mindspore-Ecosystem
/
mindspore
21
0
Fork 27
Code Issues Pull Requests 6 Projects Releases Wiki Activity

!49227 Add avgpool args 

Merge pull request !49227 from 冯一航/modify_avgpool_args
This commit is contained in:
i-robot 2023-03-07 09:55:21 +00:00 committed by Gitee
parent e2ac12b1ad 717887d2cb
commit 48cb94e5c6
No known key found for this signature in database
GPG Key ID: 173E9B9CA92EEF8F
9 changed files with 415 additions and 136 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

20
docs/api/api_python/nn/mindspore.nn.AvgPool1d.rst
View File

@ -1,7 +1,7 @@
mindspore.nn.AvgPool1d
=======================
.. py:class:: mindspore.nn.AvgPool1d(kernel_size=1, stride=1, pad_mode='valid')
.. py:class:: mindspore.nn.AvgPool1d(kernel_size=1, stride=1, pad_mode="valid", padding=0, ceil_mode=False, count_include_pad=True)
在一个输入Tensor上应用1D平均池化运算，可被视为组成一个1D输入平面。
@ -12,25 +12,29 @@ mindspore.nn.AvgPool1d
\text{output}(N_i, C_j, l) = \frac{1}{k} \sum_{n=0}^{k-1}
\text{input}(N_i, C_j, stride \times l + n)
.. note::
pad_mode仅支持"same"和"valid"。
参数：
- **kernel_size** (int) - 指定池化核尺寸大小，数据类型为整型。默认值：1。
- **stride** (int) - 池化操作的移动步长，数据类型为整型。默认值：1。
- **pad_mode** (str) - 指定池化的填充方式，可选值为"same"或"valid"，不区分大小写。默认值："valid"。
- **pad_mode** (str) - 指定池化的填充方式，可选值为"same"，"valid"或"pad"，不区分大小写。默认值："valid"。
- **same** - 输出的shape与输入整数 `stride` 后的值相同。
- **valid** - 在不填充的前提下返回有效计算所得的输出。不满足计算的多余像素会被丢弃。
- **pad** - 对输入进行填充。在输入的左右两端填充 `padding` 大小的0。如果设置此模式， `padding` 必须大于或等于0。
- **padding** (Union(int, tuple[int], list[int])) - 池化填充值。默认值：0。 `padding` 只能是一个整数或者包含一个整数的tuple/list，设定后，则会在输入的左边和右边填充 `padding` 次或者 `padding[0]` 次。
- **ceil_mode** (bool) - 若为True，使用ceil来计算输出shape。若为False，使用floor来计算输出shape。默认值：False。
- **count_include_pad** (bool) - 如果为True，平均计算将包括零填充。默认值：True。
输入：
- **x** (Tensor) - shape为 :math:`(N, C_{in}, L_{in})` 的Tensor。
- **x** (Tensor) - shape为 :math:`(N, C_{in}, L_{in})` 或 :math:`(C_{in}, L_{in})` 的Tensor。
输出：
shape为 :math:`(N, C_{out}, L_{out})` 的Tensor。
shape为 :math:`(N, C_{out}, L_{out})` 或 :math:`(C_{out}, L_{out})` 的Tensor。
异常：
- **TypeError** - `kernel_size` 或 `stride` 不是int。
- **ValueError** - `pad_mode` 既不是"valid"，也不是"same"，不区分大小写。
- **ValueError** - `pad_mode` 既不是"valid"，也不是"same" 或者 "pad"，不区分大小写。
- **ValueError** - `kernel_size` 或 `stride` 小于1。
- **ValueError** - `x` 的shape长度不等于3。
- **ValueError** - `padding` 为tuple/list时长度不为1。
- **ValueError** - `x` 的shape长度不等于2或3。

22
docs/api/api_python/nn/mindspore.nn.AvgPool2d.rst
View File

@ -1,7 +1,7 @@
mindspore.nn.AvgPool2d
=======================
.. py:class:: mindspore.nn.AvgPool2d(kernel_size=1, stride=1, pad_mode='valid', data_format='NCHW')
.. py:class:: mindspore.nn.AvgPool2d(kernel_size=1, stride=1, pad_mode='valid', padding=0, ceil_mode=False, count_include_pad=True, divisor_override=None, data_format='NCHW')
在输入Tensor上应用2D平均池化运算，可视为二维输入平面的组合。
@ -11,28 +11,32 @@ mindspore.nn.AvgPool2d
\text{output}(N_i, C_j, h, w) = \frac{1}{kH * kW} \sum_{m=0}^{kH-1} \sum_{n=0}^{kW-1}
\text{input}(N_i, C_j, stride[0] \times h + m, stride[1] \times w + n)
.. note::
pad_mode仅支持"same"和"valid"。
参数：
- **kernel_size** (Union[int, tuple[int]]) - 指定池化核尺寸大小。如果为整数，则代表池化核的高和宽。如果为tuple，其值必须包含两个整数值分别表示池化核的高和宽。默认值：1。
- **stride** (Union[int, tuple[int]]) - 池化操作的移动步长，如果为整数，则代表池化核的高和宽方向的移动步长。如果为tuple，其值必须包含两个整数值分别表示池化核的高和宽的移动步长。默认值：1。
- **pad_mode** (str) - 指定池化填充模式，可选值为"same"或"valid"，不区分大小写。默认值："valid"。
- **pad_mode** (str) - 指定池化的填充方式，可选值为"same"，"valid"或"pad"，不区分大小写。默认值："valid"。
- same: 输出的宽度与输入整数 `stride` 后的值相同。
- valid: 在不填充的前提下返回有效计算所得的输出。不满足计算的多余像素会被丢弃。
- pad: 对输入进行填充。在输入的上下左右分别填充 `padding` 大小的0。如果设置此模式， `padding` 必须大于或等于0。
- **padding** (Union(int, tuple[int], list[int])) - 池化填充值。默认值：0。 `padding` 只能是一个整数或者包含一个或两个整数的元组，若 `padding` 为一个整数或者包含一个整数的tuple/list，则会分别在输入的上下左右四个方向进行 `padding` 次的填充，若 `padding` 为一个包含两个整数的tuple/list，则会在输入的上下进行 `padding[0]` 次的填充，在输入的左右进行 `padding[1]` 次的填充。
- **ceil_mode** (bool) - 若为True，使用ceil来计算输出shape。若为False，使用floor来计算输出shape。默认值：False。
- **count_include_pad** (bool) - 平均计算是否包括零填充。默认值：True。
- **divisor_override** (int) - 如果被指定为非0参数，该参数将会在平均计算中被用作除数，否则将会使用 `kernel_size` 作为除数，默认值：None。
- **data_format** (str) - 输入数据格式可为'NHWC'或'NCHW'。默认值：'NCHW'。
输入：
- **x** (Tensor) - 输入数据的shape为 :math:`(N,C_{in},H_{in},W_{in})` 的Tensor。
- **x** (Tensor) - 输入数据的shape为 :math:`(N,C_{in},H_{in},W_{in})` 或 :math:`C_{in},H_{in},W_{in})` 的Tensor。
输出：
输出数据的shape为 :math:`(N,C_{out},H_{out},W_{out})` 的Tensor。
输出数据的shape为 :math:`(N,C_{out},H_{out},W_{out})` 或 :math:`(C_{out},H_{out},W_{out})` 的Tensor。
异常：
- **TypeError** - `kernel_size` 或 `strides` 既不是整数也不是元组。
- **ValueError** - `pad_mode` 既不是'valid'，也不是'same'，不区分大小写。
- **ValueError** - `pad_mode` 既不是"valid"，也不是"same" 或者 "pad"，不区分大小写。
- **ValueError** - `data_format` 既不是'NCHW'，也不是'NHWC'。
- **ValueError** - `data_format` 为 'NHWC' 时，使用了 `padding` 或者 `ceil_mode` 或者 `count_include_pad` 或者 `divisor_override` 或者 `pad_mode` 为 `pad`。
- **ValueError** - `kernel_size` 或 `stride` 小于1。
- **ValueError** - `x` 的shape长度不等于4。
- **ValueError** - `padding` 为tuple/list时长度不为1或2。
- **ValueError** - `x` 的shape长度不等于3或4。

21
docs/api/api_python/nn/mindspore.nn.AvgPool3d.rst
View File

@ -1,7 +1,7 @@
mindspore.nn.AvgPool3d
=======================
.. py:class:: mindspore.nn.AvgPool3d(kernel_size, stride=None, padding=0, ceil_mode=False, count_include_pad=True, divisor_override=None)
.. py:class:: mindspore.nn.AvgPool3d(kernel_size=1, stride=1, pad_mode='valid', padding=0, ceil_mode=False, count_include_pad=True, divisor_override=None)
在一个输入Tensor上应用3D平均池化运算，输入Tensor可看成是由一系列3D平面组成的。
@ -16,9 +16,15 @@ mindspore.nn.AvgPool3d
\text{input}(N_i, C_j, s_0 \times d + l, s_1 \times h + m, s_2 \times w + n)
参数：
- **kernel_size** (Union[int, tuple[int]]) - 指定池化核尺寸大小。如果为int，则同时代表池化核的深度，高度和宽度。如果为tuple，其值必须包含三个int，分别表示池化核的深度，高度和宽度。取值必须为正整数。
- **stride** (Union[int, tuple[int]]) - 池化操作的移动步长。如果为int，则同时代表池化核的深度，高度和宽度方向上的移动步长。如果为tuple，其值必须包含三个整数值，分别表示池化核的深度，高度和宽度方向上的移动步长。取值必须为正整数。如果值为None，则使用默认值 `kernel_size`。
- **padding** (Union(int, tuple[int])) - 需要填充的pad值。取值不能为负数。如果 `padding` 为整数，则分别在头，尾，上，下，左，右都填充padding，如果 `padding` 是一个六个整数的元组，则分别在头，尾，上，下，左，右填充padding[0]，padding[1]，padding[2]，padding[3]，padding[4]，padding[5]。默认值：0。
- **kernel_size** (Union[int, tuple[int]]) - 指定池化核尺寸大小。如果为int，则同时代表池化核的深度，高度和宽度。如果为tuple，其值必须包含三个int，分别表示池化核的深度，高度和宽度。取值必须为正整数。默认值：1。
- **stride** (Union[int, tuple[int]]) - 池化操作的移动步长。如果为int，则同时代表池化核的深度，高度和宽度方向上的移动步长。如果为tuple，其值必须包含三个整数值，分别表示池化核的深度，高度和宽度方向上的移动步长。取值必须为正整数。默认值：1。
- **pad_mode** (str) - 指定池化的填充方式，可选值为"same"，"valid"或"pad"，不区分大小写。默认值："valid"。
- same: 输出的宽度与输入整数 `stride` 后的值相同。
- valid: 在不填充的前提下返回有效计算所得的输出。不满足计算的多余像素会被丢弃。
- pad: 对输入进行填充。在输入的前后上下左右分别填充 `padding` 大小的0。如果设置此模式， `padding` 必须大于或等于0。
- **padding** (Union(int, tuple[int], list[int])) - 池化填充值。默认值：0。 `padding` 只能是一个整数或者包含一个或三个整数的tuple/list，若 `padding` 为一个整数或包含一个整数的tuple/list，则会分别在输入的前后上下左右六个方向进行 `padding` 次的填充，若 `padding` 为一个包含三个整数的tuple/list，则会在输入的前后进行 `padding[0]` 次的填充，上下进行 `padding[1]` 次的填充，在输入的左右进行 `padding[2]` 次的填充。
- **ceil_mode** (bool) - 若为True，使用ceil来计算输出shape。若为False，使用floor来计算输出shape。默认值：False。
- **count_include_pad** (bool) - 平均计算是否包括零填充。默认值：True。
- **divisor_override** (int) - 如果被指定为非0参数，该参数将会在平均计算中被用作除数，否则将会使用 `kernel_size` 作为除数，默认值：None。
@ -27,15 +33,14 @@ mindspore.nn.AvgPool3d
- **x** (Tensor) - shape为 :math:`(N, C, D_{in}, H_{in}, W_{in})` 或者 :math:`(C, D_{in}, H_{in}, W_{in})` 的Tensor。数据类型必须为float16或者float32。
输出：
shape为 :math:`(N, C, D_{out}, H_{out}, W_{out})` 或者 :math:`(C, D_{in}, H_{in}, W_{in})` 的Tensor。数据类型与 `x` 一致。
shape为 :math:`(N, C, D_{out}, H_{out}, W_{out})` 或者 :math:`(C, D_{out}, H_{out}, W_{out})` 的Tensor。数据类型与 `x` 一致。
异常：
- **TypeError** - `kernel_size` ，`stride` 或 `padding` 既不是整数也不是元组。
- **TypeError** - `ceil_mode` 或 `count_include_pad` 不是布尔类型。
- **TypeError** - `data_format` 不是字符串。
- **TypeError** - `divisor_override` 不是整数。
- **ValueError** - `kernel_size` 或者 `stride` 中的数字不是正数。
- **ValueError** - `kernel_size` 或 `stride` 是一个长度不为3的tuple。
- **ValueError** - `padding` 是一个长度不为6的tuple。
- **ValueError** - `padding` 为一个tuple/list时，长度不为1或者3。
- **ValueError** - `padding` 的元素小于0。
- **ValueError** - `x` 的shape长度不等于5。
- **ValueError** - `x` 的shape长度不等于4或5。

