Apply 2D conv with un-shared weights. Arguments: inputs: 4D tensor with shape: (batch_size, filters, new_rows, new_cols) if data_format='channels_first' or 4D tensor with shape: (batch_size, new_rows, new_cols, filters) if data_format='chann
(inputs,
kernel,
kernel_size,
strides,
output_shape,
data_format=None)
| 5334 | |
| 5335 | @keras_export('keras.backend.local_conv2d') |
| 5336 | def local_conv2d(inputs, |
| 5337 | kernel, |
| 5338 | kernel_size, |
| 5339 | strides, |
| 5340 | output_shape, |
| 5341 | data_format=None): |
| 5342 | """Apply 2D conv with un-shared weights. |
| 5343 | |
| 5344 | Arguments: |
| 5345 | inputs: 4D tensor with shape: |
| 5346 | (batch_size, filters, new_rows, new_cols) |
| 5347 | if data_format='channels_first' |
| 5348 | or 4D tensor with shape: |
| 5349 | (batch_size, new_rows, new_cols, filters) |
| 5350 | if data_format='channels_last'. |
| 5351 | kernel: the unshared weight for convolution, |
| 5352 | with shape (output_items, feature_dim, filters). |
| 5353 | kernel_size: a tuple of 2 integers, specifying the |
| 5354 | width and height of the 2D convolution window. |
| 5355 | strides: a tuple of 2 integers, specifying the strides |
| 5356 | of the convolution along the width and height. |
| 5357 | output_shape: a tuple with (output_row, output_col). |
| 5358 | data_format: the data format, channels_first or channels_last. |
| 5359 | |
| 5360 | Returns: |
| 5361 | A 4D tensor with shape: |
| 5362 | (batch_size, filters, new_rows, new_cols) |
| 5363 | if data_format='channels_first' |
| 5364 | or 4D tensor with shape: |
| 5365 | (batch_size, new_rows, new_cols, filters) |
| 5366 | if data_format='channels_last'. |
| 5367 | """ |
| 5368 | return local_conv(inputs, |
| 5369 | kernel, |
| 5370 | kernel_size, |
| 5371 | strides, |
| 5372 | output_shape, |
| 5373 | data_format) |
| 5374 | |
| 5375 | |
| 5376 | @keras_export('keras.backend.bias_add') |
nothing calls this directly
no test coverage detected