Python gen_nn_ops.conv2d方法代碼示例

# 需要導入模塊: from tensorflow.python.ops import gen_nn_ops [as 別名]
# 或者: from tensorflow.python.ops.gen_nn_ops import conv2d [as 別名]
def _with_space_to_batch_base_paddings(filter_shape, num_spatial_dims,
                                       rate_or_const_rate):
  """Helper function to compute base_paddings."""
  # Spatial dimensions of the filters and the upsampled filters in which we
  # introduce (rate - 1) zeros between consecutive filter values.
  filter_spatial_shape = filter_shape[:num_spatial_dims]
  dilated_filter_spatial_shape = (filter_spatial_shape +
                                  (filter_spatial_shape - 1) *
                                  (rate_or_const_rate - 1))
  pad_extra_shape = dilated_filter_spatial_shape - 1

  # When full_padding_shape is odd, we pad more at end, following the same
  # convention as conv2d.
  pad_extra_start = pad_extra_shape // 2
  pad_extra_end = pad_extra_shape - pad_extra_start
  base_paddings = array_ops.stack([[pad_extra_start[i], pad_extra_end[i]]
                                   for i in range(num_spatial_dims)])
  return base_paddings 

# 需要導入模塊: from tensorflow.python.ops import gen_nn_ops [as 別名]
# 或者: from tensorflow.python.ops.gen_nn_ops import conv2d [as 別名]
def __init__(self,
               input_shape,
               filter_shape,  # pylint: disable=redefined-builtin
               padding, data_format=None,
               strides=None, name=None):
    filter_shape = filter_shape.with_rank(input_shape.ndims)
    self.padding = padding
    self.name = name
    input_shape = input_shape.with_rank(filter_shape.ndims)
    if input_shape.ndims is None:
      raise ValueError("Rank of convolution must be known")
    if input_shape.ndims < 3 or input_shape.ndims > 5:
      raise ValueError(
          "`input` and `filter` must have rank at least 3 and at most 5")
    conv_dims = input_shape.ndims - 2
    if strides is None:
      strides = [1] * conv_dims
    elif len(strides) != conv_dims:
      raise ValueError("len(strides)=%d, but should be %d" %
                       (len(strides), conv_dims))
    if conv_dims == 1:
      # conv1d uses the 2-d data format names
      if data_format is None or data_format == "NWC":
        data_format_2d = "NHWC"
      elif data_format == "NCW":
        data_format_2d = "NCHW"
      else:
        raise ValueError("data_format must be \"NWC\" or \"NCW\".")
      self.strides = strides[0]
      self.data_format = data_format_2d
      self.conv_op = self._conv1d
    elif conv_dims == 2:
      if data_format is None or data_format == "NHWC":
        data_format = "NHWC"
        strides = [1] + list(strides) + [1]
      elif data_format == "NCHW":
        strides = [1, 1] + list(strides)
      else:
        raise ValueError("data_format must be \"NHWC\" or \"NCHW\".")
      self.strides = strides
      self.data_format = data_format
      self.conv_op = gen_nn_ops.conv2d
    elif conv_dims == 3:
      if data_format is None or data_format == "NDHWC":
        strides = [1] + list(strides) + [1]
      elif data_format == "NCDHW":
        strides = [1, 1] + list(strides)
      else:
        raise ValueError("data_format must be \"NDHWC\" or \"NCDHW\". Have: %s"
                         % data_format)
      self.strides = strides
      self.data_format = data_format
      self.conv_op = gen_nn_ops.conv3d

  # Note that we need this adapter since argument names for conv1d don't match
  # those for gen_nn_ops.conv2d and gen_nn_ops.conv3d.
  # pylint: disable=redefined-builtin