Speaker: Dr. Peng Wu, Engineering Manager, META
https://www.linkedin.com/in/peng-wu-411b5b/
Title: PyTorch 2.0 and its Compiler Technologies
Abstract
PyTorch 2.0 (PT2) was unveiled in March 2023 and brought substantial performance improvements across a diverse array of models, often by adding just a couple of lines of code to user models. But do not mistake it as the end of the story.
The first release of PT2 began a long-term technical roadmap to improving PyTorch’s execution efficiency via compiled mode. This talk will delve into the design and development of the PyTorch Compiler. We will examine critical aspects through the lens of a three-year timeframe. We will highlight our unique approach to creating an industry-leading ML compiler in a highly competitive ML landscape. To attest to the fast-moving nature of ML, we will share how we responded to the rise of single-purpose ML frameworks in the Generative AI and LLM domains.
Bio
Dr. Peng Wu is the engineering manager of the PyTorch Compiler team at Meta. Dr. Wu spent over a decade at IBM research, working on many aspects of programming systems. She then founded the Programming Technologies Lab at Huawei and led its growth for six years. At Meta, she supported the team’s search for the right compiler solutions for PyTorch over the last four years, culminating in the release of PyTorch 2.0 in 2023. She holds a PhD in Computer Science from the University of Illinois, Urbana-Champaign.
Date and Time: November 1, 2024, 11am-noon ET / 10-11am CT
Location: Coordinated Science Lab Auditorium (B02), 1308 W. Main St., Urbana IL 61801
This will be a wonderful opportunity to connect with other experts and engage in fruitful discussions.
https://go.illinois.edu/IIDAISeminar