.. _l-onnx-doc-LpPool: ====== LpPool ====== .. contents:: :local: .. _l-onnx-op-lppool-18: LpPool - 18 =========== **Version** * **name**: `LpPool (GitHub) `_ * **domain**: **main** * **since_version**: **18** * **function**: False * **support_level**: SupportType.COMMON * **shape inference**: True This version of the operator has been available **since version 18**. **Summary** LpPool consumes an input tensor X and applies Lp pooling across the tensor according to kernel sizes, stride sizes, and pad lengths. Lp pooling consisting of computing the Lp norm on all values of a subset of the input tensor according to the kernel size and downsampling the data into the output tensor Y for further processing. The output spatial shape will be following: :: output_spatial_shape[i] = floor((input_spatial_shape[i] + pad_shape[i] - {kernelSpatialShape}) / strides_spatial_shape[i] + 1) or :: output_spatial_shape[i] = ceil((input_spatial_shape[i] + pad_shape[i] - {kernelSpatialShape}) / strides_spatial_shape[i] + 1) if ceil_mode is enabled :: * pad_shape[i] is sum of pads along axis i `auto_pad` is a DEPRECATED attribute. If you are using them currently, the output spatial shape will be following: :: VALID: output_spatial_shape[i] = ceil((input_spatial_shape[i] - {kernelSpatialShape} + 1) / strides_spatial_shape[i]) SAME_UPPER or SAME_LOWER: output_spatial_shape[i] = ceil(input_spatial_shape[i] / strides_spatial_shape[i]) And pad shape will be following if `SAME_UPPER` or `SAME_LOWER`: :: pad_shape[i] = (output_spatial_shape[i] - 1) * strides_spatial_shape[i] + {kernelSpatialShape} - input_spatial_shape[i] **Attributes** * **auto_pad**: auto_pad must be either NOTSET, SAME_UPPER, SAME_LOWER or VALID. Where default value is NOTSET, which means explicit padding is used. SAME_UPPER or SAME_LOWER mean pad the input so that `output_shape[i] = ceil(input_shape[i] / strides[i])` for each axis `i`. The padding is split between the two sides equally or almost equally (depending on whether it is even or odd). In case the padding is an odd number, the extra padding is added at the end for SAME_UPPER and at the beginning for SAME_LOWER. Default value is ``'NOTSET'``. * **ceil_mode**: Whether to use ceil or floor (default) to compute the output shape. Default value is ``0``. * **dilations**: dilation value along each spatial axis of the filter. If not present, the dilation defaults is 1 along each spatial axis. * **kernel_shape** (required): The size of the kernel along each axis. * **p**: p value of the Lp norm used to pool over the input data. Default value is ``2``. * **pads**: Padding for the beginning and ending along each spatial axis, it can take any value greater than or equal to 0. The value represent the number of pixels added to the beginning and end part of the corresponding axis. `pads` format should be as follow [x1_begin, x2_begin...x1_end, x2_end,...], where xi_begin the number of pixels added at the beginning of axis `i` and xi_end, the number of pixels added at the end of axis `i`. This attribute cannot be used simultaneously with auto_pad attribute. If not present, the padding defaults to 0 along start and end of each spatial axis. * **strides**: Stride along each spatial axis. If not present, the stride defaults to 1 along each spatial axis. **Inputs** * **X** (heterogeneous) - **T**: Input data tensor from the previous operator; dimensions for image case are (N x C x H x W), where N is the batch size, C is the number of channels, and H and W are the height and the width of the data. For non image case, the dimensions are in the form of (N x C x D1 x D2 ... Dn), where N is the batch size. **Outputs** * **Y** (heterogeneous) - **T**: Output data tensor from Lp pooling across the input tensor. Dimensions will vary based on various kernel, stride, and pad sizes. **Type Constraints** * **T** in ( tensor(double), tensor(float), tensor(float16) ): Constrain input and output types to float tensors. **Examples** **Differences** .. raw:: html
00LpPool consumes an input tensor X and applies Lp pooling acrossLpPool consumes an input tensor X and applies Lp pooling across
11the tensor according to kernel sizes, stride sizes, and pad lengths.the tensor according to kernel sizes, stride sizes, and pad lengths.
22Lp pooling consisting of computing the Lp norm on all values of a subsetLp pooling consisting of computing the Lp norm on all values of a subset
33of the input tensor according to the kernel size and downsampling theof the input tensor according to the kernel size and downsampling the
4data into the output tensor Y for further processing. The output spatial shape will be following:
5::
6
7 output_spatial_shape[i] = floor((input_spatial_shape[i] + pad_shape[i] - {kernelSpatialShape}) / strides_spatial_shape[i] + 1)
8
9or
10::
11
12 output_spatial_shape[i] = ceil((input_spatial_shape[i] + pad_shape[i] - {kernelSpatialShape}) / strides_spatial_shape[i] + 1)
13
14if ceil_mode is enabled
15
16::
17
18 * pad_shape[i] is sum of pads along axis i
19
420data into the output tensor Y for further processing.auto_pad is a DEPRECATED attribute. If you are using them currently, the output spatial shape will be following:
21::
522
23 VALID: output_spatial_shape[i] = ceil((input_spatial_shape[i] - {kernelSpatialShape} + 1) / strides_spatial_shape[i])
24 SAME_UPPER or SAME_LOWER: output_spatial_shape[i] = ceil(input_spatial_shape[i] / strides_spatial_shape[i])
25
26And pad shape will be following if SAME_UPPER or SAME_LOWER:
27::
28
29 pad_shape[i] = (output_spatial_shape[i] - 1) * strides_spatial_shape[i] + {kernelSpatialShape} - input_spatial_shape[i]
30
631**Attributes****Attributes**
732
833* **auto_pad**:* **auto_pad**:
934 auto_pad must be either NOTSET, SAME_UPPER, SAME_LOWER or VALID. auto_pad must be either NOTSET, SAME_UPPER, SAME_LOWER or VALID.
1035 Where default value is NOTSET, which means explicit padding is used. Where default value is NOTSET, which means explicit padding is used.
1136 SAME_UPPER or SAME_LOWER mean pad the input so that output_shape[i] SAME_UPPER or SAME_LOWER mean pad the input so that output_shape[i]
1237 = ceil(input_shape[i] / strides[i]) for each axis i. The padding = ceil(input_shape[i] / strides[i]) for each axis i. The padding
1338 is split between the two sides equally or almost equally (depending is split between the two sides equally or almost equally (depending
1439 on whether it is even or odd). In case the padding is an odd number, on whether it is even or odd). In case the padding is an odd number,
1540 the extra padding is added at the end for SAME_UPPER and at the the extra padding is added at the end for SAME_UPPER and at the
1641 beginning for SAME_LOWER. Default value is 'NOTSET'. beginning for SAME_LOWER. Default value is 'NOTSET'.
42* **ceil_mode**:
43 Whether to use ceil or floor (default) to compute the output shape. Default value is 0.
44* **dilations**:
45 dilation value along each spatial axis of the filter. If not
46 present, the dilation defaults is 1 along each spatial axis.
1747* **kernel_shape** (required):* **kernel_shape** (required):
1848 The size of the kernel along each axis. The size of the kernel along each axis.
1949* **p**:* **p**:
2050 p value of the Lp norm used to pool over the input data. Default value is 2. p value of the Lp norm used to pool over the input data. Default value is 2.
2151* **pads**:* **pads**:
2252 Padding for the beginning and ending along each spatial axis, it can Padding for the beginning and ending along each spatial axis, it can
2353 take any value greater than or equal to 0. The value represent the take any value greater than or equal to 0. The value represent the
2454 number of pixels added to the beginning and end part of the number of pixels added to the beginning and end part of the
2555 corresponding axis. pads format should be as follow [x1_begin, corresponding axis. pads format should be as follow [x1_begin,
2656 x2_begin...x1_end, x2_end,...], where xi_begin the number of pixels x2_begin...x1_end, x2_end,...], where xi_begin the number of pixels
2757 added at the beginning of axis i and xi_end, the number of pixels added at the beginning of axis i and xi_end, the number of pixels
2858 added at the end of axis i. This attribute cannot be used added at the end of axis i. This attribute cannot be used
2959 simultaneously with auto_pad attribute. If not present, the padding simultaneously with auto_pad attribute. If not present, the padding
3060 defaults to 0 along start and end of each spatial axis. defaults to 0 along start and end of each spatial axis.
3161* **strides**:* **strides**:
3262 Stride along each spatial axis. If not present, the stride defaults Stride along each spatial axis. If not present, the stride defaults
3363 to 1 along each spatial axis. to 1 along each spatial axis.
3464
3565**Inputs****Inputs**
3666
3767* **X** (heterogeneous) - **T**:* **X** (heterogeneous) - **T**:
3868 Input data tensor from the previous operator; dimensions for image Input data tensor from the previous operator; dimensions for image
3969 case are (N x C x H x W), where N is the batch size, C is the number case are (N x C x H x W), where N is the batch size, C is the number
4070 of channels, and H and W are the height and the width of the data. of channels, and H and W are the height and the width of the data.
4171 For non image case, the dimensions are in the form of (N x C x D1 x For non image case, the dimensions are in the form of (N x C x D1 x
4272 D2 ... Dn), where N is the batch size. D2 ... Dn), where N is the batch size.
4373
4474**Outputs****Outputs**
4575
4676* **Y** (heterogeneous) - **T**:* **Y** (heterogeneous) - **T**:
4777 Output data tensor from Lp pooling across the input tensor. Output data tensor from Lp pooling across the input tensor.
4878 Dimensions will vary based on various kernel, stride, and pad sizes. Dimensions will vary based on various kernel, stride, and pad sizes.
4979
5080**Type Constraints****Type Constraints**
5181
5282* **T** in (* **T** in (
5383 tensor(double), tensor(double),
5484 tensor(float), tensor(float),
5585 tensor(float16) tensor(float16)
5686 ): ):
5787 Constrain input and output types to float tensors. Constrain input and output types to float tensors.
.. _l-onnx-op-lppool-11: LpPool - 11 =========== **Version** * **name**: `LpPool (GitHub) `_ * **domain**: **main** * **since_version**: **11** * **function**: False * **support_level**: SupportType.COMMON * **shape inference**: True This version of the operator has been available **since version 11**. **Summary** LpPool consumes an input tensor X and applies Lp pooling across the tensor according to kernel sizes, stride sizes, and pad lengths. Lp pooling consisting of computing the Lp norm on all values of a subset of the input tensor according to the kernel size and downsampling the data into the output tensor Y for further processing. **Attributes** * **auto_pad**: auto_pad must be either NOTSET, SAME_UPPER, SAME_LOWER or VALID. Where default value is NOTSET, which means explicit padding is used. SAME_UPPER or SAME_LOWER mean pad the input so that `output_shape[i] = ceil(input_shape[i] / strides[i])` for each axis `i`. The padding is split between the two sides equally or almost equally (depending on whether it is even or odd). In case the padding is an odd number, the extra padding is added at the end for SAME_UPPER and at the beginning for SAME_LOWER. Default value is ``'NOTSET'``. * **kernel_shape** (required): The size of the kernel along each axis. * **p**: p value of the Lp norm used to pool over the input data. Default value is ``2``. * **pads**: Padding for the beginning and ending along each spatial axis, it can take any value greater than or equal to 0. The value represent the number of pixels added to the beginning and end part of the corresponding axis. `pads` format should be as follow [x1_begin, x2_begin...x1_end, x2_end,...], where xi_begin the number of pixels added at the beginning of axis `i` and xi_end, the number of pixels added at the end of axis `i`. This attribute cannot be used simultaneously with auto_pad attribute. If not present, the padding defaults to 0 along start and end of each spatial axis. * **strides**: Stride along each spatial axis. If not present, the stride defaults to 1 along each spatial axis. **Inputs** * **X** (heterogeneous) - **T**: Input data tensor from the previous operator; dimensions for image case are (N x C x H x W), where N is the batch size, C is the number of channels, and H and W are the height and the width of the data. For non image case, the dimensions are in the form of (N x C x D1 x D2 ... Dn), where N is the batch size. **Outputs** * **Y** (heterogeneous) - **T**: Output data tensor from Lp pooling across the input tensor. Dimensions will vary based on various kernel, stride, and pad sizes. **Type Constraints** * **T** in ( tensor(double), tensor(float), tensor(float16) ): Constrain input and output types to float tensors. **Differences** .. raw:: html
00LpPool consumes an input tensor X and applies Lp pooling acrossLpPool consumes an input tensor X and applies Lp pooling across
11the tensor according to kernel sizes, stride sizes, and pad lengths.the tensor according to kernel sizes, stride sizes, and pad lengths.
22Lp pooling consisting of computing the Lp norm on all values of a subsetLp pooling consisting of computing the Lp norm on all values of a subset
33of the input tensor according to the kernel size and downsampling theof the input tensor according to the kernel size and downsampling the
44data into the output tensor Y for further processing.data into the output tensor Y for further processing.
55
66**Attributes****Attributes**
77
88* **auto_pad**:* **auto_pad**:
99 auto_pad must be either NOTSET, SAME_UPPER, SAME_LOWER or VALID. auto_pad must be either NOTSET, SAME_UPPER, SAME_LOWER or VALID.
1010 Where default value is NOTSET, which means explicit padding is used. Where default value is NOTSET, which means explicit padding is used.
1111 SAME_UPPER or SAME_LOWER mean pad the input so that the output SAME_UPPER or SAME_LOWER mean pad the input so that output_shape[i]
12 spatial size match the input.In case of odd number add the extra
12 = ceil(input_shape[i] / strides[i]) for each axis i. The padding
13 is split between the two sides equally or almost equally (depending
14 on whether it is even or odd). In case the padding is an odd number,
1315 padding at the end for SAME_UPPER and at the beginning for the extra padding is added at the end for SAME_UPPER and at the
1416 SAME_LOWER. VALID mean no padding. Default value is 'NOTSET'. beginning for SAME_LOWER. Default value is 'NOTSET'.
1517* **kernel_shape** (required):* **kernel_shape** (required):
1618 The size of the kernel along each axis. The size of the kernel along each axis.
1719* **p**:* **p**:
1820 p value of the Lp norm used to pool over the input data. Default value is 2. p value of the Lp norm used to pool over the input data. Default value is 2.
1921* **pads**:* **pads**:
2022 Padding for the beginning and ending along each spatial axis, it can Padding for the beginning and ending along each spatial axis, it can
2123 take any value greater than or equal to 0. The value represent the take any value greater than or equal to 0. The value represent the
2224 number of pixels added to the beginning and end part of the number of pixels added to the beginning and end part of the
2325 corresponding axis. pads format should be as follow [x1_begin, corresponding axis. pads format should be as follow [x1_begin,
2426 x2_begin...x1_end, x2_end,...], where xi_begin the number of pixels x2_begin...x1_end, x2_end,...], where xi_begin the number of pixels
2527 added at the beginning of axis i and xi_end, the number of pixels added at the beginning of axis i and xi_end, the number of pixels
2628 added at the end of axis i. This attribute cannot be used added at the end of axis i. This attribute cannot be used
2729 simultaneously with auto_pad attribute. If not present, the padding simultaneously with auto_pad attribute. If not present, the padding
2830 defaults to 0 along start and end of each spatial axis. defaults to 0 along start and end of each spatial axis.
2931* **strides**:* **strides**:
3032 Stride along each spatial axis. Stride along each spatial axis. If not present, the stride defaults
33 to 1 along each spatial axis.
3134
3235**Inputs****Inputs**
3336
3437* **X** (heterogeneous) - **T**:* **X** (heterogeneous) - **T**:
3538 Input data tensor from the previous operator; dimensions for image Input data tensor from the previous operator; dimensions for image
3639 case are (N x C x H x W), where N is the batch size, C is the number case are (N x C x H x W), where N is the batch size, C is the number
3740 of channels, and H and W are the height and the width of the data. of channels, and H and W are the height and the width of the data.
3841 For non image case, the dimensions are in the form of (N x C x D1 x For non image case, the dimensions are in the form of (N x C x D1 x
3942 D2 ... Dn), where N is the batch size. D2 ... Dn), where N is the batch size.
4043
4144**Outputs****Outputs**
4245
4346* **Y** (heterogeneous) - **T**:* **Y** (heterogeneous) - **T**:
4447 Output data tensor from Lp pooling across the input tensor. Output data tensor from Lp pooling across the input tensor.
4548 Dimensions will vary based on various kernel, stride, and pad sizes. Dimensions will vary based on various kernel, stride, and pad sizes.
4649
4750**Type Constraints****Type Constraints**
4851
4952* **T** in (* **T** in (
5053 tensor(double), tensor(double),
5154 tensor(float), tensor(float),
5255 tensor(float16) tensor(float16)
5356 ): ):
5457 Constrain input and output types to float tensors. Constrain input and output types to float tensors.
.. _l-onnx-op-lppool-2: LpPool - 2 ========== **Version** * **name**: `LpPool (GitHub) `_ * **domain**: **main** * **since_version**: **2** * **function**: False * **support_level**: SupportType.COMMON * **shape inference**: True This version of the operator has been available **since version 2**. **Summary** LpPool consumes an input tensor X and applies Lp pooling across the tensor according to kernel sizes, stride sizes, and pad lengths. Lp pooling consisting of computing the Lp norm on all values of a subset of the input tensor according to the kernel size and downsampling the data into the output tensor Y for further processing. **Attributes** * **auto_pad**: auto_pad must be either NOTSET, SAME_UPPER, SAME_LOWER or VALID. Where default value is NOTSET, which means explicit padding is used. SAME_UPPER or SAME_LOWER mean pad the input so that the output spatial size match the input.In case of odd number add the extra padding at the end for SAME_UPPER and at the beginning for SAME_LOWER. VALID mean no padding. Default value is ``'NOTSET'``. * **kernel_shape** (required): The size of the kernel along each axis. * **p**: p value of the Lp norm used to pool over the input data. Default value is ``2``. * **pads**: Padding for the beginning and ending along each spatial axis, it can take any value greater than or equal to 0. The value represent the number of pixels added to the beginning and end part of the corresponding axis. `pads` format should be as follow [x1_begin, x2_begin...x1_end, x2_end,...], where xi_begin the number of pixels added at the beginning of axis `i` and xi_end, the number of pixels added at the end of axis `i`. This attribute cannot be used simultaneously with auto_pad attribute. If not present, the padding defaults to 0 along start and end of each spatial axis. * **strides**: Stride along each spatial axis. **Inputs** * **X** (heterogeneous) - **T**: Input data tensor from the previous operator; dimensions for image case are (N x C x H x W), where N is the batch size, C is the number of channels, and H and W are the height and the width of the data. For non image case, the dimensions are in the form of (N x C x D1 x D2 ... Dn), where N is the batch size. **Outputs** * **Y** (heterogeneous) - **T**: Output data tensor from Lp pooling across the input tensor. Dimensions will vary based on various kernel, stride, and pad sizes. **Type Constraints** * **T** in ( tensor(double), tensor(float), tensor(float16) ): Constrain input and output types to float tensors. **Differences** .. raw:: html
00LpPool consumes an input tensor X and applies Lp pooling across theLpPool consumes an input tensor X and applies Lp pooling across
11the tensor according to kernel sizes, stride sizes, and pad lengths.the tensor according to kernel sizes, stride sizes, and pad lengths.
22Lp pooling consisting of computing the Lp norm on all values of a subsetLp pooling consisting of computing the Lp norm on all values of a subset
33of the input tensor according to the kernel size and downsampling theof the input tensor according to the kernel size and downsampling the
44data into the output tensor Y for further processing.data into the output tensor Y for further processing.
55
66**Attributes****Attributes**
77
88* **auto_pad**:* **auto_pad**:
99 auto_pad must be either NOTSET, SAME_UPPER, SAME_LOWER or VALID. auto_pad must be either NOTSET, SAME_UPPER, SAME_LOWER or VALID.
1010 Where default value is NOTSET, which means explicit padding is used. Where default value is NOTSET, which means explicit padding is used.
1111 SAME_UPPER or SAME_LOWER mean pad the input so that the output size SAME_UPPER or SAME_LOWER mean pad the input so that the output
1212 match the input.In case of odd number add the extra padding at the spatial size match the input.In case of odd number add the extra
1313 end for SAME_UPPER and at the beginning for SAME_LOWER. VALID mean padding at the end for SAME_UPPER and at the beginning for
14 SAME_LOWER. VALID mean no padding. Default value is 'NOTSET'.
1415 no padding. DEPRECATION NOTE: auto_pad is only intended to support* **kernel_shape** (required):
15 legacy uses, and for framework authors, one is explicitly encouraged
16 to use explicit padding specified in the pads attribute. Default value is 'NOTSET'.
17* **kernel_shape**:
1816 The size of the kernel along each axis. The size of the kernel along each axis.
1917* **p**:* **p**:
2018 p value of the Lp norm used to pool over the input data, default is p value of the Lp norm used to pool over the input data. Default value is 2.
21 2.0. Default value is 2.0.
