/usr/local/lib64/python3.6/site-packages/torch/nn/modules/__pycache__
Edit: /usr/local/lib64/python3.6/site-packages/torch/nn/modules/__pycache__/conv.cpython-36.pyc (57789B)
3
Eg @ s d dl Z d dlZd dlZd dlmZ d dlmZmZ ddlmZ ddlm
Z
ddl m Z dd l
mZ dd
lmZmZmZmZ d d lmZ dd lmZmZmZ d d
lmZm Z m Z m Z dddZ G dd deZ G dd de Z!G dd de Z"G dd de Z#G dd de Z$G dd d e$Z%G d d d e$Z&G d d d e$Z'G d!d" d"e$Z(G d#d$ d$e Z)G d%d& d&e)e!Z*G d'd( d(e)e"Z+G d)d* d*e)e#Z,G d+d, d,e)e%Z-G d-d. d.e)e&Z.G d/d0 d0e)e'Z/dS )1 N)Tensor) ParameterUninitializedParameter )
functional)init )LazyModuleMixin)Module)_single_pair_triple_reverse_repeat_tuple)reproducibility_notes) _size_1_t _size_2_t _size_3_t)OptionalListTupleUniona * :attr:`groups` controls the connections between inputs and outputs.
:attr:`in_channels` and :attr:`out_channels` must both be divisible by
:attr:`groups`. For example,
* At groups=1, all inputs are convolved to all outputs.
* At groups=2, the operation becomes equivalent to having two conv
layers side by side, each seeing half the input channels
and producing half the output channels, and both subsequently
concatenated.
* At groups= :attr:`in_channels`, each input channel is convolved with
its own set of filters (of size
:math:`\frac{\text{out\_channels}}{\text{in\_channels}}`).a When `groups == in_channels` and `out_channels == K * in_channels`,
where `K` is a positive integer, this operation is also known as a "depthwise convolution".
In other words, for an input of size :math:`(N, C_{in}, L_{in})`,
a depthwise convolution with a depthwise multiplier `K` can be performed with the arguments
:math:`(C_\text{in}=C_\text{in}, C_\text{out}=C_\text{in} \times \text{K}, ..., \text{groups}=C_\text{in})`.)Z groups_noteZdepthwise_separable_notec s. e Zd ZU ddddddddd g Zd
eej iZeeee ed d d
Ze
e e e
ee df ee df eeee df f ee df eee df e eeee de e ee df ee df ee df ee df eee df e eedd fdd
Z ddddZ dd Z fddZ Z S )_ConvNdstridepaddingdilationgroups padding_modeoutput_padding in_channels out_channels kernel_sizebias)inputweightr! returnc C s d S )N )selfr" r# r! r% r% A/usr/local/lib64/python3.6/site-packages/torch/nn/modules/conv.py
_conv_forward2 s z_ConvNd._conv_forward.N) r r r r r r
transposedr r r! r r$ c s | |
d}t t| j || dkr,td|| dkr@tdddh}t|tr||krjtdj|||dkrtdd |D rtd
d d d
dh}| |krtdj|| || _|| _ || _
|| _ || _ || _
|| _|| _| | _| | _t| j tr~ddgt| | _|dkrxtt||tt|d ddD ]D\}}}||d }|d }|| jd| < || | jd| d < q4W nt| j d| _|rttj||| f|f|| _n ttj||| f|f|| _|
rttj|f|| _n | j dd | j d S )N)devicedtyper z'in_channels must be divisible by groupsz(out_channels must be divisible by groupsZsameZvalidz0Invalid padding string {!r}, should be one of {}c s s | ]}|d kV qdS )r Nr% ).0sr% r% r'
] s z#_ConvNd.__init__..z8padding='same' is not supported for strided convolutionszerosZreflectZ replicateZcircularz9padding_mode must be one of {}, but got padding_mode='{}'r r r! r0 ) superr __init__
ValueError
isinstancestrformatanyr r r r r r r) r r r len _reversed_padding_repeated_twicezipranger r torchemptyr# r! Zregister_parameterreset_parameters)r& r r r r r r r) r r r! r r* r+ factory_kwargsZvalid_padding_stringsZvalid_padding_modesdkiZ
total_paddingZleft_pad) __class__r% r' r2 C s^
z_ConvNd.__init__)r$ c C sT t j| jtjdd | jd k rPt j| j\}}dtj| }t j| j| | d S )N )ar )r Zkaiming_uniform_r# mathsqrtr! Z _calculate_fan_in_and_fan_outZuniform_)r& Zfan_in_boundr% r% r' r> s
z_ConvNd.reset_parametersc C s d}| j d t| j kr |d7 }| jd t| j kr<|d7 }| jd
t| j krX|d7 }| jdkrj|d7 }| jd kr||d7 }| jd kr|d
7 }|jf | jS )NzI{in_channels}, {out_channels}, kernel_size={kernel_size}, stride={stride}r z, padding={padding}r z, dilation={dilation}z!, output_padding={output_padding}z, groups={groups}z , bias=Falser/ z , padding_mode={padding_mode})r )r )r ) r r8 r r r r! r r6 __dict__)r& r- r% r% r'
extra_repr s
z_ConvNd.extra_reprc s$ t t| j| t| ds d| _d S )Nr r/ )r1 r __setstate__hasattrr )r& state)rC r% r' rL s
z_ConvNd.__setstate__)NN)!__name__
__module__ __qualname__Z
__constants__r r< r __annotations__r( intZ _in_channelsr r9 r r r r r r5 r r boolr) r r r r# r! r2 r> rK rL
__classcell__r% r% )rC r' r + s.
R<
r c s| e Zd Zddjf ee d Zdeeeee e
ef eee e
dd
fd
d
Z e
e
ee
d d
dZe
e
dddZ ZS )Conv1da Applies a 1D convolution over an input signal composed of several input
planes.
In the simplest case, the output value of the layer with input size
:math:`(N, C_{\text{in}}, L)` and output :math:`(N, C_{\text{out}}, L_{\text{out}})` can be
precisely described as:
.. math::
\text{out}(N_i, C_{\text{out}_j}) = \text{bias}(C_{\text{out}_j}) +
\sum_{k = 0}^{C_{in} - 1} \text{weight}(C_{\text{out}_j}, k)
\star \text{input}(N_i, k)
where :math:`\star` is the valid `cross-correlation`_ operator,
:math:`N` is a batch size, :math:`C` denotes a number of channels,
:math:`L` is a length of signal sequence.
u-
This module supports :ref:`TensorFloat32`.
