@@ -858,7 +858,7 @@ def convolution(inputs,
858
858
with "NC".
859
859
num_outputs: Integer, the number of output filters.
860
860
kernel_size: A sequence of N positive integers specifying the spatial
861
- dimensions of of the filters. Can be a single integer to specify the same
861
+ dimensions of the filters. Can be a single integer to specify the same
862
862
value for all spatial dimensions.
863
863
stride: A sequence of N positive integers specifying the stride at which to
864
864
compute output. Can be a single integer to specify the same value for all
@@ -1097,7 +1097,7 @@ def convolution2d_transpose(
1097
1097
`[batch, height, width, in_channels]` for `NHWC` data format or
1098
1098
`[batch, in_channels, height, width]` for `NCHW` data format.
1099
1099
num_outputs: Integer, the number of output filters.
1100
- kernel_size: A list of length 2 holding the [kernel_height, kernel_width] of
1100
+ kernel_size: A list of length 2 holding the [kernel_height, kernel_width]
1101
1101
of the filters. Can be an int if both values are the same.
1102
1102
stride: A list of length 2: [stride_height, stride_width].
1103
1103
Can be an int if both strides are the same. Note that presently
@@ -1858,7 +1858,7 @@ def separable_convolution2d(
1858
1858
inputs: A tensor of size [batch_size, height, width, channels].
1859
1859
num_outputs: The number of pointwise convolution output filters. If is
1860
1860
None, then we skip the pointwise convolution stage.
1861
- kernel_size: A list of length 2: [kernel_height, kernel_width] of
1861
+ kernel_size: A list of length 2: [kernel_height, kernel_width]
1862
1862
of the filters. Can be an int if both values are the same.
1863
1863
depth_multiplier: The number of depthwise convolution output channels for
1864
1864
each input channel. The total number of depthwise convolution output
0 commit comments