ёia:SSKJr SSKJrJr SSKrSSKJr SSK J r SSK J r SSK Jr SS KJr \(a SS KJr SS KJr SS KJrJrJrJrJr SS KJr /r"SS\5r"SS\5r "SS\5r!"SS\ 5r""SS\5r#"SS\ 5r$g)) annotations) TYPE_CHECKINGLiteralN)Layer)_update_padding_nd)Normal)convert_to_list) functional)Sequence)Tensor) ParamAttrLikeSize2Size3Size4Size6)_PaddingSizeModec^\rSrSr%S\S'S\S'S S U4SjjjrS SjrS SjrSrU=r $) _Conv3D,r weightbiasc>^[TT]5 U SLdS5eU TlU TlUTlUTlUTlU TlUTlU Tl UTl U S:XdS5eUS:XdS5eUS;dS5eS 1nX;a[S US U S 35eU S :HnS n[UUS5Tl [UUS5Tl[UUS5TlUTlU Tl[%UUU5uTlTl/TRQTR PTR PnU4SjnTR+UTRU"5S9TlTR+TRTR /SS9Tlg)NF(weight_attr should not be False in Conv.zeros,Currently, only support padding_mode='zeros' Currently, only support groups=1Nigemm;The value of 'backend' in Conv3D should be None or 'igemm'.NDHWCdata_format must be one of , but got data_format=''stridedilation kernel_sizec>[R"TR5TR-nSU- S-n[ SU5$Ng@g?gnpprod _kernel_size _in_channelsrfilter_elem_numstdselfs [/var/www/html/banglarbhumi/venv/lib/python3.13/site-packages/paddle/sparse/nn/layer/conv.py_get_default_param_initializer8_Conv3D.__init__.._get_default_param_initializerq> ggd&7&784;L;LLO(S0C#s# #shapeattrdefault_initializerTr<r;is_biassuper__init__ _param_attr _bias_attr_groupsr0 _out_channels _data_format_subm_key_backend ValueErrorr _stride _dilationr/_padding _padding_moder_updated_padding_padding_algorithmcreate_parameterrrr4 in_channels out_channelsr)r'paddingr(groupssubmkey padding_mode weight_attr bias_attr data_formatbackend valid_format channel_lastdims filter_shaper6 __class__s` r5rB_Conv3D.__init__0s" %' 6 ''# ')'   w& : &{>>>{   I I I  y  *-l^;RS^R__`a #g- &vtX> (4D+K}M )9K \4: 6t6             $ ++!! > @,  ))););(>aS3t~~#66 6 0 0H <<1  , ,H22///r9rJrDrGrMrEr0r/rIrFrNrQrOrCrLrHrPrr) rrrrFNrNNr"N)rTintrUrr)rr'rrV2_PaddingSizeMode | Size3 | Size6 | Sequence[Size2]r(rrW Literal[1]rXboolrY str | NonerZLiteral['zeros']r[ParamAttrLike | Noner\rr]Literal['NDHWC']r^Literal['igemm'] | NonereturnNonermr rr rstr __name__ __module__ __qualname____firstlineno____annotations__rBror{__static_attributes__ __classcell__rcs@r5rr,s N LFG)0,0*.(/+/M M M  M  M D M M M M M 'M *M (M &M )M !M M ^B 0 0r9rc^\rSrSr%S\S'S\S'S S U4SjjjrS SjrS SjrSrU=r $) _Conv2Dr rrc>^[TT]5 U SLdS5eU TlU TlUTlUTlUTlU TlUTlU Tl UTl U S:XdS5eUS:XdS5eUS;dS5eS 1nX;a[S US U S 35eU S :HnS n[UUS5Tl [UUS5Tl[UUS5TlUTlU Tl[%UUU5uTlTl/TRQTR PTR PnU4SjnTR+UTRU"5S9TlTR+TRTR /SS9Tlg)NFrrrrrrr!NHWCr#r$r%r r'r(r)c>[R"TR5TR-nSU- S-n[ SU5$r+r,r1s r5r68_Conv2D.__init__.._get_default_param_initializerr8r9r:Tr>r@rSs` r5rB_Conv2D.__init__s" %' 6 ''# ')'   w& : &{>>>{   I I I x  *-l^;RS^R__`a #f, &vtX> (4D+K}M )9K \4: 6t6             $ ++!! > @,  ))););(>> import paddle >>> indices = [ ... [0, 0, 0, 0], ... [0, 0, 0, 0], ... [0, 0, 1, 2], ... [1, 3, 2, 3], ... ] >>> values = [[1], [2], [3], [4]] >>> indices = paddle.to_tensor(indices, dtype='int32') >>> values = paddle.to_tensor(values, dtype='float32') >>> dense_shape = [1, 1, 3, 4, 1] >>> sparse_x = paddle.sparse.sparse_coo_tensor(indices, values, dense_shape, stop_gradient=True) >>> conv = paddle.sparse.nn.Conv3D(1, 1, (1, 3, 3)) >>> y = conv(sparse_x) >>> print(y.shape) paddle.Size([1, 1, 1, 2, 1]) c 8>[T U]UUUUUUUSSUU U U S9 gNF) r'rVr(rWrXrYrZr[r\r]rArB r4rTrUr)r'rVr(rWrZr[r\r]rcs r5rBConv3D.