* :attr:`stride` controls the stride for the cross-correlation, a single
number or a one-element tuple.
* :attr:`padding` controls the amount of padding applied to the input. It
can be either a string {{'valid', 'same'}} or a tuple of ints giving the
amount of implicit padding applied on both sides.
* :attr:`dilation` controls the spacing between the kernel points; also
known as the à trous algorithm. It is harder to describe, but this `link`_
has a nice visualization of what :attr:`dilation` does.
{groups_note}
Note:
{depthwise_separable_note}
Note:
{cudnn_reproducibility_note}
Note:
``padding='valid'`` is the same as no padding. ``padding='same'`` pads
the input so the output has the shape as the input. However, this mode
doesn't support any stride values other than 1.
Args:
in_channels (int): Number of channels in the input image
out_channels (int): Number of channels produced by the convolution
kernel_size (int or tuple): Size of the convolving kernel
stride (int or tuple, optional): Stride of the convolution. Default: 1
padding (int, tuple or str, optional): Padding added to both sides of
the input. Default: 0
padding_mode (string, optional): ``'zeros'``, ``'reflect'``,
``'replicate'`` or ``'circular'``. Default: ``'zeros'``
dilation (int or tuple, optional): Spacing between kernel
elements. Default: 1
groups (int, optional): Number of blocked connections from input
channels to output channels. Default: 1
bias (bool, optional): If ``True``, adds a learnable bias to the
output. Default: ``True``
a:
Shape:
- Input: :math:`(N, C_{in}, L_{in})`
- Output: :math:`(N, C_{out}, L_{out})` where
.. math::
L_{out} = \left\lfloor\frac{L_{in} + 2 \times \text{padding} - \text{dilation}
\times (\text{kernel\_size} - 1) - 1}{\text{stride}} + 1\right\rfloor
Attributes:
weight (Tensor): the learnable weights of the module of shape
:math:`(\text{out\_channels},
\frac{\text{in\_channels}}{\text{groups}}, \text{kernel\_size})`.
The values of these weights are sampled from
:math:`\mathcal{U}(-\sqrt{k}, \sqrt{k})` where
:math:`k = \frac{groups}{C_\text{in} * \text{kernel\_size}}`
bias (Tensor): the learnable bias of the module of shape
(out_channels). If :attr:`bias` is ``True``, then the values of these weights are
sampled from :math:`\mathcal{U}(-\sqrt{k}, \sqrt{k})` where
:math:`k = \frac{groups}{C_\text{in} * \text{kernel\_size}}`
Examples::
>>> m = nn.Conv1d(16, 33, 3, stride=2)
>>> input = torch.randn(20, 16, 50)
>>> output = m(input)
.. _cross-correlation:
https://en.wikipedia.org/wiki/Cross-correlation
.. _link:
https://github.com/vdumoulin/conv_arithmetic/blob/master/README.md
r r Tr/ N)
r r r r r r r r! r r$ c sh |
| d} t |}
t |}t|tr(|nt |}t |}tt| j|||
|||dt d||| f | d S )N)r* r+ Fr )r r4 r5 r1 rV r2 )r& r r r r r r r r! r r* r+ r? kernel_size_stride_padding_ dilation_)rC r% r' r2 s
zConv1d.__init__)r" r# r! c C sX | j dkr:tjtj|| j| j d||| jtd| j| jS tj|||| j| j | j| jS )Nr/ )moder )
r FZconv1dpadr9 r r r r r )r& r" r# r! r% r% r' r( $ s
zConv1d._conv_forward)r" r$ c C s | j || j| jS )N)r( r# r! )r& r" r% r% r' forward, s zConv1d.forward)r r r r Tr/ NN)rO rP rQ r6 r convolution_notes__doc__rS r r r5 rT r2 r r r( r^ rU r% r% )rC r' rV s ;! rV c s| e Zd Zddjf ee d Zdeeeee e
ef eee e
dd
fd
d
Z e
e
ee
d d
dZe
e
dddZ ZS )Conv2da Applies a 2D convolution over an input signal composed of several input
planes.
In the simplest case, the output value of the layer with input size
:math:`(N, C_{\text{in}}, H, W)` and output :math:`(N, C_{\text{out}}, H_{\text{out}}, W_{\text{out}})`
can be precisely described as:
.. math::
\text{out}(N_i, C_{\text{out}_j}) = \text{bias}(C_{\text{out}_j}) +
\sum_{k = 0}^{C_{\text{in}} - 1} \text{weight}(C_{\text{out}_j}, k) \star \text{input}(N_i, k)
where :math:`\star` is the valid 2D `cross-correlation`_ operator,
:math:`N` is a batch size, :math:`C` denotes a number of channels,
:math:`H` is a height of input planes in pixels, and :math:`W` is
width in pixels.
u
This module supports :ref:`TensorFloat32`.
* :attr:`stride` controls the stride for the cross-correlation, a single
number or a tuple.
* :attr:`padding` controls the amount of padding applied to the input. It
can be either a string {{'valid', 'same'}} or a tuple of ints giving the
amount of implicit padding applied on both sides.
* :attr:`dilation` controls the spacing between the kernel points; also
known as the à trous algorithm. It is harder to describe, but this `link`_
has a nice visualization of what :attr:`dilation` does.
{groups_note}
The parameters :attr:`kernel_size`, :attr:`stride`, :attr:`padding`, :attr:`dilation` 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
Note:
{depthwise_separable_note}
Note:
{cudnn_reproducibility_note}
Note:
``padding='valid'`` is the same as no padding. ``padding='same'`` pads
the input so the output has the shape as the input. However, this mode
doesn't support any stride values other than 1.
Args:
in_channels (int): Number of channels in the input image
out_channels (int): Number of channels produced by the convolution
kernel_size (int or tuple): Size of the convolving kernel
stride (int or tuple, optional): Stride of the convolution. Default: 1
padding (int, tuple or str, optional): Padding added to all four sides of
the input. Default: 0
padding_mode (string, optional): ``'zeros'``, ``'reflect'``,
``'replicate'`` or ``'circular'``. Default: ``'zeros'``
dilation (int or tuple, optional): Spacing between kernel elements. Default: 1
groups (int, optional): Number of blocked connections from input
channels to output channels. Default: 1
bias (bool, optional): If ``True``, adds a learnable bias to the
output. Default: ``True``
a
Shape:
- Input: :math:`(N, C_{in}, H_{in}, W_{in})`
- Output: :math:`(N, C_{out}, H_{out}, W_{out})` where
.. math::
H_{out} = \left\lfloor\frac{H_{in} + 2 \times \text{padding}[0] - \text{dilation}[0]
\times (\text{kernel\_size}[0] - 1) - 1}{\text{stride}[0]} + 1\right\rfloor
.. math::
W_{out} = \left\lfloor\frac{W_{in} + 2 \times \text{padding}[1] - \text{dilation}[1]
\times (\text{kernel\_size}[1] - 1) - 1}{\text{stride}[1]} + 1\right\rfloor
Attributes:
weight (Tensor): the learnable weights of the module of shape
:math:`(\text{out\_channels}, \frac{\text{in\_channels}}{\text{groups}},`
:math:`\text{kernel\_size[0]}, \text{kernel\_size[1]})`.