327
mindspore/python/mindspore/nn/layer/pooling.py
View File

@ -41,7 +41,7 @@ class _PoolNd(Cell):
"""Initialize _PoolNd."""
super(_PoolNd, self).__init__()
validator.check_value_type('pad_mode', pad_mode, [str], self.cls_name)
self.pad_mode = validator.check_string(pad_mode.upper(), ['VALID', 'SAME'], 'pad_mode', self.cls_name)
self.pad_mode = validator.check_string(pad_mode.upper(), ['VALID', 'SAME', 'PAD'], 'pad_mode', self.cls_name)
self.format = validator.check_string(data_format, ['NCHW', 'NHWC'], 'format', self.cls_name)
if context.get_context("device_target") != "GPU" and self.format == "NHWC":
raise ValueError(f"For '{self.cls_name}, the 'NHWC' format only support in GPU target, but got device "
@ -50,17 +50,17 @@ class _PoolNd(Cell):
def _check_int_or_tuple(arg_name, arg_value):
validator.check_value_type(arg_name, arg_value, [int, tuple], self.cls_name)
error_msg = f"For '{self.cls_name}', the '{arg_name}' must be an positive int number or " \
f"a tuple of two positive int numbers, but got {arg_value}"
f"a tuple, but got {arg_value}"
if isinstance(arg_value, int):
if arg_value <= 0:
raise ValueError(error_msg)
elif len(arg_value) == 2:
else:
for item in arg_value:
if isinstance(item, int) and item > 0:
continue
raise ValueError(error_msg)
else:
raise ValueError(error_msg)
if len(arg_value) == 1:
return arg_value[0]
return arg_value
self.kernel_size = _check_int_or_tuple('kernel_size', kernel_size)
@ -488,7 +488,37 @@ class MaxPool1d(_PoolNd):
return output
class AvgPool3d(Cell):
def _cal_padding(padding, cls_name, nd):
"""Calculate padding before call primitive"""
validator.check_value_type('padding', padding, (int, tuple, list), cls_name)
if isinstance(padding, int):
padding = (0, 0) * (3 - nd) + (padding,) * nd * 2
elif isinstance(padding, (tuple, list)):
validator.check_non_negative_int_sequence(padding, "padding", cls_name)
if len(padding) == nd:
padding_start = (0, 0) * (3 - nd)
padding_end = tuple(padding[i // 2] for i in range(nd * 2))
padding = padding_start + padding_end
elif len(padding) == 1:
padding = (0, 0) * (3 - nd) + padding * nd * 2
else:
if nd == 1:
raise ValueError(f"For {cls_name}, the padding must be a int or tuple contains one int, "
f"but got tuple with length:{len(padding)}.")
raise ValueError(f"For {cls_name}, the padding must be a int or tuple contains 1 or {nd} int, "
f"but got tuple with length:{len(padding)}.")
return padding
def _check_tuple_length(arg_name, prim_name, length, cls_name):
"""check the tuple length"""
if len(arg_name) != length:
raise ValueError(f"For {cls_name}, the length of {prim_name} must be equal to {length}, "
f"but got {len(arg_name)}.")
return arg_name
class AvgPool3d(_PoolNd):
r"""
Applies a 3D average pooling over an input Tensor which can be regarded as a composition of 3D input planes.
Typically the input is of shape :math:`(N, C, D_{in}, H_{in}, W_{in})`, and AvgPool3D outputs
@ -507,16 +537,28 @@ class AvgPool3d(Cell):
kernel_size (Union[int, tuple[int]]): The size of kernel used to take the average value,
can be 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.
The value must be a positive integer. Default: 1.
stride (Union[int, tuple[int]]): The distance of kernel moving, can be an int number that represents
the depth, height and width of movement, or a tuple of three int numbers that
represent depth, height and width of movement respectively. The value must be a positive integer.
If the value is None, the default value `kernel_size` is used.
padding (Union(int, tuple[int])): The padding value to be filled. Default: 0. The value cannot be negative.
If `padding` is an integer, the paddings of head, tail, top, bottom, left and right are the same,
equal to padding.
If `padding` is a tuple of six integers, the padding of head, tail, top, bottom, left and right
equal to padding[0], padding[1], padding[2], padding[3], padding[4] and padding[5] correspondingly.
If the value is None, the default value `kernel_size` is used. Default: 1.
pad_mode (str): Specifies the padding method of pooling, optional values are "same", "valid" or "pad",
case insensitive. Default: "valid".
- same: The output width is the same as the integer value after the input is multiplied by the stride.
- valid: Returns the output obtained by effective calculation without padding.
The excess pixels that do not meet the calculation will be discarded.
- pad: Pads the input. Fill the front, back, top, and bottom of the input with 0s of size `padding`.
If this mode is set, `padding` must be greater than or equal to 0.
padding (Union(int, tuple[int], list[int])): Pooling padding value. Default: 0.
`padding` can only be an integer or a tuple/list containing one or three integers.
If `padding` is an integer or a tuple/list containing one integer, it will be padded in six directions of
front, back, top, bottom, left and right of the input. If `padding` is a tuple/list containing three
integers, it will be padded in front and back of the input `padding[0]` times, up and down `padding[1]`
times, and left and right of the input `padding[2]` times.
ceil_mode (bool): If True, use ceil to compute the output shape instead of floor. Default: False.
count_include_pad (bool): If True, averaging calculation will include the zero-padding. Default: True.
