# Copyright (c) 2026 PaddlePaddle Authors. All Rights Reserved. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ 函数式池化操作单元测试 / Functional Pooling Unit Tests 测试目标 / Test Target: paddle.nn.functional.pooling 模块 (覆盖率约82.4%) 覆盖的模块 / Covered Modules: - paddle.nn.functional.avg_pool1d/2d/3d: 平均池化 - paddle.nn.functional.max_pool1d/2d/3d: 最大池化 - paddle.nn.functional.adaptive_avg_pool1d/2d/3d: 自适应平均池化 - paddle.nn.functional.adaptive_max_pool1d/2d/3d: 自适应最大池化 作用 / Purpose: 覆盖各种池化操作的代码路径，测试各种参数组合。 """ import unittest import paddle import paddle.nn.functional as F paddle.disable_static() class TestAvgPool(unittest.TestCase): """测试平均池化 / Test average pooling""" def test_avg_pool1d(self): """测试1D平均池化 / Test 1D average pooling""" x = paddle.randn([4, 3, 16]) y = F.avg_pool1d(x, kernel_size=2, stride=2) self.assertEqual(y.shape, [4, 3, 8]) def test_avg_pool1d_padding(self): """测试带填充的1D平均池化 / Test 1D avg pool with padding""" x = paddle.randn([4, 3, 16]) y = F.avg_pool1d(x, kernel_size=3, stride=1, padding=1) self.assertEqual(y.shape, [4, 3, 16]) def test_avg_pool2d(self): """测试2D平均池化 / Test 2D average pooling""" x = paddle.randn([4, 3, 16, 16]) y = F.avg_pool2d(x, kernel_size=2, stride=2) self.assertEqual(y.shape, [4, 3, 8, 8]) def test_avg_pool2d_same(self): """测试same填充2D平均池化 / Test 2D avg pool with same padding""" x = paddle.randn([4, 3, 16, 16]) y = F.avg_pool2d(x, kernel_size=3, stride=1, padding=1) self.assertEqual(y.shape, [4, 3, 16, 16]) def test_avg_pool2d_with_padding(self): """测试带填充的2D平均池化 / Test 2D avg pool with padding""" x = paddle.randn([4, 3, 16, 16]) y = F.avg_pool2d(x, kernel_size=3, stride=1, padding=1) self.assertEqual(y.shape, [4, 3, 16, 16]) def test_avg_pool3d(self): """测试3D平均池化 / Test 3D average pooling""" x = paddle.randn([2, 3, 8, 8, 8]) y = F.avg_pool3d(x, kernel_size=2, stride=2) self.assertEqual(y.shape, [2, 3, 4, 4, 4]) class TestMaxPool(unittest.TestCase): """测试最大池化 / Test max pooling""" def test_max_pool1d(self): """测试1D最大池化 / Test 1D max pooling""" x = paddle.randn([4, 3, 16]) y = F.max_pool1d(x, kernel_size=2, stride=2) self.assertEqual(y.shape, [4, 3, 8]) def test_max_pool1d_with_indices(self): """测试带索引的1D最大池化 / Test 1D max pool with indices""" x = paddle.randn([4, 3, 16]) y, indices = F.max_pool1d(x, kernel_size=2, stride=2, return_mask=True) self.assertEqual(y.shape, [4, 3, 8]) self.assertEqual(indices.shape, [4, 3, 8]) def test_max_pool2d(self): """测试2D最大池化 / Test 2D max pooling""" x = paddle.randn([4, 3, 16, 16]) y = F.max_pool2d(x, kernel_size=2, stride=2) self.assertEqual(y.shape, [4, 3, 8, 8]) def test_max_pool2d_with_dilation(self): """测试空洞2D最大池化 / Test dilated 2D max pooling""" x = paddle.randn([4, 3, 16, 16]) y = F.max_pool2d(x, kernel_size=2, stride=1, dilation=2) self.assertEqual(y.shape, [4, 3, 14, 14]) def test_max_pool2d_ceil_mode(self): """测试ceil模式2D最大池化 / Test ceil mode 2D max pooling""" x = paddle.randn([4, 3, 15, 15]) y = F.max_pool2d(x, kernel_size=2, stride=2, ceil_mode=True) self.assertEqual(y.shape, [4, 3, 8, 8]) def test_max_pool3d(self): """测试3D最大池化 / Test 3D max pooling""" x = paddle.randn([2, 3, 8, 8, 8]) y = F.max_pool3d(x, kernel_size=2, stride=2) self.assertEqual(y.shape, [2, 3, 4, 4, 4]) class TestAdaptivePool(unittest.TestCase): """测试自适应池化 / Test adaptive pooling""" def test_adaptive_avg_pool1d(self): """测试1D自适应平均池化 / Test 1D adaptive avg pooling""" x = paddle.randn([4, 3, 16]) y = F.adaptive_avg_pool1d(x, output_size=8) self.assertEqual(y.shape, [4, 3, 8]) def test_adaptive_avg_pool2d(self): """测试2D自适应平均池化 / Test 2D adaptive avg pooling""" x = paddle.randn([4, 3, 16, 16]) y = F.adaptive_avg_pool2d(x, output_size=(8, 8)) self.assertEqual(y.shape, [4, 3, 8, 8]) def test_adaptive_avg_pool2d_global(self): """测试全局自适应平均池化 / Test global adaptive avg pooling""" x = paddle.randn([4, 3, 16, 16]) y = F.adaptive_avg_pool2d(x, output_size=(1, 1)) self.assertEqual(y.shape, [4, 3, 1, 1]) def test_adaptive_avg_pool3d(self): """测试3D自适应平均池化 / Test 3D adaptive avg pooling""" x = paddle.randn([2, 3, 8, 8, 8]) y = F.adaptive_avg_pool3d(x, output_size=(4, 4, 4)) self.assertEqual(y.shape, [2, 3, 4, 4, 4]) def test_adaptive_max_pool2d(self): """测试2D自适应最大池化 / Test 2D adaptive max pooling""" x = paddle.randn([4, 3, 16, 16]) y = F.adaptive_max_pool2d(x, output_size=(8, 8)) self.assertEqual(y.shape, [4, 3, 8, 8]) def test_adaptive_max_pool2d_with_indices(self): """测试带索引的2D自适应最大池化 / Test adaptive max pool with indices""" x = paddle.randn([4, 3, 16, 16]) y, indices = F.adaptive_max_pool2d( x, output_size=(8, 8), return_mask=True ) self.assertEqual(y.shape, [4, 3, 8, 8]) self.assertEqual(indices.shape, [4, 3, 8, 8]) class TestSpecialPooling(unittest.TestCase): """测试特殊池化操作 / Test special pooling operations""" def test_max_unpool2d(self): """测试2D最大反池化 / Test 2D max unpooling""" x = paddle.randn([4, 3, 16, 16]) y, indices = F.max_pool2d(x, kernel_size=2, stride=2, return_mask=True) result = F.max_unpool2d(y, indices, kernel_size=2, stride=2) self.assertEqual(result.shape, [4, 3, 16, 16]) def test_lp_pool2d(self): """测试Lp池化 / Test Lp pooling""" x = paddle.randn([4, 3, 16, 16]) y = F.lp_pool2d(x, norm_type=2, kernel_size=2, stride=2) self.assertEqual(y.shape, [4, 3, 8, 8]) if __name__ == '__main__': unittest.main()