The values of these weights are sampled from
:math:`\mathcal{U}(-\sqrt{k}, \sqrt{k})` where
:math:`k = \frac{groups}{C_\text{in} * \prod_{i=0}^{1}\text{kernel\_size}[i]}`
bias (Tensor): the learnable bias of the module of shape
(out_channels). If :attr:`bias` is ``True``,
then the values of these weights are
sampled from :math:`\mathcal{U}(-\sqrt{k}, \sqrt{k})` where
:math:`k = \frac{groups}{C_\text{in} * \prod_{i=0}^{1}\text{kernel\_size}[i]}`
Examples:
>>> # With square kernels and equal stride
>>> m = nn.Conv2d(16, 33, 3, stride=2)
>>> # non-square kernels and unequal stride and with padding
>>> m = nn.Conv2d(16, 33, (3, 5), stride=(2, 1), padding=(4, 2))
>>> # non-square kernels and unequal stride and with padding and dilation
>>> m = nn.Conv2d(16, 33, (3, 5), stride=(2, 1), padding=(4, 2), dilation=(3, 1))
>>> input = torch.randn(20, 16, 50, 100)
>>> output = m(input)
.. _cross-correlation:
https://en.wikipedia.org/wiki/Cross-correlation
.. _link:
https://github.com/vdumoulin/conv_arithmetic/blob/master/README.md
r r Tr/ N)
r r r r r r r r! r r$ c sh |
| d} t |}
t |}t|tr(|nt |}t |}tt| j|||
|||dt d||| f | d S )N)r* r+ Fr )r r4 r5 r1 ra r2 )r& r r r r r r r r! r r* r+ r? rW rX rY rZ )rC r% r' r2 s
zConv2d.__init__)r" r# r! c C sX | j dkr:tjtj|| j| j d||| jtd| j| jS tj|||| j| j | j| jS )Nr/ )r[ r )
r r\ Zconv2dr] r9 r r r r r )r& r" r# r! r% r% r' r( s
zConv2d._conv_forward)r" r$ c C s | j || j| jS )N)r( r# r! )r& r" r% r% r' r^ s zConv2d.forward)r r r r Tr/ NN)rO rP rQ r6 r r_ r` rS r r r5 rT r2 r r r( r^ rU r% r% )rC r' ra 0 s A+
ra c s| e Zd Zddjf ee d Zdeeeee e
ef eee e
dd
fd
d
Z e
e
ee
d d
dZe
e
dddZ ZS )Conv3da Applies a 3D convolution over an input signal composed of several input
planes.
In the simplest case, the output value of the layer with input size :math:`(N, C_{in}, D, H, W)`
and output :math:`(N, C_{out}, D_{out}, H_{out}, W_{out})` can be precisely described as:
.. math::
out(N_i, C_{out_j}) = bias(C_{out_j}) +
\sum_{k = 0}^{C_{in} - 1} weight(C_{out_j}, k) \star input(N_i, k)
where :math:`\star` is the valid 3D `cross-correlation`_ operator
uk
This module supports :ref:`TensorFloat32`.
* :attr:`stride` controls the stride for the cross-correlation.
* :attr:`padding` controls the amount of padding applied to the input. It
can be either a string {{'valid', 'same'}} or a tuple of ints giving the
amount of implicit padding applied on both sides.
* :attr:`dilation` controls the spacing between the kernel points; also known as the à trous algorithm.
It is harder to describe, but this `link`_ has a nice visualization of what :attr:`dilation` does.
{groups_note}
The parameters :attr:`kernel_size`, :attr:`stride`, :attr:`padding`, :attr:`dilation` can either be:
- a single ``int`` -- in which case the same value is used for the depth, height and width dimension
- a ``tuple`` of three ints -- in which case, the first `int` is used for the depth dimension,
the second `int` for the height dimension and the third `int` for the width dimension
Note:
{depthwise_separable_note}
Note:
{cudnn_reproducibility_note}
Note:
``padding='valid'`` is the same as no padding. ``padding='same'`` pads
the input so the output has the shape as the input. However, this mode
doesn't support any stride values other than 1.
Args:
in_channels (int): Number of channels in the input image
out_channels (int): Number of channels produced by the convolution
kernel_size (int or tuple): Size of the convolving kernel
stride (int or tuple, optional): Stride of the convolution. Default: 1
padding (int, tuple or str, optional): Padding added to all six sides of
the input. Default: 0
padding_mode (string, optional): ``'zeros'``, ``'reflect'``, ``'replicate'`` or ``'circular'``. Default: ``'zeros'``
dilation (int or tuple, optional): Spacing between kernel elements. Default: 1
groups (int, optional): Number of blocked connections from input channels to output channels. Default: 1
bias (bool, optional): If ``True``, adds a learnable bias to the output. Default: ``True``
a
Shape:
- Input: :math:`(N, C_{in}, D_{in}, H_{in}, W_{in})`
- Output: :math:`(N, C_{out}, D_{out}, H_{out}, W_{out})` where
.. math::
D_{out} = \left\lfloor\frac{D_{in} + 2 \times \text{padding}[0] - \text{dilation}[0]
\times (\text{kernel\_size}[0] - 1) - 1}{\text{stride}[0]} + 1\right\rfloor
.. math::
H_{out} = \left\lfloor\frac{H_{in} + 2 \times \text{padding}[1] - \text{dilation}[1]
\times (\text{kernel\_size}[1] - 1) - 1}{\text{stride}[1]} + 1\right\rfloor
.. math::
W_{out} = \left\lfloor\frac{W_{in} + 2 \times \text{padding}[2] - \text{dilation}[2]
\times (\text{kernel\_size}[2] - 1) - 1}{\text{stride}[2]} + 1\right\rfloor
Attributes:
weight (Tensor): the learnable weights of the module of shape
:math:`(\text{out\_channels}, \frac{\text{in\_channels}}{\text{groups}},`
:math:`\text{kernel\_size[0]}, \text{kernel\_size[1]}, \text{kernel\_size[2]})`.