divisor_override (int): If specified, it will be used as divisor in the averaging calculation,
@ -533,14 +575,13 @@ class AvgPool3d(Cell):
Raises:
TypeError: If `kernel_size`, `stride` or `padding` is neither an int nor a tuple.
TypeError: If `ceil_mode` or `count_include_pad` is not a bool.
TypeError: If `data_format` is not a string.
TypeError: If `ceil_mode` or `count_include_pad` is not a bool.=
TypeError: If `divisor_override` is not an int.
ValueError: If numbers in `kernel_size` or `stride` are not positive.
ValueError: If `kernel_size` or `stride` is a tuple whose length is not equal to 3.
ValueError: If `padding` is a tuple whose length is not equal to 6.
ValueError: If `padding` is a tuple/list whose length is neither 1 nor 3.
ValueError: If element of `padding` is less than 0.
ValueError: If length of shape of `x` is not equal to 5.
ValueError: If length of shape of `x` is neither 4 nor 5.
Supported Platforms:
``Ascend`` ``GPU`` ``CPU``
@ -548,34 +589,36 @@ class AvgPool3d(Cell):
Examples:
>>> import mindspore as ms
>>> import mindspore.nn as nn
>>> import numpy as np
>>> import mindspore.ops as ops
>>> pool = nn.AvgPool3d(kernel_size=3, stride=1)
>>> x = ms.Tensor(np.random.randint(0, 10, [1, 2, 4, 4, 5]), ms.float32)
>>> x = ops.randn(1, 2, 4, 4, 5).astype(ms.float32)
>>> output = pool(x)
>>> print(output.shape)
(1, 2, 2, 2, 3)
>>> x1 = ops.randn(6, 5, 7, 7, 5).astype(ms.float32)
>>> pool2 = nn.AvgPool3d(4, stride=2, pad_mode='pad', padding=(2, 2, 1), divisor_override=10)
>>> output2 = pool2(x1)
>>> print(output2.shape)
(6, 5, 4, 4, 2)
"""
def __init__(self, kernel_size, stride=None, padding=0, ceil_mode=False, count_include_pad=True,
def __init__(self, kernel_size=1, stride=1, pad_mode="valid", padding=0, ceil_mode=False, count_include_pad=True,
divisor_override=None):
"""Initialize AvgPool3d."""
super(AvgPool3d, self).__init__()
stride = stride if (stride is not None) else kernel_size
if not divisor_override:
divisor_override = 0
self.avg_pool = P.AvgPool3D(kernel_size, stride, "pad", padding, ceil_mode, count_include_pad,
super(AvgPool3d, self).__init__(kernel_size, stride, pad_mode)
padding = _cal_padding(padding, self.cls_name, 3)
divisor_override = 0 if divisor_override is None else divisor_override
self.avg_pool = P.AvgPool3D(self.kernel_size, self.stride, pad_mode, padding, ceil_mode, count_include_pad,
divisor_override)
self.squeeze = P.Squeeze(0)
self.expand_dims = P.ExpandDims()
def construct(self, x):
_is_squeeze = False
expand_batch = False
if len(x.shape) == 4:
x = self.expand_dims(x, 0)
_is_squeeze = True
x = x.unsqueeze(0)
expand_batch = True
out = self.avg_pool(x)
if _is_squeeze:
out = self.squeeze(out)
if expand_batch:
out = out.squeeze(0)
return out
@ -591,9 +634,6 @@ class AvgPool2d(_PoolNd):
\text{output}(N_i, C_j, h, w) = \frac{1}{h_{ker} * w_{ker}} \sum_{m=0}^{h_{ker}-1} \sum_{n=0}^{w_{ker}-1}
\text{input}(N_i, C_j, s_0 \times h + m, s_1 \times w + n)
Note:
pad_mode for training only supports "same" and "valid".
Args:
kernel_size (Union[int, tuple[int]]): The size of kernel used to take the average value.
The data type of kernel_size must be int and the value represents the height and width,
@ -602,58 +642,121 @@ class AvgPool2d(_PoolNd):
stride (Union[int, tuple[int]]): The distance of kernel moving, an int number that represents
the height and width of movement are both strides, or a tuple of two int numbers that
represent height and width of movement respectively. Default: 1.
pad_mode (str): The optional value for pad mode, is "same" or "valid", not case sensitive.
Default: "valid".
pad_mode (str) - Specifies the padding method of pooling, optional values are "same", "valid" or "pad",
case insensitive. Default: "valid".
- same: Adopts the way of completion. The height and width of the output will be the same as
the input. The total number of padding will be calculated in horizontal and vertical
directions and evenly distributed to top and bottom, left and right if possible.
Otherwise, the last extra padding will be done from the bottom and the right side.
- same: The output width is the same as the integer value after the input is multiplied by the stride.
- valid: Adopts the way of discarding. The possible largest height and width of output
will be returned without padding. Extra pixels will be discarded.
- valid: Returns the output obtained by effective calculation without padding.
The excess pixels that do not meet the calculation will be discarded.
- pad: pads the input. Pads the top, bottom, left, and right sides of the input with `padding` number of
zeros. If this mode is set, `padding` must be greater than or equal to 0.
padding (Union[int, tuple[int], list[int]]): Specifies the padding value of the pooling operation. Default: 0.
`padding` can only be an integer or a tuple/list containing one or two integers. If `padding` is an integer
or a tuple/list containing one integer, it will be padded `padding` times in the four directions of the
input. If `padding` is a tuple/list containing two integers, it will be padded `padding[0]` times in the
up-down direction of the input and `padding[1]` times in the left-right direction of the input.
ceil_mode (bool): If True, use ceil to compute the output shape instead of floor. Default: False.
