The Toronto ML Systems Group: Transformer Accounting

​Come hang out with the Toronto ML Systems Group for an evening on Transformer Accounting. Here, we intend to build a deep intuition around transformer math and what parameters cost in data, compute, memory, and GPU time.

This is the fifth session of the Toronto ML Systems Group, a recurring meetup for people who want to get serious about the full stack: chips, compilers, kernels, model architectures, distributed training, inference optimization, and everything in between.

​Relevant resources: https://jax-ml.github.io/scaling-book/transformers/

⏱️ Schedule
6:00 PM - 6:30 PM Check-in, Food, and networking

​6:30 PM - 7:15 PM Lecture: Transformer Accounting, Speaker: Shashank Shekhar

​7:15 PM - 7:30 PM Discussion
7:30 PM - 8:30 PM Networking and wrap-up

​If you’re interested in speaking at or sponsoring our future events, let us know:)