The values of these weights are sampled from
:math:`\mathcal{U}(-\sqrt{k}, \sqrt{k})` where
:math:`k = \frac{groups}{C_\text{in} * \prod_{i=0}^{2}\text{kernel\_size}[i]}`
bias (Tensor): the learnable bias of the module of shape (out_channels). If :attr:`bias` is ``True``,
then the values of these weights are
sampled from :math:`\mathcal{U}(-\sqrt{k}, \sqrt{k})` where
:math:`k = \frac{groups}{C_\text{in} * \prod_{i=0}^{2}\text{kernel\_size}[i]}`
Examples::
>>> # With square kernels and equal stride
>>> m = nn.Conv3d(16, 33, 3, stride=2)
>>> # non-square kernels and unequal stride and with padding
>>> m = nn.Conv3d(16, 33, (3, 5, 2), stride=(2, 1, 1), padding=(4, 2, 0))
>>> input = torch.randn(20, 16, 10, 50, 100)
>>> output = m(input)
.. _cross-correlation:
https://en.wikipedia.org/wiki/Cross-correlation
.. _link:
https://github.com/vdumoulin/conv_arithmetic/blob/master/README.md
r r Tr/ N)
r r r r r r r r! r r$ c sh |
| d} t |}
t |}t|tr(|nt |}t |}tt| j|||
|||dt d||| f | d S )N)r* r+ Fr )r
r4 r5 r1 rb r2 )r& r r r r r r r r! r r* r+ r? rW rX rY rZ )rC r% r' r2 % s
zConv3d.__init__)r" r# r! c C sX | j dkr:tjtj|| j| j d||| jtd| j| jS tj|||| j| j | j| jS )Nr/ )r[ r )
r r\ Zconv3dr] r9 r r
r r r )r& r" r# r! r% r% r' r( < s
zConv3d._conv_forward)r" r$ c C s | j || j| jS )N)r( r# r! )r& r" r% r% r' r^ M s zConv3d.forward)r r r r Tr/ NN)rO rP rQ r6 r r_ r` rS r r r5 rT r2 r r r( r^ rU r% r% )rC r' rb s 7,
rb c s` e Zd Zddd fdd
Zd eeee ee ee ee eee ee dddZ Z S )
_ConvTransposeNdN)r$ c sP | dkrt dj| jj| |
d}tt| j||||||||| |
| f | d S )Nr/ z-Only "zeros" padding mode is supported for {})r* r+ )r3 r6 rC rO r1 rc r2 )r& r r r r r r r) r r r! r r* r+ r? )rC r% r' r2 S s
z_ConvTransposeNd.__init__)r" output_sizer r r r r$ c C s |d krt | j}n|j d }t||d kr>|dd }t||krdtdj||d t|tjjt t
g } tjjt t
g }
x~t |D ]r} |j | d d || d|| |d k r|| nd|| d d } | j
| |
j
| | || d qW xbt t|D ]R}
||
}| |
}|
|
}||k sF||krtdj|| |
|j dd qW tjjt t
g }x(t |D ] } |j
|| | | qW |}|S )Nr z0output_size must have {} or {} elements (got {})r zXrequested an output size of {}, but valid sizes range from {} to {} (for an input of {}))r r Zdimr8 r3 r6 r< ZjitZannotater rS r; sizeappend)r& r" rd r r r r retrA Z min_sizesZ max_sizesr@ Zdim_sizerB re Zmin_sizemax_sizeresr% r% r' _output_paddinga s8 J
z _ConvTransposeNd._output_padding)NN)N)
rO rP rQ r2 r r r rS rj rU r% r% )rC r' rc R s rc c
sh e Zd Zdjf eed Zdeeeeeeee ee
dd fd d
Z de e
ee e d d d
Z ZS )ConvTranspose1du Applies a 1D transposed convolution operator over an input image
composed of several input planes.
This module can be seen as the gradient of Conv1d with respect to its input.
It is also known as a fractionally-strided convolution or
a deconvolution (although it is not an actual deconvolution operation as it does
not compute a true inverse of convolution). For more information, see the visualizations
`here`_ and the `Deconvolutional Networks`_ paper.
This module supports :ref:`TensorFloat32`.
* :attr:`stride` controls the stride for the cross-correlation.
* :attr:`padding` controls the amount of implicit zero padding on both
sides for ``dilation * (kernel_size - 1) - padding`` number of points. See note
below for details.
* :attr:`output_padding` controls the additional size added to one side
of the output shape. See note below for details.
* :attr:`dilation` controls the spacing between the kernel points; also known as the à trous algorithm.
It is harder to describe, but the link `here`_ has a nice visualization of what :attr:`dilation` does.
{groups_note}
Note:
The :attr:`padding` argument effectively adds ``dilation * (kernel_size - 1) - padding``
amount of zero padding to both sizes of the input. This is set so that
when a :class:`~torch.nn.Conv1d` and a :class:`~torch.nn.ConvTranspose1d`
are initialized with same parameters, they are inverses of each other in
regard to the input and output shapes. However, when ``stride > 1``,
:class:`~torch.nn.Conv1d` maps multiple input shapes to the same output
shape. :attr:`output_padding` is provided to resolve this ambiguity by
effectively increasing the calculated output shape on one side. Note
that :attr:`output_padding` is only used to find output shape, but does
not actually add zero-padding to output.
Note:
In some circumstances when using the CUDA backend with CuDNN, this operator
may select a nondeterministic algorithm to increase performance. If this is
undesirable, you can try to make the operation deterministic (potentially at
a performance cost) by setting ``torch.backends.cudnn.deterministic =
True``.
Please see the notes on :doc:`/notes/randomness` for background.
Args:
in_channels (int): Number of channels in the input image
out_channels (int): Number of channels produced by the convolution
kernel_size (int or tuple): Size of the convolving kernel
stride (int or tuple, optional): Stride of the convolution. Default: 1
padding (int or tuple, optional): ``dilation * (kernel_size - 1) - padding`` zero-padding
will be added to both sides of the input. Default: 0
output_padding (int or tuple, optional): Additional size added to one side
of the output shape. Default: 0
groups (int, optional): Number of blocked connections from input channels to output channels. Default: 1
bias (bool, optional): If ``True``, adds a learnable bias to the output. Default: ``True``
dilation (int or tuple, optional): Spacing between kernel elements. Default: 1
aC
Shape:
- Input: :math:`(N, C_{in}, L_{in})`
- Output: :math:`(N, C_{out}, L_{out})` where
.. math::
L_{out} = (L_{in} - 1) \times \text{stride} - 2 \times \text{padding} + \text{dilation}
\times (\text{kernel\_size} - 1) + \text{output\_padding} + 1
Attributes:
weight (Tensor): the learnable weights of the module of shape
:math:`(\text{in\_channels}, \frac{\text{out\_channels}}{\text{groups}},`
:math:`\text{kernel\_size})`.
The values of these weights are sampled from
:math:`\mathcal{U}(-\sqrt{k}, \sqrt{k})` where
:math:`k = \frac{groups}{C_\text{out} * \text{kernel\_size}}`
bias (Tensor): the learnable bias of the module of shape (out_channels).
If :attr:`bias` is ``True``, then the values of these weights are
sampled from :math:`\mathcal{U}(-\sqrt{k}, \sqrt{k})` where
:math:`k = \frac{groups}{C_\text{out} * \text{kernel\_size}}`
.. _`here`:
https://github.com/vdumoulin/conv_arithmetic/blob/master/README.md
.. _`Deconvolutional Networks`:
https://www.matthewzeiler.com/mattzeiler/deconvolutionalnetworks.pdf
r r Tr/ N) r r r r r r r r! r r r$ c
s^ | | d}
t |}t |}t |}t | } t |}tt| j|||||| d||||
f |
d S )N)r* r+ T)r r1 rk r2 )r& r r r r r r r r! r r r* r+ r? )rC r% r' r2 s
zConvTranspose1d.__init__)r" rd r$ c C sb | j dkrtdt| jts"t| j||| j| j| j| j }t
j || j | j
| j| j|| j| j S )Nr/ z:Only `zeros` padding mode is supported for ConvTranspose1d)r r3 r4 r tupleAssertionErrorrj r r r r\ Zconv_transpose1dr# r! r )r& r" rd r r% r% r' r^ s
zConvTranspose1d.forward) r r r r Tr r/ NN)N)rO rP rQ r6 r r_ r` rS r rT r5 r2 r r r r^ rU r% r% )rC r' rk s ; rk c
sh e Zd Zdjf eed Zdeeeeeeee ee
dd fd d
Z de e
ee e d d d
Z ZS )ConvTranspose2du Applies a 2D transposed convolution operator over an input image
composed of several input planes.
This module can be seen as the gradient of Conv2d with respect to its input.
It is also known as a fractionally-strided convolution or
a deconvolution (although it is not an actual deconvolution operation as it does
not compute a true inverse of convolution). For more information, see the visualizations
`here`_ and the `Deconvolutional Networks`_ paper.
This module supports :ref:`TensorFloat32`.
* :attr:`stride` controls the stride for the cross-correlation.
* :attr:`padding` controls the amount of implicit zero padding on both
sides for ``dilation * (kernel_size - 1) - padding`` number of points. See note
below for details.
* :attr:`output_padding` controls the additional size added to one side
of the output shape. See note below for details.
* :attr:`dilation` controls the spacing between the kernel points; also known as the à trous algorithm.
It is harder to describe, but the link `here`_ has a nice visualization of what :attr:`dilation` does.
{groups_note}
The parameters :attr:`kernel_size`, :attr:`stride`, :attr:`padding`, :attr:`output_padding`
can either be:
- a single ``int`` -- in which case the same value is used for the height and width dimensions
- 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
Note:
The :attr:`padding` argument effectively adds ``dilation * (kernel_size - 1) - padding``
amount of zero padding to both sizes of the input. This is set so that
when a :class:`~torch.nn.Conv2d` and a :class:`~torch.nn.ConvTranspose2d`
are initialized with same parameters, they are inverses of each other in
regard to the input and output shapes. However, when ``stride > 1``,
:class:`~torch.nn.Conv2d` maps multiple input shapes to the same output
shape. :attr:`output_padding` is provided to resolve this ambiguity by
effectively increasing the calculated output shape on one side. Note
that :attr:`output_padding` is only used to find output shape, but does
not actually add zero-padding to output.
Note:
{cudnn_reproducibility_note}
Args:
in_channels (int): Number of channels in the input image
out_channels (int): Number of channels produced by the convolution
kernel_size (int or tuple): Size of the convolving kernel
stride (int or tuple, optional): Stride of the convolution. Default: 1
padding (int or tuple, optional): ``dilation * (kernel_size - 1) - padding`` zero-padding
will be added to both sides of each dimension in the input. Default: 0
output_padding (int or tuple, optional): Additional size added to one side
of each dimension in the output shape. Default: 0
groups (int, optional): Number of blocked connections from input channels to output channels. Default: 1
bias (bool, optional): If ``True``, adds a learnable bias to the output. Default: ``True``
dilation (int or tuple, optional): Spacing between kernel elements. Default: 1
a
Shape:
- Input: :math:`(N, C_{in}, H_{in}, W_{in})`
- Output: :math:`(N, C_{out}, H_{out}, W_{out})` where
.. math::
H_{out} = (H_{in} - 1) \times \text{stride}[0] - 2 \times \text{padding}[0] + \text{dilation}[0]
\times (\text{kernel\_size}[0] - 1) + \text{output\_padding}[0] + 1
.. math::
W_{out} = (W_{in} - 1) \times \text{stride}[1] - 2 \times \text{padding}[1] + \text{dilation}[1]
\times (\text{kernel\_size}[1] - 1) + \text{output\_padding}[1] + 1
Attributes:
weight (Tensor): the learnable weights of the module of shape
:math:`(\text{in\_channels}, \frac{\text{out\_channels}}{\text{groups}},`
:math:`\text{kernel\_size[0]}, \text{kernel\_size[1]})`.