count_include_pad (bool): If True, averaging calculation will include the zero-padding. Default: True.
divisor_override (int): If specified, it will be used as divisor in the averaging calculation,
otherwise kernel_size will be used. Default: None.
data_format (str): The optional value for data format, is 'NHWC' or 'NCHW'.
Default: 'NCHW'.
Inputs:
- **x** (Tensor) - Tensor of shape :math:`(N, C_{in}, H_{in}, W_{in})`.
- **x** (Tensor) - Tensor of shape :math:`(N, C_{in}, H_{in}, W_{in})` or :math:`C_{in},H_{in},W_{in})`.
Outputs:
Tensor of shape :math:`(N, C_{out}, H_{out}, W_{out})`.
Tensor of shape :math:`(N, C_{out}, H_{out}, W_{out})` or :math:`(C_{out},H_{out},W_{out})`.
Raises:
TypeError: If `kernel_size` or `strides` is neither int nor tuple.
ValueError: If `pad_mode` is neither 'valid' nor 'same' with not case sensitive.
ValueError: If `pad_mode` is not 'valid' ,'same' or 'pad' with not case sensitive.
ValueError: If `data_format` is neither 'NCHW' nor 'NHWC'.
ValueError: If `padding`, `ceil_mode`, `count_include_pad`, or `divisor_override` is used
or `pad_mode` is `pad` when `data_format` is 'NHWC'.
ValueError: If `kernel_size` or `strides` is less than 1.
ValueError: If length of shape of `x` is not equal to 4.
ValueError: If length of `padding` tuple/list is not 1 or 2.
ValueError: If length of shape of `x` is not equal to 3 or 4.
Supported Platforms:
``Ascend`` ``GPU`` ``CPU``
Examples:
>>> import mindspore as ms
>>> import mindspore.nn as nn
>>> import mindspore.ops as ops
>>> import numpy as np
>>> pool = nn.AvgPool2d(kernel_size=3, stride=1)
>>> x = Tensor(np.random.randint(0, 10, [1, 2, 4, 4]), mindspore.float32)
>>> x = ms.Tensor(np.random.randint(0, 10, [1, 2, 4, 4]), ms.float32)
>>> output = pool(x)
>>> print(output.shape)
(1, 2, 2, 2)
>>> x = ops.randn(6, 6, 8, 8)
>>> pool2 = nn.AvgPool2d(4, stride=1, pad_mode='pad', padding=2, divisor_override=5)
>>> output2 = pool2(x)
>>> print(output2.shape)
(6, 6, 9, 9)
"""
def __init__(self,
kernel_size=1,
stride=1,
pad_mode="valid",
padding=0,
ceil_mode=False,
count_include_pad=True,
divisor_override=None,
data_format="NCHW"):
"""Initialize AvgPool2d."""
super(AvgPool2d, self).__init__(kernel_size, stride, pad_mode, data_format)
self.avg_pool = P.AvgPool(kernel_size=self.kernel_size,
strides=self.stride,
pad_mode=self.pad_mode,
data_format=self.format)
if pad_mode == 'pad' or padding != 0 or ceil_mode or not count_include_pad or divisor_override is not None:
if self.format == "NHWC":
raise ValueError(f"For '{self.cls_name}, the 'NHWC' format are not support when 'padding' is not 0 "
f"or 'ceil_mode' is not False or 'count_include_pad' is not True "
f"or divisor_override is not None, but got padding:{padding}, ceil_mode:{ceil_mode}, "
f"count_include_pad:{count_include_pad}, divisor_override:{divisor_override}.")
self.is_expand = True
divisor_override = 0 if divisor_override is None else divisor_override
padding = _cal_padding(padding, self.cls_name, 2)
if isinstance(self.kernel_size, tuple):
_check_tuple_length(self.kernel_size, 'kernel_size', 2, self.cls_name)
kernel_size = (1,) + self.kernel_size
elif isinstance(self.kernel_size, int):
kernel_size = (1, self.kernel_size, self.kernel_size)
if isinstance(self.stride, tuple):
_check_tuple_length(self.stride, 'stride', 2, self.cls_name)
stride = (1,) + self.stride
elif isinstance(self.stride, int):
stride = (1, self.stride, self.stride)
self.avg_pool = P.AvgPool3D(kernel_size=kernel_size, strides=stride, pad_mode=pad_mode, pad=padding,
ceil_mode=ceil_mode,
count_include_pad=count_include_pad, divisor_override=divisor_override)
else:
self.is_expand = False
self.avg_pool = P.AvgPool(kernel_size=self.kernel_size,
strides=self.stride,
pad_mode=self.pad_mode,
data_format=self.format)
def construct(self, x):
return self.avg_pool(x)
expand_batch = False
if x.ndim == 3:
x = x.unsqueeze(0)
expand_batch = True
if self.is_expand:
x = x.unsqueeze(2)
out = self.avg_pool(x)
res = out.squeeze(2)
else:
res = self.avg_pool(x)
if expand_batch:
res = res.squeeze(0)
return res
class AvgPool1d(_PoolNd):
@ -675,76 +778,111 @@ class AvgPool1d(_PoolNd):
kernel_size (int): The size of kernel window used to take the average value, Default: 1.
stride (int): The distance of kernel moving, an int number that represents
the width of movement is strides, Default: 1.
pad_mode (str): The optional value for pad mode, is "same" or "valid", not case sensitive.
Default: "valid".
pad_mode (str) - Specifies the padding method of pooling, optional values are "same", "valid" or "pad",
case insensitive. Default: "valid".
- same: Adopts the way of completion. The height and width of the output will be the same as
the input. The total number of padding will be calculated in horizontal and vertical
directions and evenly distributed to top and bottom, left and right if possible.
Otherwise, the last extra padding will be done from the bottom and the right side.
- same: The output width is the same as the integer value after the input is multiplied by the stride.
- valid: Adopts the way of discarding. The possible largest height and width of output
will be returned without padding. Extra pixels will be discarded.
- valid: Returns the output obtained by effective calculation without padding.
The excess pixels that do not meet the calculation will be discarded.
