AWS SageMaker Neo Introduction
Run your training models anywhere with double the performance
ML (Machine Learning) makes it possible to accomplish business goals by manipulating information collected from a server. SageMaker Neo is an Amazon Web Services machine learning(ML) resource that makes it easier to use artificial intelligence to super power your business production through smarter predictions. You’ll save time and still be accurate when using your ML training models. Neo is automatically optimized to give you double the operating power for use on any ML framework and target hardware platform. The open source documentation gives developers the ability to open the hood of Neo and customize their ML environment. Neo offers numerous benefits that can be used on any server.
Neo is a resource that makes it possible to train a model once, and then run that model on your cloud servers or on the edge location servers offered through AWS. With the optimized nature of Neo, it is possible to run functions on AWS edge locations, which saves time and memory without losing data. Edge locations are special data centers that have servers located closer to the users who call your data. More power is possible because a tenth of the memory footprint is used through traveling shorter distances to call servers at edge locations. The operational power of your apps can be increased by double through Neo.
A lot of traditional problems faced with manual tuning models are fixed with the automatic optimization provided through NEO. Real time, low-latency predictions made by ML models make it faster to learn off of new data. This is important IoT (Internet-of-Things) devices because they operate at the edge. Imagine a self-driving car or a deep lens camera that had to travel to the furthest server to retrieve data instead of an edge location server. That would be a nightmare scenario where the device relies on the freshest, and truest information to continue to adapt to its environment. Tesla cars use ML to decide how to change lanes. Companies that ship millions of products rely on ML to keep a smooth stream of supply. Ride share apps use data to update customers on the estimated time of arrival for their drivers. The fully managed nature of SageMaker Neo makes it possible for developers to build better solutions by capitalizing on the data collected at the edge. Browse the open source documentation to learn more about Neo.
How it works:
- Choose a framework to build an ML model
- AWS SageMaker trains and tunes model automatically
- Choose target hardware platform
- Neo optimizes the trained model for target hardware platform
- Deploy model on cloud or edge