The values of these weights are sampled from
:math:`\mathcal{U}(-\sqrt{k}, \sqrt{k})` where
:math:`k = \frac{groups}{C_\text{out} * \prod_{i=0}^{1}\text{kernel\_size}[i]}`
bias (Tensor): the learnable bias of the module of shape (out_channels)
If :attr:`bias` is ``True``, then the values of these weights are
sampled from :math:`\mathcal{U}(-\sqrt{k}, \sqrt{k})` where
:math:`k = \frac{groups}{C_\text{out} * \prod_{i=0}^{1}\text{kernel\_size}[i]}`
Examples::
>>> # With square kernels and equal stride
>>> m = nn.ConvTranspose2d(16, 33, 3, stride=2)
>>> # non-square kernels and unequal stride and with padding
>>> m = nn.ConvTranspose2d(16, 33, (3, 5), stride=(2, 1), padding=(4, 2))
>>> input = torch.randn(20, 16, 50, 100)
>>> output = m(input)
>>> # exact output size can be also specified as an argument
>>> input = torch.randn(1, 16, 12, 12)
>>> downsample = nn.Conv2d(16, 16, 3, stride=2, padding=1)
>>> upsample = nn.ConvTranspose2d(16, 16, 3, stride=2, padding=1)
>>> h = downsample(input)
>>> h.size()
torch.Size([1, 16, 6, 6])
>>> output = upsample(h, output_size=input.size())
>>> output.size()
torch.Size([1, 16, 12, 12])
.. _`here`:
https://github.com/vdumoulin/conv_arithmetic/blob/master/README.md
.. _`Deconvolutional Networks`:
https://www.matthewzeiler.com/mattzeiler/deconvolutionalnetworks.pdf
r r Tr/ N) r r r r r r r r! r r r$ c
s^ | | d}
t |}t |}t |}t | } t |}tt| j|||||| d||||
f |
d S )N)r* r+ T)r r1 rn r2 )r& r r r r r r r r! r r r* r+ r? )rC r% r' r2 x s
zConvTranspose2d.__init__)r" rd r$ c C sb | j dkrtdt| jts"t| j||| j| j| j| j }t
j || j | j
| j| j|| j| j S )Nr/ z:Only `zeros` padding mode is supported for ConvTranspose2d)r r3 r4 r rl rm rj r r r r\ Zconv_transpose2dr# r! r )r& r" rd r r% r% r' r^ s
zConvTranspose2d.forward) r r r r Tr r/ NN)N)rO rP rQ r6 r r_ r` rS r rT r5 r2 r r r r^ rU r% r% )rC r' rn s <1 rn c
sh e Zd Zdjf eed Zdeeeeeeee ee
dd fd d
Z de e
ee e d d d
Z ZS )ConvTranspose3du
Applies a 3D transposed convolution operator over an input image composed of several input
planes.
The transposed convolution operator multiplies each input value element-wise by a learnable kernel,
and sums over the outputs from all input feature planes.
This module can be seen as the gradient of Conv3d with respect to its input.
It is also known as a fractionally-strided convolution or
a deconvolution (although it is not an actual deconvolution operation as it does
not compute a true inverse of convolution). For more information, see the visualizations
`here`_ and the `Deconvolutional Networks`_ paper.
This module supports :ref:`TensorFloat32`.
* :attr:`stride` controls the stride for the cross-correlation.
* :attr:`padding` controls the amount of implicit zero padding on both
sides for ``dilation * (kernel_size - 1) - padding`` number of points. See note
below for details.
* :attr:`output_padding` controls the additional size added to one side
of the output shape. See note below for details.
* :attr:`dilation` controls the spacing between the kernel points; also known as the à trous algorithm.
It is harder to describe, but the link `here`_ has a nice visualization of what :attr:`dilation` does.
{groups_note}
The parameters :attr:`kernel_size`, :attr:`stride`, :attr:`padding`, :attr:`output_padding`
can either be:
- a single ``int`` -- in which case the same value is used for the depth, height and width dimensions
- a ``tuple`` of three ints -- in which case, the first `int` is used for the depth dimension,
the second `int` for the height dimension and the third `int` for the width dimension
Note:
The :attr:`padding` argument effectively adds ``dilation * (kernel_size - 1) - padding``
amount of zero padding to both sizes of the input. This is set so that
when a :class:`~torch.nn.Conv3d` and a :class:`~torch.nn.ConvTranspose3d`
are initialized with same parameters, they are inverses of each other in
regard to the input and output shapes. However, when ``stride > 1``,
:class:`~torch.nn.Conv3d` maps multiple input shapes to the same output
shape. :attr:`output_padding` is provided to resolve this ambiguity by
effectively increasing the calculated output shape on one side. Note
that :attr:`output_padding` is only used to find output shape, but does
not actually add zero-padding to output.
Note:
{cudnn_reproducibility_note}
Args:
in_channels (int): Number of channels in the input image
out_channels (int): Number of channels produced by the convolution
kernel_size (int or tuple): Size of the convolving kernel
stride (int or tuple, optional): Stride of the convolution. Default: 1
padding (int or tuple, optional): ``dilation * (kernel_size - 1) - padding`` zero-padding
will be added to both sides of each dimension in the input. Default: 0
output_padding (int or tuple, optional): Additional size added to one side
of each dimension in the output shape. Default: 0
groups (int, optional): Number of blocked connections from input channels to output channels. Default: 1
bias (bool, optional): If ``True``, adds a learnable bias to the output. Default: ``True``
dilation (int or tuple, optional): Spacing between kernel elements. Default: 1
a
Shape:
- Input: :math:`(N, C_{in}, D_{in}, H_{in}, W_{in})`
- Output: :math:`(N, C_{out}, D_{out}, H_{out}, W_{out})` where
.. math::
D_{out} = (D_{in} - 1) \times \text{stride}[0] - 2 \times \text{padding}[0] + \text{dilation}[0]
\times (\text{kernel\_size}[0] - 1) + \text{output\_padding}[0] + 1
.. math::
H_{out} = (H_{in} - 1) \times \text{stride}[1] - 2 \times \text{padding}[1] + \text{dilation}[1]
\times (\text{kernel\_size}[1] - 1) + \text{output\_padding}[1] + 1
.. math::
W_{out} = (W_{in} - 1) \times \text{stride}[2] - 2 \times \text{padding}[2] + \text{dilation}[2]
\times (\text{kernel\_size}[2] - 1) + \text{output\_padding}[2] + 1
Attributes:
weight (Tensor): the learnable weights of the module of shape
:math:`(\text{in\_channels}, \frac{\text{out\_channels}}{\text{groups}},`
:math:`\text{kernel\_size[0]}, \text{kernel\_size[1]}, \text{kernel\_size[2]})`.