- pad: Performs padding on the input. Adds padding size of zeros to both ends of the input.
If this mode is set, padding must be greater than or equal to 0.
padding (Union(int, tuple[int], list[int])): Padding value for the pooling. Default value is 0.
padding can only be an integer or a tuple/list containing a single integer, in which case padding times or
padding[0] times are padded on both sides of the input.
ceil_mode (bool): If True, use ceil to compute the output shape instead of floor. Default: False.
count_include_pad (bool): If True, averaging calculation will include the zero-padding. Default: True.
Inputs:
- **x** (Tensor) - Tensor of shape :math:`(N, C_{in}, L_{in})`.
- **x** (Tensor) - Tensor of shape :math:`(N, C_{in}, L_{in})` or :math:`(C_{in}, L_{in})`.
Outputs:
Tensor of shape :math:`(N, C_{out}, L_{out})`.
Tensor of shape :math:`(N, C_{out}, L_{out})` or :math:`(C_{out}, L_{out})`.
Raises:
TypeError: If `kernel_size` or `stride` is not an int.
ValueError: If `pad_mode` is neither 'same' nor 'valid' with not case sensitive.
ValueError: If `pad_mode` is not 'valid' ,'same' or 'pad' with not case sensitive.
ValueError: If `kernel_size` or `strides` is less than 1.
ValueError: If length of shape of `x` is not equal to 3.
ValueError: If length of `padding` tuple/list is not 1.
ValueError: If length of shape of `x` is not equal to 2 or 3.
Supported Platforms:
``Ascend`` ``GPU`` ``CPU``
Examples:
>>> import mindspore as ms
>>> import mindspore.nn as nn
>>> import mindspore.ops as ops
>>> import numpy as np
>>> pool = nn.AvgPool1d(kernel_size=6, stride=1)
>>> x = Tensor(np.random.randint(0, 10, [1, 3, 6]), mindspore.float32)
>>> x = ms.Tensor(np.random.randint(0, 10, [1, 3, 6]), ms.float32)
>>> output = pool(x)
>>> result = output.shape
>>> print(result)
(1, 3, 1)
>>> pool2 = nn.AvgPool1d(4, stride=1, ceil_mode=True, pad_mode='pad', padding=2)
>>> x1 = ops.randn(6, 6, 8)
>>> output = pool2(x1)
>>> print(output.shape)
(6, 6, 9)
"""
def __init__(self,
kernel_size=1,
stride=1,
pad_mode="valid"):
pad_mode="valid",
padding=0,
ceil_mode=False,
count_include_pad=True):
"""Initialize AvgPool1d."""
validator.check_value_type('kernel_size', kernel_size, [int], self.cls_name)
validator.check_value_type('stride', stride, [int], self.cls_name)
validator.check_value_type('pad_mode', pad_mode, [str], self.cls_name)
self.pad_mode = validator.check_string(pad_mode.upper(), ['VALID', 'SAME'], 'pad_mode', self.cls_name)
validator.check_int(kernel_size, 1, Rel.GE, "kernel_size", self.cls_name)
validator.check_int(stride, 1, Rel.GE, "stride", self.cls_name)
super(AvgPool1d, self).__init__(kernel_size, stride, pad_mode)
self.kernel_size = (1, kernel_size)
self.stride = (1, stride)
self.avg_pool = P.AvgPool(kernel_size=self.kernel_size,
strides=self.stride,
pad_mode=self.pad_mode)
self.shape = F.shape
self.reduce_mean = P.ReduceMean(keep_dims=True)
self.slice = P.Slice()
self.expand = P.ExpandDims()
self.squeeze = P.Squeeze(2)
validator.check_int(self.kernel_size, 1, Rel.GE, "kernel_size", self.cls_name)
validator.check_int(self.stride, 1, Rel.GE, "stride", self.cls_name)
if pad_mode == 'pad' or padding != 0 or ceil_mode or not count_include_pad:
padding = _cal_padding(padding, self.cls_name, 1)
self.is_expand_3d = True
kernel_size = (1, 1, self.kernel_size)
stride = (1, 1, self.stride)
self.avg_pool = P.AvgPool3D(kernel_size=kernel_size, strides=stride, pad_mode=pad_mode, pad=padding,
ceil_mode=ceil_mode,
count_include_pad=count_include_pad)
else:
self.is_expand_3d = False
self.kernel_size = (1, self.kernel_size)
self.stride = (1, self.stride)
self.avg_pool = P.AvgPool(kernel_size=self.kernel_size,
strides=self.stride,
pad_mode=self.pad_mode)
self.shape = F.shape
self.reduce_mean = P.ReduceMean(keep_dims=True)
self.slice = P.Slice()
self.expand = P.ExpandDims()
self.squeeze = P.Squeeze(2)
def construct(self, x):
batch, channel, width = self.shape(x)
if width == self.kernel_size[1]:
x = self.reduce_mean(x, 2)
elif width - self.kernel_size[1] < self.stride[1]:
x = self.slice(x, (0, 0, 0), (batch, channel, self.kernel_size[1]))
x = self.reduce_mean(x, 2)
else:
x = self.expand(x, 2)
expand_batch = False
if x.ndim == 2:
x = x.unsqueeze(0)
expand_batch = True
if self.is_expand_3d:
x = x.unsqueeze(2).unsqueeze(3)
x = self.avg_pool(x)
x = self.squeeze(x)
x = x.squeeze(3).squeeze(2)
else:
batch, channel, width = self.shape(x)
if width == self.kernel_size[1]:
x = self.reduce_mean(x, 2)
elif width - self.kernel_size[1] < self.stride[1]:
x = self.slice(x, (0, 0, 0), (batch, channel, self.kernel_size[1]))
x = self.reduce_mean(x, 2)
else:
x = self.expand(x, 2)
x = self.avg_pool(x)
x = self.squeeze(x)
if expand_batch:
x = x.squeeze(0)
return x
@ -1443,6 +1581,7 @@ class MaxUnpool1d(Cell):
>>> print(output.asnumpy())
[[0. 2. 0. 4. 0. 6. 0. 8.]]
"""
def __init__(self, kernel_size, stride=None, padding=0):
"""Initialize MaxUnpool1d."""
super(MaxUnpool1d, self).__init__()
@ -1530,6 +1669,7 @@ class MaxUnpool2d(Cell):
[[[[0. 1.]
[8. 9.]]]]
"""
def __init__(self, kernel_size, stride=None, padding=0):
"""Initialize MaxUnpool2d."""
super(MaxUnpool2d, self).__init__()
@ -1620,6 +1760,7 @@ class MaxUnpool3d(Cell):
[[[[[0. 1.]
[8. 9.]]]]]
"""
def __init__(self, kernel_size, stride=None, padding=0):
super(MaxUnpool3d, self).__init__()
if stride is None:

