Administrator/opencv-MIRROR
1
0
Fork 0
mirror of https://github.com/opencv/opencv.git synced 2026-11-09 18:04:29 -06:00
Code Issues Packages Projects Releases Wiki Activity

dnn(onnx): enable ONNX AveragePool with dilations conformance tests

Browse Source 
- Accept plural attribute names (kernel_shape, strides, dilations, pads) in AveragePoolLayer
- Un-denylist 6 official ONNX AveragePool dilation conformance tests in parser denylist
- Conformance tests verified on CPU: 2D and 3D dilated average pooling passing (6/6)
This commit is contained in:
Tushar Burla
2026-11-02 18:25:22 +05:30
parent 739c1e60fd
commit ee27ba2059
2 changed files with 8 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
modules/dnn/src/layers/avgpool_layer.cpp
View File

@@ -587,9 +587,17 @@ public:
setParamsFrom(params);
auto_pad = getAutoPadding(params);
kernel_shape = params.getVector<int>("kernel_size");
if (kernel_shape.empty() && params.has("kernel_shape"))
kernel_shape = params.getVector<int>("kernel_shape");
strides = params.getVector<int>("stride");
if (strides.empty() && params.has("strides"))
strides = params.getVector<int>("strides");
dilations = params.getVector<int>("dilation");
if (dilations.empty() && params.has("dilations"))
dilations = params.getVector<int>("dilations");
pads = params.getVector<int>("pad");
if (pads.empty() && params.has("pads"))
pads = params.getVector<int>("pads");
ceil_mode = params.get<bool>("ceil_mode", false);
count_include_pad = params.get<bool>("count_include_pad", false);
}

6
modules/dnn/test/test_onnx_conformance_layer_parser_denylist.inl.hpp
View File

@@ -18,12 +18,6 @@
"test_ai_onnx_ml_tree_ensemble_single_tree",
"test_attention_4d_diff_heads_mask4d_padded_kv", // needs the nonpad_kv_seqlen input (padded-KV variant) which is not supported yet
"test_averagepool_2d_ceil_last_window_starts_on_pad",
"test_averagepool_2d_dilations",
"test_averagepool_3d_dilations_large_count_include_pad_is_0_ceil_mode_is_False",
"test_averagepool_3d_dilations_large_count_include_pad_is_0_ceil_mode_is_True",
"test_averagepool_3d_dilations_large_count_include_pad_is_1_ceil_mode_is_False",
"test_averagepool_3d_dilations_large_count_include_pad_is_1_ceil_mode_is_True",
"test_averagepool_3d_dilations_small",
"test_basic_deform_conv_with_padding",
"test_basic_deform_conv_without_padding",
"test_batchnorm_epsilon_training_mode", // ---- same as above ---