bug描述 Describe the Bug

input = paddle.randn([1, 30, 15, 20])
pool = paddle.nn.MaxPool2D(kernel_size=[3, 3], stride=[5, 6], padding=[0, 0], ceil_mode=True)
output = pool(input)
print(output.shape)

输出

[1, 30, 4, 4]

对于同样的代码，pytorch输出的形状不同

input = torch.randn([1, 30, 15, 20])
pool = torch.nn.MaxPool2d(kernel_size=[3, 3], stride=[5, 6], padding=[0, 0], dilation=[1, 1], ceil_mode=True)
output = pool(input)
print(output.shape)

输出

torch.Size([1, 30, 3, 4])

其他补充信息 Additional Supplementary Information

paddle版本2.6.1
torch版本2.1.1