fix(quantize): Overhaul INT8 static quantization to use QDQ format, unlocking TensorRT/NPU hardware acceleration#307

Open

jay7-tech wants to merge 2 commits intoopencv:mainfrom

jay7-tech:feature/gsoc-qdq-hardware-accel

Commits on Mar 4, 2026

fix(benchmark): Auto-route INT8 precision to NPU/TIM-VX backend on aarch64 hardware to resolve quantization paradox
jay7-tech
committed
fix(quantize): Overhaul INT8 static quantization to use QDQ format, unlocking TensorRT/NPU hardware acceleration
jay7-tech
committed