tf.compat.v1.nn.conv2d_backprop_input

View source on GitHub

Computes the gradients of convolution with respect to the input.

input_sizes A Tensor of type int32. An integer vector representing the shape of input, where input is a 4-D [batch, height, width, channels] tensor.
filter A Tensor. Must be one of the following types: half, bfloat16, float32, float64. 4-D with shape [filter_height, filter_width, in_channels, out_channels].
out_backprop A Tensor. Must have the same type as filter. 4-D with shape [batch, out_height, out_width, out_channels]. Gradients w.r.t. the output of the convolution.
strides A list of ints. The stride of the sliding window for each dimension of the input of the convolution. Must be in the same order as the dimension specified with format.
padding Either the string"SAME"or"VALID"indicating the type of padding algorithm to use, or a list indicating the explicit paddings at the start and end of each dimension. When explicit padding is used and data_format is"NHWC", this should be in the form[[0, 0], [pad_top, pad_bottom], [pad_left, pad_right], [0, 0]]. When explicit padding used and data_format is"NCHW", this should be in the form[[0, 0], [0, 0], [pad_top, pad_bottom], [pad_left, pad_right]]. </td> </tr><tr> <td>use_cudnn_on_gpu</td> <td> An optionalbool. Defaults toTrue. </td> </tr><tr> <td>data_format</td> <td> An optionalstringfrom:"NHWC", "NCHW". Defaults to"NHWC". Specify the data format of the input and output data. With the default format "NHWC", the data is stored in the order of: [batch, in_height, in_width, in_channels]. Alternatively, the format could be "NCHW", the data storage order of: [batch, in_channels, in_height, in_width]. </td> </tr><tr> <td>dilations</td> <td> An optional list ofints. Defaults to[1, 1, 1, 1]. 1-D tensor of length 4. The dilation factor for each dimension ofinput. If set to k > 1, there will be k-1 skipped cells between each filter element on that dimension. The dimension order is determined by the value ofdata_format, see above for details. Dilations in the batch and depth dimensions must be 1. </td> </tr><tr> <td>name</td> <td> A name for the operation (optional). </td> </tr><tr> <td>filters` Alias for filter.

A Tensor. Has the same type as filter.