__init__>    %##  r9rt)rrrrrNNr")rTrrUrr)rr'rrVrr(rrWrrZrr[rr\rr]rrrrrrr__doc__rBrrrs@r5rr4sjbFG)0,0*.(/       D    ' * ( &    r9rcr^\rSrSrSrSSU4SjjjrSrU=r$)Conv2Dia **Sparse Convolution2d Layer** The Sparse convolution2d layer calculates the output based on the input, filter and strides, paddings, dilations, groups parameters. Input(Input) and Output(Output) are multidimensional SparseCooTensors with a shape of :math:`[N, H, W, C]` . Where N is batch size, C is the number of channels, H is the height of the feature, and W is the width of the feature. If bias attribution is provided, bias is added to the output of the convolution. For each input :math:`X`, the equation is: .. math:: Out = W \ast X + b In the above equation: * :math:`X`: Input value, a tensor with NHWC format. * :math:`W`: Filter value, a tensor with HWCM format. * :math:`\\ast`: Convolution operation. * :math:`b`: Bias value, a 1-D tensor with shape [M]. * :math:`Out`: Output value, the shape of :math:`Out` and :math:`X` may be different. Parameters: in_channels(int): The number of input channels in the input image. out_channels(int): The number of output channels produced by the convolution. kernel_size(int|list|tuple): The size of the convolving kernel. stride(int|list|tuple, optional): The stride size. If stride is a list/tuple, it must contain three integers, (stride_H, stride_W). Otherwise, the stride_H = stride_W = stride. The default value is 1. padding(int|str|tuple|list, optional): The padding size. Padding couple be in one of the following forms. 1. a string in ['valid', 'same']. 2. an int, which means each spatial dimension(height, width) is zero padded by size of `padding` 3. a list[int] or tuple[int] whose length is the number of spatial dimensions, which contains the amount of padding on each side for each spatial dimension. It has the form [pad_d1, pad_d2, ...]. 4. a list[int] or tuple[int] whose length is 2 * number of spatial dimensions. It has the form [pad_before, pad_after, pad_before, pad_after, ...] for all spatial dimensions. 5. a list or tuple of pairs of ints. It has the form [[pad_before, pad_after], [pad_before, pad_after], ...]. Note that, the batch dimension and channel dimension are also included. Each pair of integers correspond to the amount of padding for a dimension of the input. Padding in batch dimension and channel dimension should be [0, 0] or (0, 0). The default value is 0. dilation(int|list|tuple, optional): The dilation size. If dilation is a list/tuple, it must contain three integers, (dilation_H, dilation_W). Otherwise, the dilation_H = dilation_W = dilation. The default value is 1. groups(int, optional): The groups number of the Conv2D Layer. According to grouped convolution in Alex Krizhevsky's Deep CNN paper: when group=2, the first half of the filters is only connected to the first half of the input channels, while the second half of the filters is only connected to the second half of the input channels. The default value is 1, currently, only support groups=1. padding_mode(str, optional): ``'zeros'``, ``'reflect'``, ``'replicate'`` or ``'circular'``. Currently only support ``'zeros'``. weight_attr(ParamAttr, optional): The parameter attribute for learnable parameters/weights of conv2d. If it is set to None or one attribute of ParamAttr, conv2d will create ParamAttr as