The values of these weights are sampled from
:math:`\mathcal{U}(-\sqrt{k}, \sqrt{k})` where
:math:`k = \frac{groups}{C_\text{out} * \prod_{i=0}^{2}\text{kernel\_size}[i]}`
bias (Tensor): the learnable bias of the module of shape (out_channels)
If :attr:`bias` is ``True``, then the values of these weights are
sampled from :math:`\mathcal{U}(-\sqrt{k}, \sqrt{k})` where
:math:`k = \frac{groups}{C_\text{out} * \prod_{i=0}^{2}\text{kernel\_size}[i]}`
Examples::
>>> # With square kernels and equal stride
>>> m = nn.ConvTranspose3d(16, 33, 3, stride=2)
>>> # non-square kernels and unequal stride and with padding
>>> m = nn.ConvTranspose3d(16, 33, (3, 5, 2), stride=(2, 1, 1), padding=(0, 4, 2))
>>> input = torch.randn(20, 16, 10, 50, 100)
>>> output = m(input)
.. _`here`:
https://github.com/vdumoulin/conv_arithmetic/blob/master/README.md
.. _`Deconvolutional Networks`:
https://www.matthewzeiler.com/mattzeiler/deconvolutionalnetworks.pdf
r r Tr/ N) r r r r r r r r! r r r$ c
s^ | | d}
t |}t |}t |}t | } t |}tt| j|||||| d||||
f |
d S )N)r* r+ T)r
r1 ro r2 )r& r r r r r r r r! r r r* r+ r? )rC r% r' r2 s
zConvTranspose3d.__init__)r" rd r$ c C sb | j dkrtdt| jts"t| j||| j| j| j| j }t
j || j | j
| j| j|| j| j S )Nr/ z:Only `zeros` padding mode is supported for ConvTranspose3d)r r3 r4 r rl rm rj r r r r\ Zconv_transpose3dr# r! r )r& r" rd r r% r% r' r^ $ s
zConvTranspose3d.forward) r r r r Tr r/ NN)N)rO rP rQ r6 r r_ r` rS r rT r5 r2 r r r r^ rU r% r% )rC r' ro s >+ ro c s e Zd Z fddZ ZS )_ConvTransposeMixinc s t jd tt| j|| d S )NzV_ConvTransposeMixin is a deprecated internal class. Please consider using public APIs.)warningswarnr1 rp r2 )r& argskwargs)rC r% r' r2 C s z _ConvTransposeMixin.__init__)rO rP rQ r2 rU r% r% )rC r' rp B s rp c sV e Zd ZU eeeee edf
e e
dd fdd ZddddZ ZS )_LazyConvXdMixin.N)r$ c s" | j r | jdkr t j d S )Nr )has_uninitialized_paramsr r1 r> )r& )rC r% r' r> X s z!_LazyConvXdMixin.reset_parametersc C s | j r|jd | _| j| j dkr,tdt| jts )r& r" r% r% r' initialize_parametersa s
z&_LazyConvXdMixin.initialize_parameters)rO rP rQ rS r rT r) r r r r r r# r! r> rx rU r% r% )rC r' ru O s
ru c s< e Zd ZdZeZd
eeeeeeee dd fdd
Z
Z S )
LazyConv1da A :class:`torch.nn.Conv1d` module with lazy initialization of
the ``in_channels`` argument of the :class:`Conv1d` that is inferred from
the ``input.size(1)``.
The attributes that will be lazily initialized are `weight` and `bias`.
Check the :class:`torch.nn.modules.lazy.LazyModuleMixin` for further documentation
on lazy modules and their limitations.
Args:
out_channels (int): Number of channels produced by the convolution
kernel_size (int or tuple): Size of the convolving kernel
stride (int or tuple, optional): Stride of the convolution. Default: 1
padding (int or tuple, optional): Zero-padding added to both sides of
the input. Default: 0
padding_mode (string, optional): ``'zeros'``, ``'reflect'``,
``'replicate'`` or ``'circular'``. Default: ``'zeros'``
dilation (int or tuple, optional): Spacing between kernel
elements. Default: 1
groups (int, optional): Number of blocked connections from input
channels to output channels. Default: 1
bias (bool, optional): If ``True``, adds a learnable bias to the
output. Default: ``True``
.. seealso:: :class:`torch.nn.Conv1d` and :class:`torch.nn.modules.lazy.LazyModuleMixin`
r r Tr/ N) r r r r r r r! r r$ c
sP | |
d} t jdd|||||d|f | tf | | _|| _|rLtf | | _d S )N)r* r+ r F)r1 r2 r r# r r! ) r& r r r r r r r! r r* r+ r? )rC r% r' r2 s
zLazyConv1d.__init__)r r r r Tr/ NN) rO rP rQ r` rV
cls_to_becomerS r rT r5 r2 rU r% r% )rC r' ry u s ry c s< e Zd ZdZeZd
eeeeeeee dd fdd
Z
Z S )
LazyConv2da A :class:`torch.nn.Conv2d` module with lazy initialization of
the ``in_channels`` argument of the :class:`Conv2d` that is inferred from
the ``input.size(1)``.
The attributes that will be lazily initialized are `weight` and `bias`.
Check the :class:`torch.nn.modules.lazy.LazyModuleMixin` for further documentation
on lazy modules and their limitations.
Args:
out_channels (int): Number of channels produced by the convolution
kernel_size (int or tuple): Size of the convolving kernel
stride (int or tuple, optional): Stride of the convolution. Default: 1
padding (int or tuple, optional): Zero-padding added to both sides of
the input. Default: 0
padding_mode (string, optional): ``'zeros'``, ``'reflect'``,
``'replicate'`` or ``'circular'``. Default: ``'zeros'``
dilation (int or tuple, optional): Spacing between kernel
elements. Default: 1
groups (int, optional): Number of blocked connections from input
channels to output channels. Default: 1
bias (bool, optional): If ``True``, adds a learnable bias to the
output. Default: ``True``
.. seealso:: :class:`torch.nn.Conv2d` and :class:`torch.nn.modules.lazy.LazyModuleMixin`
r r Tr/ N) r r r r r r r! r r$ c
sP | |
d} t jdd|||||d|f | tf | | _|| _|rLtf | | _d S )N)r* r+ r F)r1 r2 r r# r r! ) r& r r r r r r r! r r* r+ r? )rC r% r' r2 s
zLazyConv2d.__init__)r r r r Tr/ NN) rO rP rQ r` ra rz rS r rT r5 r2 rU r% r% )rC r' r{ s r{ c s< e Zd ZdZeZd
eeeeeeee dd fdd
Z
Z S )
LazyConv3da A :class:`torch.nn.Conv3d` module with lazy initialization of
the ``in_channels`` argument of the :class:`Conv3d` that is inferred from
the ``input.size(1)``.
The attributes that will be lazily initialized are `weight` and `bias`.
Check the :class:`torch.nn.modules.lazy.LazyModuleMixin` for further documentation
on lazy modules and their limitations.
