Research to Production at Amazon Scale With PyTorch (PyTorch Conference 2022 Breakout Session)
Watch Kiuk Chung & Pooja Maknikar from Amazon present their PyTorch Conference 2022 Breakout Session "Research to Production at Amazon Scale With PyTorch".
The M5 team at Amazon builds state-of-the-art semantic representations of Amazon-specific entities such as products, shopping sessions, and reviews which are used by various machine learning systems across Amazon. Due to the scale, dynamic, and diverse nature of Amazon’s AI use-cases, M5’s AI platform was designed to enable high velocity research and experimentation paired with low-friction productization. With PyTorch being the deep learning library of choice, we’ve built a Research to Production (R2P) stack servicing over a hundred teams at Amazon. In this session we dive into the journey of M5 from zero to success and discuss the challenges of using PyTorch on our infrastructure, the solutions we came up with, and lessons learned.
Visit our website: https://pytorch.org/
Read our blog: https://pytorch.org/blog/
Follow us on Twitter: / pytorch
Follow us on LinkedIn: / pyto. .
Follow us on Facebook: / pytorch
#PyTorch #ArtificialIntelligence #MachineLearning