HC2023-T1.2: ML Inference

Tutorial 1, part 2, Hot Chips 2023, Sunday, August 27, 2023.
Organizers: Nathan Kalyanasundharam, CXL Board & AMD
This tutorial gives a brief introduction to basic concepts underlying ML inference and then gives overviews of several hot areas where current research is improving the performance and capabilities of ML inference. The hot areas covered in this part of the tutorial are the latest version of the PyTorch software framework and how to utilize sparse non-zero weight values to improve inference performance.
PyTorch 2.0
Elias Ellison, Meta
Hardware Requirements for Exploiting Sparsity in ML Inference
Zhibin Xiao, Moffett AI