Pads the 2nd and 3rd dimensions of a 4D tensor. Arguments: x: Tensor or variable. padding: Tuple of 2 tuples, padding pattern. data_format: One of `channels_last` or `channels_first`. Returns: A padded 4D tensor. Raises: ValueError: if `data_format` is neither
(x, padding=((1, 1), (1, 1)), data_format=None)
| 2999 | |
| 3000 | @keras_export('keras.backend.spatial_2d_padding') |
| 3001 | def spatial_2d_padding(x, padding=((1, 1), (1, 1)), data_format=None): |
| 3002 | """Pads the 2nd and 3rd dimensions of a 4D tensor. |
| 3003 | |
| 3004 | Arguments: |
| 3005 | x: Tensor or variable. |
| 3006 | padding: Tuple of 2 tuples, padding pattern. |
| 3007 | data_format: One of `channels_last` or `channels_first`. |
| 3008 | |
| 3009 | Returns: |
| 3010 | A padded 4D tensor. |
| 3011 | |
| 3012 | Raises: |
| 3013 | ValueError: if `data_format` is neither |
| 3014 | `channels_last` or `channels_first`. |
| 3015 | """ |
| 3016 | assert len(padding) == 2 |
| 3017 | assert len(padding[0]) == 2 |
| 3018 | assert len(padding[1]) == 2 |
| 3019 | if data_format is None: |
| 3020 | data_format = image_data_format() |
| 3021 | if data_format not in {'channels_first', 'channels_last'}: |
| 3022 | raise ValueError('Unknown data_format: ' + str(data_format)) |
| 3023 | |
| 3024 | if data_format == 'channels_first': |
| 3025 | pattern = [[0, 0], [0, 0], list(padding[0]), list(padding[1])] |
| 3026 | else: |
| 3027 | pattern = [[0, 0], list(padding[0]), list(padding[1]), [0, 0]] |
| 3028 | return array_ops.pad(x, pattern) |
| 3029 | |
| 3030 | |
| 3031 | @keras_export('keras.backend.spatial_3d_padding') |
nothing calls this directly
no test coverage detected