Google Cloud AI is faster and more flexible
The Google Cloud AI platform has been updated with new features aimed at simplifying the deployment of the models. Custom containers are now available, and you can choose between different virtual machine configurations.
Google has introduced important updates for its Cloud AI platform dedicated to the development of applications based on machine learning algorithms, making it faster and above all more flexible than in the past.
Google Cloud AI supports hardware acceleration on GPU
One of the main advantages of the Google Cloud AI platform is the ability to train AI models in a serverless environment, directly on the cloud, avoiding developers having to deal with infrastructure management: the developers just need to indicate which of the four available configurations to use for the calculation, and the system allocates the resources and configure the virtual machines.
With the new update to the four available configurations, new ones have been added, and acceleration is now available on NVIDIA T4 GPU, which can speed up some types of calculation enormously – for example: the one that require to process text, audio or images.
The new backend is based on a microservice architecture built around the Google Kubernetes Engine and now allows you to record online prediction requests on BigQuery, so that you can analyze them carefully and evaluate if and how much it is necessary to carry out a new training of the model of AI.
Google Cloud AI and Wildlife Insights: A concrete example
Among the examples of real use of Google Cloud AI, it is worth mentioning that of Wildlife Insights, an American company that has set itself the ambitious goal of analyzing and cataloging information on wild animal species in real time.
” Wildlife Insights transforms millions of animal images into a dataset that helps us understand the behaviors of animals, so as to protect them and preserve their species ” – commented Eric H. Fegraus, Senior Director of Conservation Technology – ” The Google platform Cloud AI helps us generate reliable machine learning models and easily integrate their predictions into our applications. Fast predictions within a scalable GPU-based environment are critical to the user experience. ”
More information on the news of Google Cloud AI is available on the official blog of the platform.