param_attr. If it is set to None, the parameter is initialized with :math:`Normal(0.0, std)`, and the :math:`std` is :math:`(\frac{2.0 }{filter\_elem\_num})^{0.5}`. The default value is None. bias_attr(ParamAttr|bool, optional): The parameter attribute for the bias of conv2d. If it is set to False, no bias will be added to the output units. If it is set to None or one attribute of ParamAttr, conv2d will create ParamAttr as bias_attr. If the Initializer of the bias_attr is not set, the bias is initialized zero. The default value is None. data_format(str, optional): Data format that specifies the layout of input. It can be "NCHW" or "NHWC". Currently, only support "NHWC". Attribute: **weight** (Parameter): the learnable weights of filters of this layer. **bias** (Parameter): the learnable bias of this layer. Shape: - x: :math:`(N, H_{in}, W_{in}, C_{in})` - weight: :math:`(K_{h}, K_{w}, C_{in}, C_{out})` - bias: :math:`(C_{out})` - output: :math:`(N, H_{out}, W_{out}, C_{out})` Where .. math:: H_{out}&= \frac{(H_{in} + 2 * paddings[0] - (dilations[0] * (kernel\_size[0] - 1) + 1))}{strides[0]} + 1 W_{out}&= \frac{(W_{in} + 2 * paddings[1] - (dilations[1] * (kernel\_size[1] - 1) + 1))}{strides[1]} + 1 Examples: .. code-block:: pycon >>> import paddle >>> indices = [[0, 0, 0, 0], [0, 0, 1, 2], [1, 3, 2, 3]] >>> values = [[1], [2], [3], [4]] >>> indices = paddle.to_tensor(indices, dtype='int32') >>> values = paddle.to_tensor(values, dtype='float32') >>> dense_shape = [1, 3, 4, 1] >>> sparse_x = paddle.sparse.sparse_coo_tensor(indices, values, dense_shape, stop_gradient=True) >>> conv = paddle.sparse.nn.Conv2D(1, 1, (3, 3)) >>> y = conv(sparse_x) >>> print(y.shape) paddle.Size([1, 1, 2, 1]) c 8>[T U]UUUUUUUSSUU U U S9 grrrs r5rBConv2D.__init__(rr9rt)rrrrrNNr)rTrrUrr)rr'rrVrr(rrWrrZrr[rr\rr]rrrrrs@r5rrseXFG)0,0*.'-       D    ' * ( %    r9rc~^\rSrSrSrSSU4SjjjrSrU=r$) SubmConv3DiGa1 **Submanifold Sparse Convolution3d Layer** The submanifold sparse convolution3d layer calculates the output based on the input, filter and strides, paddings, dilations, groups parameters. Input(Input) and Output(Output) are multidimensional SparseCooTensors with a shape of :math:`[N, D, H, W, C]` . Where N is batch size, C is the number of channels, D is the depth of the feature, H is the height of the feature, and W is the width of the feature. If bias attribution is provided, bias is added to the output of the convolution. For each input :math:`X`, the equation is: .. math:: Out = W \ast X + b In the above equation: * :math:`X`: Input value, a tensor with NDHWC format. * :math:`W`: Filter value, a tensor with DHWCM format. * :math:`\\ast`: Submanifold Convolution operation, refer to the paper: https://arxiv.org/abs/1706.01307. * :math:`b`: Bias value, a 1-D tensor with shape [M]. * :math:`Out`: Output value, the shape of :math:`Out` and :math:`X` may be different. Parameters: in_channels(int): The number of input channels in the input image. out_channels(int): The number of output channels produced by the convolution. kernel_size(int|list|tuple): The size of the convolving kernel. stride(int|list|tuple, optional): The stride size. If stride is a list/tuple, it must contain three integers, (stride_D, stride_H, stride_W). Otherwise, the stride_D = stride_H = stride_W = stride. The default value is 1. padding(int|str|tuple|list, optional): The padding size. Padding couple be in one of the following forms. 