Skip to content

TensorFlow Lite delegate and Vivante NPU driver

Tomeu Vizoso requested to merge tomeu/mesa:teflon into main

These changes allow to run inference on a NPU from VeriSilicon that is closely based on the Vivante GPU IP.

Still need to do a fair amount of cleanup, but I am pushing it here for those who may be curious and may even want to comment.

I'm still not 100% sure I will be at XDC next week, but in case I am, I will love to discuss this in person!

Merge request reports