projects / platform / upstream / opencv.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: b91f173) | patch
Merge pull request #21910 from zihaomu:fast_conv_ARM
authorZihao Mu <zihaomu@outlook.com>
Fri, 1 Jul 2022 10:03:15 +0000 (18:03 +0800)
committerGitHub <noreply@github.com>
Fri, 1 Jul 2022 10:03:15 +0000 (13:03 +0300)
commit59b870a87ad22d89027e28eee3b8be1f3db1c0bf
tree9e72e9c97424678fc8cad86a4a28b0e349e2d3a4tree | snapshot
parentb91f173680af92498a8bd2517fc7c524c2eae640commit | diff
Merge pull request #21910 from zihaomu:fast_conv_ARM

DNN: Accelerating convolution

* Fast Conv of ARM, X86 and universal intrinsics.

* improve code style.

* error fixed.

* improve the License

* optimize memory allocated and Adjust the threshold.

* change FasterRCNN_vgg16 to 2GB memory.
modules/dnn/src/layers/convolution_layer.cpp diff | blob | history
modules/dnn/src/layers/fast_convolution/depthwise_convolution.cpp [new file with mode: 0644] blob
modules/dnn/src/layers/fast_convolution/fast_convolution.avx2.cpp [new file with mode: 0644] blob
modules/dnn/src/layers/fast_convolution/fast_convolution.cpp [new file with mode: 0644] blob
modules/dnn/src/layers/fast_convolution/fast_convolution.hpp [new file with mode: 0644] blob
modules/dnn/src/layers/fast_convolution/fast_convolution.simd.hpp [new file with mode: 0644] blob
modules/dnn/src/layers/fast_convolution/winograd_3x3s1_f63.cpp [new file with mode: 0644] blob
modules/dnn/test/test_backends.cpp diff | blob | history
modules/dnn/test/test_caffe_importer.cpp diff | blob | history
modules/dnn/test/test_int8_layers.cpp diff | blob | history
modules/dnn/test/test_tf_importer.cpp diff | blob | history
Domain: Multimedia / Media Vision;