1. a string in ['valid', 'same']. 2. an int, which means each spatial dimension(depth, height, width) is zero padded by size of `padding` 3. a list[int] or tuple[int] whose length is the number of spatial dimensions, which contains the amount of padding on each side for each spatial dimension. It has the form [pad_d1, pad_d2, ...]. 4. a list[int] or tuple[int] whose length is 2 * number of spatial dimensions. It has the form [pad_before, pad_after, pad_before, pad_after, ...] for all spatial dimensions. 5. a list or tuple of pairs of ints. It has the form [[pad_before, pad_after], [pad_before, pad_after], ...]. Note that, the batch dimension and channel dimension are also included. Each pair of integers correspond to the amount of padding for a dimension of the input. Padding in batch dimension and channel dimension should be [0, 0] or (0, 0). The default value is 0. dilation(int|list|tuple, optional): The dilation size. If dilation is a list/tuple, it must contain three integers, (dilation_D, dilation_H, dilation_W). Otherwise, the dilation_D = dilation_H = dilation_W = dilation. The default value is 1. groups(int, optional): The groups number of the Conv3D Layer. According to grouped convolution in Alex Krizhevsky's Deep CNN paper: when group=2, the first half of the filters is only connected to the first half of the input channels, while the second half of the filters is only connected to the second half of the input channels. The default value is 1. padding_mode(str, optional): ``'zeros'``, ``'reflect'``, ``'replicate'`` or ``'circular'``. Currently only support ``'zeros'``. key(str, optional): the key is used to save or use the same rulebook, the definition and role of rulebook refers to https://pdfs.semanticscholar.org/5125/a16039cabc6320c908a4764f32596e018ad3.pdf. The default value is None. weight_attr(ParamAttr, optional): The parameter attribute for learnable parameters/weights of conv3d. If it is set to None or one attribute of ParamAttr, conv3d will create ParamAttr as param_attr. If it is set to None, the parameter is initialized with :math:`Normal(0.0, std)`, and the :math:`std` is :math:`(\frac{2.0 }{filter\_elem\_num})^{0.5}`. The default value is None. bias_attr(ParamAttr|bool, optional): The parameter attribute for the bias of conv3d. If it is set to False, no bias will be added to the output units. If it is set to None or one attribute of ParamAttr, conv3d will create ParamAttr as bias_attr. If the Initializer of the bias_attr is not set, the bias is initialized zero. The default value is None. data_format(str, optional): Data format that specifies the layout of input. It can be "NCDHW" or "NDHWC". Currently, only support "NCDHW". Attribute: **weight** (Parameter): the learnable weights of filters of this layer. **bias** (Parameter): the learnable bias of this layer. Shape: - x: :math:`(N, D_{in}, H_{in}, W_{in}, C_{in})` - weight: :math:`(K_{d}, K_{h}, K_{w}, C_{in}, C_{out})` - bias: :math:`(C_{out})` - output: :math:`(N, D_{out}, H_{out}, W_{out}, C_{out})` Where .. math:: D_{out}&= \frac{(D_{in} + 2 * paddings[0] - (dilations[0] * (kernel\_size[0] - 1) + 1))}{strides[0]} + 1 H_{out}&= \frac{(H_{in} + 2 * paddings[1] - (dilations[1] * (kernel\_size[1] - 1) + 1))}{strides[1]} + 1 W_{out}&= \frac{(W_{in} + 2 * paddings[2] - (dilations[2] * (kernel\_size[2] - 1) + 1))}{strides[2]} + 1 Examples: .. code-block:: pycon >>> import paddle >>> indices = [ ... [0, 0, 0, 0], ... [0, 0, 0, 0], ... [0, 0, 1, 2], ... [1, 3, 2, 3], ... ] >>> values = [[1], [2], [3], [4]] >>> dense_shape = [1, 1, 3, 4, 1] >>> indices = paddle.to_tensor(indices, dtype='int32') >>> values = paddle.to_tensor(values, dtype='float32') >>> sparse_x = paddle.sparse.sparse_coo_tensor(indices, values, dense_shape, stop_gradient=True) >>> subm_conv = paddle.sparse.nn.SubmConv3D(1, 1, (1, 3, 3)) >>> y = subm_conv(sparse_x) >>> print(y.shape) paddle.Size([1, 1, 3, 4, 1]) c:>[TU]UUUUUUUSU UU U U U S9 gNT) r'rVr(rWrXrYrZr[r\r]r^rr4rTrUr)r'rVr(rWrZrYr[r\r]r^rcs r5rBSubmConv3D.__init__A    %##  r9rt) rrrrrNNNr"N)rTrrUrr)rr'rrVrr(rrWrrZrrYrr[rr\rr]rr^rrrrrs@r5rrGsnjFG)0,0*.(/+/       D    '  * ( & )    r9rc~^\rSrSrSrSSU4SjjjrSrU=r$) SubmConv2Dia **Submanifold Sparse Convolution2d Layer** The submanifold sparse convolution2d layer calculates the output based on the input, filter and strides, paddings, dilations, groups parameters. Input(Input) and Output(Output) are multidimensional SparseCooTensors with a shape of :math:`[N, H, W, C]` . Where N is batch size, C is the number of channels, H is the height of the feature, and W is the width of the feature. If bias attribution is provided, bias is added to the output of the convolution. For each input :math:`X`, the equation is: .. math:: Out = W \ast X + b In the above equation: * :math:`X`: Input value, a tensor with NDHWC format. * :math:`W`: Filter value, a tensor with DHWCM format. * :math:`\\ast`: Submanifold Convolution operation, refer to the paper: https://arxiv.org/abs/1706.01307. * :math:`b`: Bias value, a 1-D tensor with shape [M]. * :math:`Out`: Output value, the shape of :math:`Out` and :math:`X` may be different. Parameters: in_channels(int): The number of input channels in the input image. out_channels(int): The number of output channels produced by the convolution. kernel_size(int|list|tuple): The size of the convolving kernel. stride(int|list|tuple, optional): The stride size. If stride is a list/tuple, it must contain two integers, (stride_H, stride_W). Otherwise, the stride_H = stride_W = stride. The default value is 1. padding(int|str|tuple|list, optional): The padding size. Padding couple be in one of the following forms. 1. a string in ['valid', 'same']. 2. an int, which means each spatial dimension(depth, height, width) is zero padded by size of `padding` 3. a list[int] or tuple[int] whose length is the number of spatial dimensions, which contains the amount of padding on each side for each spatial dimension. It has the form [pad_d1, pad_d2, ...]. 4. a list[int] or tuple[int] whose length is 2 * number of spatial dimensions. It has the form [pad_before, pad_after, pad_before, pad_after, ...] for all spatial dimensions. 5. a list or tuple of pairs of ints. It has the form [[pad_before, pad_after], [pad_before, pad_after], ...]