Args:
out_channels (int): Number of channels produced by the convolution
kernel_size (int or tuple): Size of the convolving kernel
stride (int or tuple, optional): Stride of the convolution. Default: 1
padding (int or tuple, optional): Zero-padding added to both sides of
the input. Default: 0
padding_mode (string, optional): ``'zeros'``, ``'reflect'``,
``'replicate'`` or ``'circular'``. Default: ``'zeros'``
dilation (int or tuple, optional): Spacing between kernel
elements. Default: 1
groups (int, optional): Number of blocked connections from input
channels to output channels. Default: 1
bias (bool, optional): If ``True``, adds a learnable bias to the
output. Default: ``True``
.. seealso:: :class:`torch.nn.Conv3d` and :class:`torch.nn.modules.lazy.LazyModuleMixin`
r r Tr/ N) r r r r r r r! r r$ c
sP | |
d} t jdd|||||d|f | tf | | _|| _|rLtf | | _d S )N)r* r+ r F)r1 r2 r r# r r! ) r& r r r r r r r! r r* r+ r? )rC r% r' r2 s
zLazyConv3d.__init__)r r r r Tr/ NN) rO rP rQ r` rb rz rS r rT r5 r2 rU r% r% )rC r' r| s r| c s> e Zd ZdZeZd
eeeeeeeee dd
fdd
Z
Z S ) LazyConvTranspose1da' A :class:`torch.nn.ConvTranspose1d` module with lazy initialization of
the ``in_channels`` argument of the :class:`ConvTranspose1d` that is inferred from
the ``input.size(1)``.
The attributes that will be lazily initialized are `weight` and `bias`.
Check the :class:`torch.nn.modules.lazy.LazyModuleMixin` for further documentation
on lazy modules and their limitations.
Args:
out_channels (int): Number of channels produced by the convolution
kernel_size (int or tuple): Size of the convolving kernel
stride (int or tuple, optional): Stride of the convolution. Default: 1
padding (int or tuple, optional): ``dilation * (kernel_size - 1) - padding`` zero-padding
will be added to both sides of the input. Default: 0
output_padding (int or tuple, optional): Additional size added to one side
of the output shape. Default: 0
groups (int, optional): Number of blocked connections from input channels to output channels. Default: 1
bias (bool, optional): If ``True``, adds a learnable bias to the output. Default: ``True``
dilation (int or tuple, optional): Spacing between kernel elements. Default: 1
.. seealso:: :class:`torch.nn.ConvTranspose1d` and :class:`torch.nn.modules.lazy.LazyModuleMixin`
r r Tr/ N)
r r r r r r r! r r r$ c
sR |
| d} t jdd|||||d|| f
| tf | | _|| _|rNtf | | _d S )N)r* r+ r F)r1 r2 r r# r r! )
r& r r r r r r r! r r r* r+ r? )rC r% r' r2 W s"
z LazyConvTranspose1d.__init__) r r r r Tr r/ NN) rO rP rQ r` rk rz rS r rT r5 r2 rU r% r% )rC r' r} ; s r} c s> e Zd ZdZeZd
eeeeeeeee dd
fdd
Z
Z S ) LazyConvTranspose2daK A :class:`torch.nn.ConvTranspose2d` module with lazy initialization of
the ``in_channels`` argument of the :class:`ConvTranspose2d` that is inferred from
the ``input.size(1)``.
The attributes that will be lazily initialized are `weight` and `bias`.
Check the :class:`torch.nn.modules.lazy.LazyModuleMixin` for further documentation
on lazy modules and their limitations.
Args:
out_channels (int): Number of channels produced by the convolution
kernel_size (int or tuple): Size of the convolving kernel
stride (int or tuple, optional): Stride of the convolution. Default: 1
padding (int or tuple, optional): ``dilation * (kernel_size - 1) - padding`` zero-padding
will be added to both sides of each dimension in the input. Default: 0
output_padding (int or tuple, optional): Additional size added to one side
of each dimension in the output shape. Default: 0
groups (int, optional): Number of blocked connections from input channels to output channels. Default: 1
bias (bool, optional): If ``True``, adds a learnable bias to the output. Default: ``True``
dilation (int or tuple, optional): Spacing between kernel elements. Default: 1
.. seealso:: :class:`torch.nn.ConvTranspose2d` and :class:`torch.nn.modules.lazy.LazyModuleMixin`
r r Tr/ N)
r r r r r r r! r r r$ c
sR |
| d} t jdd|||||d|| f
| tf | | _|| _|rNtf | | _d S )N)r* r+ r F)r1 r2 r r# r r! )
r& r r r r r r r! r r r* r+ r? )rC r% r' r2 s"
z LazyConvTranspose2d.__init__) r r r r Tr r/ NN) rO rP rQ r` rn rz rS r rT r5 r2 rU r% r% )rC r' r~ | s r~ c s> e Zd ZdZeZd
eeeeeeeee dd
fdd
Z
Z S ) LazyConvTranspose3daK A :class:`torch.nn.ConvTranspose3d` module with lazy initialization of
the ``in_channels`` argument of the :class:`ConvTranspose3d` that is inferred from
the ``input.size(1)``.
The attributes that will be lazily initialized are `weight` and `bias`.
Check the :class:`torch.nn.modules.lazy.LazyModuleMixin` for further documentation
on lazy modules and their limitations.
Args:
out_channels (int): Number of channels produced by the convolution
kernel_size (int or tuple): Size of the convolving kernel
stride (int or tuple, optional): Stride of the convolution. Default: 1
padding (int or tuple, optional): ``dilation * (kernel_size - 1) - padding`` zero-padding
will be added to both sides of each dimension in the input. Default: 0
output_padding (int or tuple, optional): Additional size added to one side
of each dimension in the output shape. Default: 0
groups (int, optional): Number of blocked connections from input channels to output channels. Default: 1
bias (bool, optional): If ``True``, adds a learnable bias to the output. Default: ``True``
dilation (int or tuple, optional): Spacing between kernel elements. Default: 1
.. seealso:: :class:`torch.nn.ConvTranspose3d` and :class:`torch.nn.modules.lazy.LazyModuleMixin`
r r Tr/ N)
r r r r r r r! r r r$ c
sR |
| d} t jdd|||||d|| f
| tf | | _|| _|rNtf | | _d S )N)r* r+ r F)r1 r2 r r# r r! )
r& r r r r r r r! r r r* r+ r? )rC r% r' r2 s"
z LazyConvTranspose3d.__init__) r r r r Tr r/ NN) rO rP rQ r` ro rz rS r rT r5 r2 rU r% r% )rC r' r s r )0rF rq r< r Ztorch.nn.parameterr r r r\ r Zlazyr moduler
utilsr r r
r Ztorch._torch_docsr common_typesr r r typingr r r r r_ r rV ra rb rc rk rn ro rp ru ry r{ r| r} r~ r r% r% r% r' sH 8 #
&BBBAA