SpringML Inc. (“SpringML”), a leader in machine learning and advanced data analytics services, is proud to announce TPU accelerator packages to help customers migrate to TPU for data intensive use cases.
There is a common thread that connects Google services such as Google Search, Street View, Google Photos and Google Translate: they all use Google’s Tensor Processing Unit, or TPU, to accelerate their neural network computations behind the scenes.
We are excited about the launch of TPUs and the opportunity to help our customers take advantage of this new capability. In a recent implementation, we have seen remarkable improvements in execution speed when processing millions of images that were trained on 10 to 20 layer CNN models.
To help clients quickly migrate their models to TPU and take advantage of the processing capabilities, we are introducing a “jumpstart” package that is focused on both the design and deployment requirements for TPU models.
“SpringML is constantly looking for ways to simplify deployment of machine learning models, and we believe TPU is a great option for customers who are looking to solve business challenges that include complex vision and natural language use cases,” said Prabhu Palanisamy, Chief Strategy Officer of SpringML.