James Montantes currently works for Exxact Corporation. They are a leading provider of accelerated systems for deep learning, molecular dynamics, and other HPC applications. Aside from performing various marketing activities, he conducts GPU benchmarks for deep learning, and train deep neural networks using Exxact systems. He has about 15 years of experience in market research, product development, and prototyping. Has has held many roles from marketing to systems engineering, and his current technology interests include: human-machine interfaces (HMI), deep learning/AI research, and distributed consensus networks.
We discuss the classes that PyTorch provides for helping with Natural Language Processing (NLP) and how they can be used for related tasks using recurrent layers.
We explain deep compression for improved inference efficiency, mobile applications, and regularization as technology cozies up to the physical limits of Moore's law.