Reduced int precision for texture coordinates in q_linear op, to reduce shader register pressure.
Differential Revision: D64191093 Pull Request resolved: https://github.com/pytorch/executorch/pull/6354
T
trivedivivek committed
5d7bd718cd72d9407dddd2af4fc55946ed92f11d
Parent: 5a6a571
Committed by GitHub <noreply@github.com>
on 10/22/2024, 1:40:20 PM