class torch.nn.AvgPool2d(kernel_size, stride=None, padding=0, ceil_mode=False, count_include_pad=True, divisor_override=None)[source]

Applies a 2D average pooling over an input signal composed of several input planes.

In the simplest case, the output value of the layer with input size (N,C,H,W)(N, C, H, W), output (N,C,Hout,Wout)(N, C, H_{out}, W_{out}) and kernel_size (kH,kW)(kH, kW) can be precisely described as:

out(Ni,Cj,h,w)=1kHkWm=0kH1n=0kW1input(Ni,Cj,stride[0]×h+m,stride[1]×w+n)out(N_i, C_j, h, w) = \frac{1}{kH * kW} \sum_{m=0}^{kH-1} \sum_{n=0}^{kW-1} input(N_i, C_j, stride[0] \times h + m, stride[1] \times w + n)

If padding is non-zero, then the input is implicitly zero-padded on both sides for padding number of points.


When ceil_mode=True, sliding windows are allowed to go off-bounds if they start within the left padding or the input. Sliding windows that would start in the right padded region are ignored.

The parameters kernel_size, stride, padding can either be:

  • a single int – in which case the same value is used for the height and width dimension

  • a tuple of two ints – in which case, the first int is used for the height dimension, and the second int for the width dimension

  • kernel_size (Union[int, Tuple[int, int]]) – the size of the window

  • stride (Union[int, Tuple[int, int]]) – the stride of the window. Default value is kernel_size

  • padding (Union[int, Tuple[int, int]]) – implicit zero padding to be added on both sides

  • ceil_mode (bool) – when True, will use ceil instead of floor to compute the output shape

  • count_include_pad (bool) – when True, will include the zero-padding in the averaging calculation

  • divisor_override (Optional[int]) – if specified, it will be used as divisor, otherwise size of the pooling region will be used.

  • Input: (N,C,Hin,Win)(N, C, H_{in}, W_{in}) or (C,Hin,Win)(C, H_{in}, W_{in}).

  • Output: (N,C,Hout,Wout)(N, C, H_{out}, W_{out}) or (C,Hout,Wout)(C, H_{out}, W_{out}), where

    Hout=Hin+2×padding[0]kernel_size[0]stride[0]+1H_{out} = \left\lfloor\frac{H_{in} + 2 \times \text{padding}[0] - \text{kernel\_size}[0]}{\text{stride}[0]} + 1\right\rfloor
    Wout=Win+2×padding[1]kernel_size[1]stride[1]+1W_{out} = \left\lfloor\frac{W_{in} + 2 \times \text{padding}[1] - \text{kernel\_size}[1]}{\text{stride}[1]} + 1\right\rfloor

    Per the note above, if ceil_mode is True and (Hout1)×stride[0]Hin+padding[0](H_{out} - 1)\times \text{stride}[0]\geq H_{in} + \text{padding}[0], we skip the last window as it would start in the bottom padded region, resulting in HoutH_{out} being reduced by one.

    The same applies for WoutW_{out}.


>>> # pool of square window of size=3, stride=2
>>> m = nn.AvgPool2d(3, stride=2)
>>> # pool of non-square window
>>> m = nn.AvgPool2d((3, 2), stride=(2, 1))
>>> input = torch.randn(20, 16, 50, 32)
>>> output = m(input)


Access comprehensive developer documentation for PyTorch

View Docs


Get in-depth tutorials for beginners and advanced developers

View Tutorials


Find development resources and get your questions answered

View Resources