Ambarella, an AI vision silicon company made public in 2011, has inked a deal with Amazon Web Services (AWS) to introduce Amazon SageMaker Neo to school machine learning (ML) models once and deploy them on device equipped with an Ambarella CVflow®-powered artificial vision (AI) vision system on chip (SoC). Ambarella collaborated with AWS to streamline the process by amalgamating Amazon SageMaker Neo cloud service with the Ambarella toolchain. Apparently, developers can bring forth their schooled models to Amazon SageMaker Neo and automatically boost the model for Ambarella CVflow-powered SOCs. Reportedly, customers will have the leeway to build a machine learning (ML) model using TensorFlow, MXNet, XGBoost, PyTorch and train the model with the help of Amazon SageMaker either in the cloud or on the local machine. It is worth noting that the propelled model operates in the Amazon SageMaker Neo runtime available for the Ambarella SDK and meant for Ambarella SOCs. According to sources, the Amazon SageMaker Neo runtime holds less than 10x the memory and disk footprint of PyTorch, MXNet or TensorFlow, thereby making it apt to deploy ML models on connected cameras. Ambarella's Vice president of marketing and business development Chris Day was quoted saying that Ambarella is witnessing an influx of production with CVflow AI vision processors meant for the enterprise video security, home monitoring and automotive markets. Day went on to stress that the potential inroads of next generation of AI-enabled products would be bolstered by the collaboration that will help the end-users choose an Ambarella SoC and collate a trained ML model with a single click. Amazon SageMaker is a fully managed service which offers every data scient the innate ability to train, build and use ML models robustly, Vice President of Machine Learning & Engines at Amazon Web Services, Inc., Bratin Saha elucidated. Source credit: https://www.businesswire.com/news/home/20200102005052/en/Ambarella-Enables-Artificial-Intelligence-Wide-Range-Connected