and bug fixes are provided for up to three years. TensorFlow Enterprise can be used with containers and virtual machines, and offers automatic provisioning, optimizing and scaling.

Currently still a beta release, TensorFlow Enterprise is available at no additional cost.

TensorBoard.dev

Along with TensorFlow Enterprise, Google simultaneously launched the website TensorBoard.dev. It enables users to share the results of machine learning experiments for team collaboration, publications, etc. TensorBoard.dev is free and comes with storage for up to 10 million data points out of the box. It provides the familiar experience of TensorFlow’s visualization toolkit TensorBoard:

SEE ALSO: Reinforcement learning: a gentle introduction and industrial application

An example in Google Colaboratory, a free Jupyter notebook environment, demonstrates how to train a simple model and upload the TensorBoard logs. Additionally, Google offers a guide to get you started:

TensorBoard.dev is offered as a preview version and comes with support for the Scalars dashboard. Google plans to implement other dashboards in the future.

Author
Maika Möbus
Maika Möbus has been an editor for Software & Support Media since January 2019. She studied Sociology at Goethe University Frankfurt and Johannes Gutenberg University Mainz.

Leave a Reply

Be the First to Comment!

avatar
400
  Subscribe  
Notify of