Update all w8a8ch definitions' correctness reference, aligns with what llama.cpp low-bit kernel correctness reference style: dequant input, then calculate in full precision
ArtificialRay
009b323
Definitions, solutions, and evaluation trace updates for ncnn kernels
ArtificialRay
07f5581
- update ncnn-arm baseline and reference-scalar solution and claude-sonnet-4-6 agent eval solution trace
- update evaluation traces for conv2d, conv2d depthwise, gemm, pooling
- update low-bit kernel definition for conv2d, conv2d_depthwise, gemm
add baseline solution source at all neural network kernel definitions
ArtificialRay
c211fb1
Update definition and workload for CV model kernels:
ArtificialRay
7e7beda
- only conv2d and conv2d depthwise are kept, remove others
- conv2d and conv2d depthwise have fp32 and int8 percision kernels
- add new maxpool2d definitions
- add workloads for new conv2d, conv2d depthwise definitions, pooling, lstm definitions.
