tensorplay.functional.quantized_max_pool2d

tensorplay.functional.quantized_max_pool2d(input, kernel_size, stride=[], padding=0, dilation=1, ceil_mode=False) → Tensor[source]

Applies a 2D max pooling over an input quantized tensor composed of several input planes.

Parameters:

• input (Tensor) – quantized tensor

• kernel_size (list of int) – the size of the sliding window

• stride (list of int, optional) – the stride of the sliding window

• padding (list of int, optional) – padding to be added on both sides, must be >= 0 and <= kernel_size / 2

• dilation (list of int, optional) – The stride between elements within a sliding window, must be > 0. Default 1

• ceil_mode (bool, optional) – If True, will use ceil instead of floor to compute the output shape. Defaults to False.

Returns:

A quantized tensor with max_pool2d applied.

Return type:

Tensor

Example:

>>> qx = tensorplay.quantize_per_tensor(tensorplay.rand(2, 2, 2, 2), 1.5, 3, tensorplay.quint8)
>>> tensorplay.quantized_max_pool2d(qx, [2,2])
tensor([[[[1.5000]],

        [[1.5000]]],


        [[[0.0000]],

        [[0.0000]]]], size=(2, 2, 1, 1), dtype=tensorplay.quint8,
    quantization_scheme=tensorplay.per_tensor_affine, scale=1.5, zero_point=3)