diff --git a/mindspore/nn/layer/conv.py b/mindspore/nn/layer/conv.py index 389429bf05..13f3ee4c23 100644 --- a/mindspore/nn/layer/conv.py +++ b/mindspore/nn/layer/conv.py @@ -21,7 +21,7 @@ from mindspore.ops.primitive import constexpr from mindspore.common.parameter import Parameter from mindspore.common.initializer import initializer from mindspore.common.tensor import Tensor -from mindspore._checkparam import Rel +from mindspore._checkparam import ParamValidator as validator, Rel from mindspore._checkparam import Validator from mindspore._checkparam import check_bool, twice, check_int_positive from mindspore._extends import cell_attr_register @@ -29,10 +29,12 @@ from ..cell import Cell __all__ = ['Conv2d', 'Conv2dTranspose', 'DepthwiseConv2d', 'Conv1d', 'Conv1dTranspose'] + class _Conv(Cell): """ Applies a N-D convolution over an input signal composed of several input planes. """ + def __init__(self, in_channels, out_channels, @@ -68,16 +70,16 @@ class _Conv(Cell): self.group = check_int_positive(group) self.has_bias = has_bias if (not isinstance(kernel_size[0], int)) or (not isinstance(kernel_size[1], int)) or \ - isinstance(kernel_size[0], bool) or isinstance(kernel_size[1], bool) or \ - kernel_size[0] < 1 or kernel_size[1] < 1: + isinstance(kernel_size[0], bool) or isinstance(kernel_size[1], bool) or \ + kernel_size[0] < 1 or kernel_size[1] < 1: raise ValueError("Attr 'kernel_size' of 'Conv2D' Op passed " + str(self.kernel_size) + ", should be a int or tuple and equal to or greater than 1.") if (not isinstance(stride[0], int)) or (not isinstance(stride[1], int)) or \ - isinstance(stride[0], bool) or isinstance(stride[1], bool) or stride[0] < 1 or stride[1] < 1: + isinstance(stride[0], bool) or isinstance(stride[1], bool) or stride[0] < 1 or stride[1] < 1: raise ValueError("Attr 'stride' of 'Conv2D' Op passed " + str(self.stride) + ", should be a int or tuple and equal to or greater than 1.") if (not isinstance(dilation[0], int)) or (not isinstance(dilation[1], int)) or \ - isinstance(dilation[0], bool) or isinstance(dilation[1], bool) or dilation[0] < 1 or dilation[1] < 1: + isinstance(dilation[0], bool) or isinstance(dilation[1], bool) or dilation[0] < 1 or dilation[1] < 1: raise ValueError("Attr 'dilation' of 'Conv2D' Op passed " + str(self.dilation) + ", should be a int or tuple and equal to or greater than 1.") if in_channels % group != 0: @@ -193,6 +195,7 @@ class Conv2d(_Conv): >>> net(input).shape (1, 240, 1024, 640) """ + @cell_attr_register def __init__(self, in_channels, @@ -264,6 +267,7 @@ def _check_input_3d(input_shape): if len(input_shape) != 3: raise ValueError(f"Input should be 3d, but got shape {input_shape}") + class Conv1d(_Conv): r""" 1D convolution layer. @@ -344,6 +348,7 @@ class Conv1d(_Conv): >>> net(input).shape (1, 240, 640) """ + @cell_attr_register def __init__(self, in_channels, @@ -498,6 +503,7 @@ class Conv2dTranspose(_Conv): >>> input = Tensor(np.ones([1, 3, 16, 50]), mindspore.float32) >>> net(input) """ + def __init__(self, in_channels, out_channels, @@ -662,6 +668,7 @@ class Conv1dTranspose(_Conv): >>> input = Tensor(np.ones([1, 3, 50]), mindspore.float32) >>> net(input) """ + def __init__(self, in_channels, out_channels, @@ -809,7 +816,8 @@ class DepthwiseConv2d(Cell): filter and :math:`out_{j}` corresponds to the :math:`j`-th channel of the output. :math:`W_{ij}` is a slice of kernel and it has shape :math:`(\text{ks_h}, \text{ks_w})`, where :math:`\text{ks_h}` and :math:`\text{ks_w}` are the height and width of the convolution kernel. The full kernel has shape - :math:`(C_{out}, C_{in}, \text{ks_h}, \text{ks_w})` to split the input in the channel dimension. + :math:`(C_{out}, C_{in} // \text{group}, \text{ks_h}, \text{ks_w})`, where group is the group number + to split the input in the channel dimension. If the 'pad_mode' is set to be "valid", the output height and width will be :math:`\left \lfloor{1 + \frac{H_{in} + 2 \times \text{padding} - \text{ks_h} - @@ -855,6 +863,8 @@ class DepthwiseConv2d(Cell): be :math:`k - 1` pixels skipped for each sampling location. Its value should be greater than or equal to 1 and bounded by the height and width of the input. Default: 1. + group (int): Split filter into groups, `in_ channels` and `out_channels` should be + divisible by the number of groups. Default: 1. has_bias (bool): Specifies whether the layer uses a bias vector. Default: False. weight_init (Union[Tensor, str, Initializer, numbers.Number]): Initializer for the convolution kernel. It can be a Tensor, a string, an Initializer or a number. When a string is specified, @@ -878,6 +888,7 @@ class DepthwiseConv2d(Cell): >>> net(input).shape (1, 240, 1024, 640) """ + def __init__(self, in_channels, out_channels, @@ -886,6 +897,7 @@ class DepthwiseConv2d(Cell): pad_mode='same', padding=0, dilation=1, + group=1, has_bias=False, weight_init='normal', bias_init='zeros'): @@ -895,8 +907,12 @@ class DepthwiseConv2d(Cell): self.dilation = twice(dilation) self.in_channels = check_int_positive(in_channels) self.out_channels = check_int_positive(out_channels) + validator.check_integer('group', group, in_channels, Rel.EQ) + validator.check_integer('group', group, out_channels, Rel.EQ) + validator.check_integer('group', group, 1, Rel.GE) self.pad_mode = pad_mode self.dilation = dilation + self.group = group self.has_bias = has_bias self.weight_init = weight_init self.bias_init = bias_init @@ -928,10 +944,10 @@ class DepthwiseConv2d(Cell): def extend_repr(self): s = 'input_channels={}, output_channels={}, kernel_size={}, stride={}, ' \ - 'pad_mode={}, padding={}, dilation={}' \ + 'pad_mode={}, padding={}, dilation={}, group={}, ' \ 'has_bias={}, weight_init={}, bias_init={}'.format( self.in_channels, self.out_channels, self.kernel_size, self.stride, - self.pad_mode, self.padding, self.dilation, + self.pad_mode, self.padding, self.dilation, self.group, self.has_bias, self.weight_init, self.bias_init) if self.has_bias: