Static Memory Planning for Embedded Neural Network Inference

Peak working memory depends on the execution schedule and allocation, so the schedule can be optimized ahead of time and executed with bounded runtime arenas.

June 11, 2026 in Deep Dives9 minutes