mindspore.ops.fractional_max_pool3d

mindspore.ops.fractional_max_pool3d(input, kernel_size, output_size=None, output_ratio=None, return_indices=False, _random_samples=None)[source]

Applies the 3D FractionalMaxPool operation over input. The output Tensor shape can be determined by either output_size or output_ratio, and the step size is determined by _random_samples. output_size will take effect when output_size and output_ratio are set at the same time. And output_size and output_ratio can not be None at the same time.

Warning

This interface is deprecated and will be removed after version 2.9.0.

Refer to the paper Fractional MaxPooling by Ben Graham for more details.

The input and output data format can be "NCDHW". N is the batch size, C is the number of channels, D is the feature depth, H is the feature height, and W is the feature width.

Parameters

  • input (Tensor) – The input of FractionalMaxPool3d, which is a 4D or 5D tensor. Tensor of data type: float16, float32, double. Supported shape \((N, C, D_{in}, H_{in}, W_{in})\) or \((C, D_{in}, H_{in}, W_{in})\).

  • kernel_size (Union[int, tuple[int]]) – The size of kernel used to take the maximum value, is an int number that represents depth, height and width of the kernel, or a tuple of three int numbers that represent depth, height and width respectively. The value must be a positive integer.

  • output_size (Union[int, tuple[int]], optional) – The shape of the target output_size. It is an int number that represents depth, height and width, or a tuple of three int numbers that represent depth, height and width respectively. The value must be a positive integer. Default: None .

  • output_ratio (Union[float, tuple[float]], optional) – The ratio of target output shape to input shape. Specifying the size of the output tensor by using a ratio of the input size. Data type: float16, float32, double, and the value should be between (0, 1). Default: None .

  • return_indices (bool, optional) – Whether to return the indices of max value. Default: False .

  • _random_samples (Tensor, optional) – The random step of fractional_max_pool3d, which is a 3D tensor. Tensor of data type: float16, float32, double, and the value should be between [0, 1). Supported shape \((N, C, 3)\) or \((1, C, 3)\) . Default: None, the values of _random_samples will be randomly distributed using uniform distribution over an interval [0,1).

Returns

  • y (Tensor) - A tensor, the output of FractionalMaxPool3d. Has the same data type as input. Has the shape \((N, C, D_{out}, H_{out}, W_{out})\) or \((C, D_{out}, H_{out}, W_{out})\) , where \((D_{out}, H_{out}, W_{out})\) = output_size or \((D_{out}, H_{out}, W_{out})\) = output_ratio * \((D_{in}, H_{in}, W_{in})\) .

  • argmax (Tensor) - The indices along with the outputs, which is a Tensor, with the same shape as y and int32 data type. It will output only when return_indices is True.

Raises

  • TypeError – If input is not a 4D or 5D tensor.

  • TypeError – If _random_samples is not a 3D tensor.

  • TypeError – If data type of input is not float16, float32, double, int32, int64.

  • TypeError – If dtype of _random_samples is not float16, float32, double.

  • TypeError – If dtype of argmax is not int32, int64.

  • TypeError – If _random_samples has the different dtype as input.

  • ValueError – If output_size is a tuple and if output_size length is not 3.

  • ValueError – If kernel_size is a tuple and if kernel_size length is not 3.

  • ValueError – If numbers in output_size or kernel_size are not positive.

  • ValueError – If output_size and output_ratio are None at the same time.

  • ValueError – If the first dimension size of input and _random_samples is not equal.

  • ValueError – If the second dimension size of input and _random_samples is not equal.

  • ValueError – If the third dimension size of _random_samples is not 3.

Supported Platforms:

GPU CPU

Examples

>>> import numpy as np
>>> from mindspore import Tensor, ops
>>> from mindspore import dtype as mstype
>>> x = Tensor(np.array([1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16])
...            .reshape([1, 1, 2, 2, 4]), mstype.float32)
>>> _random_samples = Tensor(np.array([0.7, 0.7, 0.7]).reshape([1, 1, 3]), mstype.float32)
>>> output, argmax = ops.fractional_max_pool3d(x, kernel_size=(1, 1, 1), output_size=(1, 1, 3),
...                                            _random_samples=_random_samples, return_indices=True)
>>> print(output)
[[[[[13. 14. 16.]]]]]
>>> print(argmax)
[[[[[12 13 15]]]]]
>>> output, argmax = ops.fractional_max_pool3d(x, kernel_size=(1, 1, 1), output_ratio=(0.5, 0.5, 0.5),
...                                            _random_samples=_random_samples, return_indices=True)
>>> print(output)
[[[[[13. 16.]]]]]
>>> print(argmax)
[[[[[12 15]]]]]