Community Keynote

March 17th, 2023
4:05PM - 4:40PM

Whether you’re new to TVM, a longtime contributor, or anywhere in between, TVM is now easier to use than ever before! Since the official unveiling of TVM Unity at TVMCon 2021, TVM has evolved towards a unified, python-first representation of ML workloads at every level of abstraction, which has enabled composable ML compilation pipelines which interoperate seamlessly with frameworks, libraries, and hardware. This has empowered users of all backgrounds to leverage TVM Unity’s collection of optimizations and universal deployment infrastructure alongside the rest of the ML ecosystem.