2
mindspore/python/mindspore/ops/function/nn_func.py
View File

@ -5563,7 +5563,7 @@ def lp_pool1d(x, norm_type, kernel_size, stride=None, ceil_mode=False):
def lp_pool2d(x, norm_type, kernel_size, stride=None, ceil_mode=False):
r"""
Applying 2D LPPooling operation on an input Tensor can be regarded as forming a 1D input plane.
Applying 2D LPPooling operation on an input Tensor can be regarded as forming a 2D input plane.
Typically the input is of shape :math:`(N, C, H_{in}, W_{in})`, the output is of shape
:math:`(N, C, H_{in}, W_{in})`, with the same shape as input, the operation is as follows.

10
mindspore/python/mindspore/ops/operations/nn_ops.py
View File

@ -7496,10 +7496,10 @@ class AvgPool3D(Primitive):
- pad: Implicit paddings on both sides of the input in depth, height, width. The number of `pad` will
be padded to the input Tensor borders. `pad` must be greater than or equal to 0.
pad (Union(int, tuple[int])): The pad value to be filled. Default: 0. If `pad` is an integer, the paddings of
head, tail, top, bottom, left and right are the same, equal to pad. If `pad` is a tuple of six
integers, the padding of head, tail, top, bottom, left and right equal to pad[0], pad[1], pad[2],
pad[3], pad[4] and pad[5] correspondingly.
pad (Union(int, tuple[int], list[int])): The pad value to be filled. Default: 0. If `pad` is an integer,
the paddings of head, tail, top, bottom, left and right are the same, equal to pad.
If `pad` is a tuple of six integers, the padding of head, tail, top, bottom, left and right equal to
pad[0], pad[1], pad[2], pad[3], pad[4] and pad[5] correspondingly.
ceil_mode (bool): If True, ceil instead of floor to compute the output shape. Default: False.
count_include_pad (bool): If True, averaging calculation will include the zero-padding. Default: True.
divisor_override (int): If specified, it will be used as divisor in the averaging calculation,
@ -7547,7 +7547,7 @@ class AvgPool3D(Primitive):
self.add_prim_attr('kernel_size', self.kernel_size)
self.strides = _check_3d_int_or_tuple('strides', strides, self.name, ret_five=True)
self.add_prim_attr('strides', self.strides)
validator.check_value_type('pad', pad, (int, tuple), self.name)
validator.check_value_type('pad', pad, (int, tuple, list), self.name)
if isinstance(pad, int):
pad = (pad,) * 6
if len(pad) != 6:

59
tests/st/nn/test_avgpool1d.py Normal file
View File

@ -0,0 +1,59 @@
# Copyright 2022 Huawei Technologies Co., Ltd
#
# 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.
# ============================================================================
import numpy as np
import pytest
import mindspore as ms
import mindspore.nn as nn
import mindspore.ops as ops
class Net(nn.Cell):
def __init__(self, kernel_size=1, stride=1, pad_mode="valid", padding=0, ceil_mode=False, count_include_pad=True):
super(Net, self).__init__()
self.pool = nn.AvgPool1d(kernel_size=kernel_size, stride=stride, pad_mode=pad_mode, padding=padding,
ceil_mode=ceil_mode, count_include_pad=count_include_pad)
def construct(self, x):
out = self.pool(x)
return out
@pytest.mark.level0
@pytest.mark.platform_arm_ascend_training
@pytest.mark.platform_x86_ascend_training
@pytest.mark.platform_x86_gpu_training
@pytest.mark.platform_x86_cpu
@pytest.mark.platform_arm_cpu
@pytest.mark.env_onecard
@pytest.mark.parametrize('mode', [ms.GRAPH_MODE, ms.PYNATIVE_MODE])
def test_avgpool1d_normal(mode):
"""
Feature: AvgPool1d
Description: Verify the result of AvgPool1d
Expectation: success
"""
ms.set_context(mode=mode)
x1 = ms.Tensor(np.random.randint(0, 10, [1, 3, 6]), ms.float32)
pool1 = Net(kernel_size=6, stride=1)
output1 = pool1(x1)
x2 = ops.randn(6, 6, 8)
pool2 = Net(4, stride=1, ceil_mode=True, pad_mode='pad', padding=2)
output2 = pool2(x2)
assert output1.shape == (1, 3, 1)
assert output2.shape == (6, 6, 9)

61
tests/st/nn/test_avgpool2d.py Normal file
View File

@ -0,0 +1,61 @@
# Copyright 2022 Huawei Technologies Co., Ltd
#
# 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.
# ============================================================================
import numpy as np
import pytest
import mindspore as ms
import mindspore.nn as nn
import mindspore.ops as ops
class Net(nn.Cell):
def __init__(self, kernel_size=1, stride=1, pad_mode="valid", padding=0, ceil_mode=False, count_include_pad=True,
divisor_override=None, data_format='NCHW'):
super(Net, self).__init__()
self.pool = nn.AvgPool2d(kernel_size=kernel_size, stride=stride, pad_mode=pad_mode, padding=padding,
ceil_mode=ceil_mode, count_include_pad=count_include_pad,
divisor_override=divisor_override, data_format=data_format)
def construct(self, x):
out = self.pool(x)
return out
@pytest.mark.level0
@pytest.mark.platform_arm_ascend_training
@pytest.mark.platform_x86_ascend_training
@pytest.mark.platform_x86_gpu_training
@pytest.mark.platform_x86_cpu
@pytest.mark.platform_arm_cpu
@pytest.mark.env_onecard
@pytest.mark.parametrize('mode', [ms.GRAPH_MODE, ms.PYNATIVE_MODE])
def test_avgpool2d_normal(mode):
"""
Feature: AvgPool2d
Description: Verify the result of AvgPool2d
Expectation: success
"""
ms.set_context(mode=mode)
x1 = ms.Tensor(np.random.randint(0, 10, [1, 2, 4, 4]), ms.float32)
pool1 = Net(kernel_size=3, stride=1)
output1 = pool1(x1)
x2 = ops.randn(6, 6, 8, 8)
pool2 = Net(kernel_size=4, stride=1, pad_mode='pad', padding=2, divisor_override=5)
output2 = pool2(x2)
assert output1.shape == (1, 2, 2, 2)
assert output2.shape == (6, 6, 9, 9)

29
tests/st/nn/test_avgpool3d.py
View File

@ -13,17 +13,20 @@
# limitations under the License.
# ============================================================================
import numpy as np
import pytest
import mindspore as ms
import mindspore.nn as nn
import mindspore.ops as ops
class Net(nn.Cell):
def __init__(self):
def __init__(self, kernel_size=1, stride=1, pad_mode="valid", padding=0, ceil_mode=False, count_include_pad=True,
divisor_override=None):
super(Net, self).__init__()
self.pool = nn.AvgPool3d(kernel_size=3, stride=1)
self.pool = nn.AvgPool3d(kernel_size=kernel_size, stride=stride, pad_mode=pad_mode, padding=padding,
ceil_mode=ceil_mode, count_include_pad=count_include_pad,
divisor_override=divisor_override)
def construct(self, x):
out = self.pool(x)
@ -33,6 +36,7 @@ class Net(nn.Cell):
@pytest.mark.level0
@pytest.mark.platform_arm_ascend_training
@pytest.mark.platform_x86_ascend_training
@pytest.mark.platform_x86_gpu_training
@pytest.mark.platform_x86_cpu
@pytest.mark.platform_arm_cpu
@pytest.mark.env_onecard
@ -44,12 +48,13 @@ def test_avgpool3d_normal(mode):
Expectation: success
"""
ms.set_context(mode=mode)
np_array = np.arange(1 * 2 * 4 * 4 * 5).reshape((1, 2, 4, 4, 5))
net = Net()
x = ms.Tensor(np_array, ms.float32)
output = net(x)
expect_output = np.array([[[[[26.0, 27.0, 28.0], [31.0, 32.0, 33.0]], [[46.0, 47.0, 48.0], [51.0, 52.0, 53.0]]],
[[[106.0, 107.0, 108.0], [111.0, 112.0, 113.0]],
[[126.0, 127.0, 128.0], [131.0, 132.0, 133.0]]]]])
assert output.shape == (1, 2, 2, 2, 3)
assert np.allclose(output.asnumpy(), expect_output, rtol=1e-3)
x1 = ops.randn(1, 2, 4, 4, 5).astype(ms.float32)
pool1 = Net(kernel_size=3, stride=1)
output1 = pool1(x1)
x2 = ops.randn(6, 5, 7, 7, 5).astype(ms.float32)
pool2 = Net(kernel_size=4, stride=2, pad_mode='pad', padding=(2, 2, 1), divisor_override=10)
output2 = pool2(x2)
assert output1.shape == (1, 2, 2, 2, 3)
assert output2.shape == (6, 5, 4, 4, 2)