2219* **pads**:* **pads**:
2320 Padding for the beginning and ending along each axis, it can take Padding for the beginning and ending along each spatial axis, it can
2421 any value greater than or equal to 0. The value represent the number take any value greater than or equal to 0. The value represent the
2522 of pixels added to the beginning and end part of the corresponding number of pixels added to the beginning and end part of the
2623 axis. pads format should be as follow [x1_begin, corresponding axis. pads format should be as follow [x1_begin,
2724 x2_begin...x1_end, x2_end,...], where xi_begin the number of pixels x2_begin...x1_end, x2_end,...], where xi_begin the number of pixels
2825 added at the beginning of axis i and xi_end, the number of pixels added at the beginning of axis i and xi_end, the number of pixels
2926 added at the end of axis i. This attribute cannot be used added at the end of axis i. This attribute cannot be used
3027 simultaneously with auto_pad attribute. simultaneously with auto_pad attribute. If not present, the padding
28 defaults to 0 along start and end of each spatial axis.
3129* **strides**:* **strides**:
3230 Stride along each axis. Stride along each spatial axis.
3331
3432**Inputs****Inputs**
3533
3634* **X** (heterogeneous) - **T**:* **X** (heterogeneous) - **T**:
3735 Input data tensor from the previous operator; dimensions for image Input data tensor from the previous operator; dimensions for image
3836 case are (N x C x H x W), where N is the batch size, C is the number case are (N x C x H x W), where N is the batch size, C is the number
3937 of channels, and H and W are the height and the width of the data. of channels, and H and W are the height and the width of the data.
4038 For non image case, the dimension are in the form of (N x C x D1 x For non image case, the dimensions are in the form of (N x C x D1 x
4139 D2 ... Dn), where N is the batch size. D2 ... Dn), where N is the batch size.
4240
4341**Outputs****Outputs**
4442
4543* **Y** (heterogeneous) - **T**:* **Y** (heterogeneous) - **T**:
4644 Output data tensor from Lp pooling across the input tensor. Output data tensor from Lp pooling across the input tensor.
4745 Dimensions will vary based on various kernel, stride, and pad sizes. Dimensions will vary based on various kernel, stride, and pad sizes.
4846
4947**Type Constraints****Type Constraints**
5048
5149* **T** in (* **T** in (
5250 tensor(double), tensor(double),
5351 tensor(float), tensor(float),
5452 tensor(float16) tensor(float16)
5553 ): ):
5654 Constrain input and output types to float tensors. Constrain input and output types to float tensors.
.. _l-onnx-op-lppool-1: LpPool - 1 ========== **Version** * **name**: `LpPool (GitHub) `_ * **domain**: **main** * **since_version**: **1** * **function**: False * **support_level**: SupportType.COMMON * **shape inference**: False This version of the operator has been available **since version 1**. **Summary** LpPool consumes an input tensor X and applies Lp pooling across the the tensor according to kernel sizes, stride sizes, and pad lengths. Lp pooling consisting of computing the Lp norm on all values of a subset of the input tensor according to the kernel size and downsampling the data into the output tensor Y for further processing. **Attributes** * **auto_pad**: auto_pad must be either NOTSET, SAME_UPPER, SAME_LOWER or VALID. Where default value is NOTSET, which means explicit padding is used. SAME_UPPER or SAME_LOWER mean pad the input so that the output size match the input.In case of odd number add the extra padding at the end for SAME_UPPER and at the beginning for SAME_LOWER. VALID mean no padding. DEPRECATION NOTE: auto_pad is only intended to support legacy uses, and for framework authors, one is explicitly encouraged to use explicit padding specified in the pads attribute. Default value is ``'NOTSET'``. * **kernel_shape**: The size of the kernel along each axis. * **p**: p value of the Lp norm used to pool over the input data, default is 2.0. Default value is ``2.0``. * **pads**: Padding for the beginning and ending along each axis, it can take any value greater than or equal to 0. The value represent the number of pixels added to the beginning and end part of the corresponding axis. `pads` format should be as follow [x1_begin, x2_begin...x1_end, x2_end,...], where xi_begin the number of pixels added at the beginning of axis `i` and xi_end, the number of pixels added at the end of axis `i`. This attribute cannot be used simultaneously with auto_pad attribute. * **strides**: Stride along each axis. **Inputs** * **X** (heterogeneous) - **T**: Input data tensor from the previous operator; dimensions for image case are (N x C x H x W), where N is the batch size, C is the number of channels, and H and W are the height and the width of the data. For non image case, the dimension are in the form of (N x C x D1 x D2 ... Dn), where N is the batch size. **Outputs** * **Y** (heterogeneous) - **T**: Output data tensor from Lp pooling across the input tensor. Dimensions will vary based on various kernel, stride, and pad sizes. **Type Constraints** * **T** in ( tensor(double), tensor(float), tensor(float16) ): Constrain input and output types to float tensors.