[GitHub] [incubator-tvm] FrozenGene commented on a change in pull request #5754: [RFC] Improve quantized convolution performance for armv8 architectures

FrozenGene commented on a change in pull request #5754: URL: https://github.com/apache/incubator-tvm/pull/5754#discussion_r443664602 ## File path: python/tvm/relay/qnn/op/legalizations.py ## @@ -237,17 +237,23 @@ def is_fast_int8_on_arm(): target =

[GitHub] [incubator-tvm] FrozenGene commented on a change in pull request #5754: [RFC] Improve quantized convolution performance for armv8 architectures

FrozenGene commented on a change in pull request #5754: URL: https://github.com/apache/incubator-tvm/pull/5754#discussion_r442604112 ## File path: src/relay/op/nn/convolution.h ## @@ -383,6 +383,38 @@ inline bool Conv2DWinogradWeightTransformRel(const Array& types, int num_i

[GitHub] [incubator-tvm] FrozenGene commented on a change in pull request #5754: [RFC] Improve quantized convolution performance for armv8 architectures

FrozenGene commented on a change in pull request #5754: URL: https://github.com/apache/incubator-tvm/pull/5754#discussion_r440237827 ## File path: python/tvm/relay/op/nn/nn.py ## @@ -2134,6 +2202,25 @@ def contrib_conv2d_winograd_weight_transform(weight, return