BROADCASTS.com

Jeff Boudier from Hugging Face joins host Jake Smith to talk about the company\u2019s open source machine learning transformers (also known as \u201cpytorch-pretrained-bert\u201d) library. Jeff talks about how transformers have accelerated the proliferation of natural language process (NLP) models and their future use in objection detection and other machine learning tasks. He goes into detail about Optimum\u2014an open source library to train and run models on specific hardware, like Intel Xeon CPUs, and the benefits of the Intel Neural Compressor, which is designed to help deploy low-precision inference solutions. Jeff also announces Hugging Face\u2019s new Infinity solution that integrates the inference pipeline to achieve results in milliseconds wherever Docker containers can be deployed. \n\nFor more information, visit: https://hf.co/\nFollow Jake on Twitter at: https://twitter.com/jakesmithintel