. Note that, the batch dimension and channel dimension are also included. Each pair of integers correspond to the amount of padding for a dimension of the input. Padding in batch dimension and channel dimension should be [0, 0] or (0, 0). The default value is 0. dilation(int|list|tuple, optional): The dilation size. If dilation is a list/tuple, it must contain two integers, (dilation_H, dilation_W). Otherwise, the dilation_H = dilation_W = dilation. The default value is 1. groups(int, optional): The groups number of the Conv2D Layer. According to grouped convolution in Alex Krizhevsky's Deep CNN paper: when group=2, the first half of the filters is only connected to the first half of the input channels, while the second half of the filters is only connected to the second half of the input channels. The default value is 1. padding_mode(str, optional): ``'zeros'``, ``'reflect'``, ``'replicate'`` or ``'circular'``. Currently only support ``'zeros'``. key(str, optional): the key is used to save or use the same rulebook, the definition and role of rulebook refers to https://pdfs.semanticscholar.org/5125/a16039cabc6320c908a4764f32596e018ad3.pdf. The default value is None. weight_attr(ParamAttr, optional): The parameter attribute for learnable parameters/weights of conv2d. If it is set to None or one attribute of ParamAttr, conv2d will create ParamAttr as param_attr. If it is set to None, the parameter is initialized with :math:`Normal(0.0, std)`, and the :math:`std` is :math:`(\frac{2.0 }{filter\_elem\_num})^{0.5}`. The default value is None. bias_attr(ParamAttr|bool, optional): The parameter attribute for the bias of conv2d. If it is set to False, no bias will be added to the output units. If it is set to None or one attribute of ParamAttr, conv2d will create ParamAttr as bias_attr. If the Initializer of the bias_attr is not set, the bias is initialized zero. The default value is None. data_format(str, optional): Data format that specifies the layout of input. It can be "NCHW" or "NHWC". Currently, only support "NHWC". Attribute: **weight** (Parameter): the learnable weights of filters of this layer. **bias** (Parameter): the learnable bias of this layer. Shape: - x: :math:`(N, H_{in}, W_{in}, C_{in})` - weight: :math:`(K_{h}, K_{w}, C_{in}, C_{out})` - bias: :math:`(C_{out})` - output: :math:`(N, H_{out}, W_{out}, C_{out})` Where .. math:: H_{out}&= \frac{(H_{in} + 2 * paddings[0] - (dilations[0] * (kernel\_size[0] - 1) + 1))}{strides[0]} + 1 W_{out}&= \frac{(W_{in} + 2 * paddings[1] - (dilations[1] * (kernel\_size[1] - 1) + 1))}{strides[1]} + 1 Examples: .. code-block:: pycon >>> import paddle >>> indices = [[0, 0, 0, 0], [0, 0, 1, 2], [1, 3, 2, 3]] >>> values = [[1], [2], [3], [4]] >>> dense_shape = [1, 3, 4, 1] >>> indices = paddle.to_tensor(indices, dtype='int32') >>> values = paddle.to_tensor(values, dtype='float32') >>> sparse_x = paddle.sparse.sparse_coo_tensor(indices, values, dense_shape, stop_gradient=True) >>> subm_conv = paddle.sparse.nn.SubmConv2D(1, 1, (3, 3)) >>> y = subm_conv(sparse_x) >>> print(y.shape) paddle.Size([1, 3, 4, 1]) c:>[TU]UUUUUUUSU UU U U U S9 grrrs r5rBSubmConv2D.__init__Frr9rt) rrrrrNNNrN)rTrrUrr)rr'rrVrr(rrWrrZrrYrr[rr\rr]rr^rrrrrs@r5rrsi`FG)0,0*.'-+/       D    '  * ( % )    r9r)% __future__rtypingrrnumpyr- paddle.nnrpaddle.nn.functional.convrpaddle.nn.initializerr paddle.utilsr r rhcollections.abcr paddler paddle._typingrrrrrpaddle.nn.functional.commonr__all__rrrrrrrtr9r5rs#)8(((= A0eA0HA0eA0HI WI XD WD NP P fK K r9