The :class:`UnStack` class is a layer for unstacking the given dimension of a rank-R tensor into rank-(R-1) tensors., see `tf.unstack() `__. Parameters ---------- num : int or None The length of the dimension axis. Auto
| 65 | |
| 66 | |
| 67 | class UnStack(Layer): |
| 68 | """ |
| 69 | The :class:`UnStack` class is a layer for unstacking the given dimension of a rank-R tensor into rank-(R-1) tensors., see `tf.unstack() <https://www.tensorflow.org/api_docs/python/tf/unstack>`__. |
| 70 | |
| 71 | Parameters |
| 72 | ---------- |
| 73 | num : int or None |
| 74 | The length of the dimension axis. Automatically inferred if None (the default). |
| 75 | axis : int |
| 76 | Dimension along which axis to concatenate. |
| 77 | name : str |
| 78 | A unique layer name. |
| 79 | |
| 80 | Returns |
| 81 | ------- |
| 82 | list of :class:`Layer` |
| 83 | The list of layer objects unstacked from the input. |
| 84 | |
| 85 | Examples |
| 86 | -------- |
| 87 | >>> ni = Input([4, 10], name='input') |
| 88 | >>> nn = Dense(n_units=5)(ni) |
| 89 | >>> nn = UnStack(axis=1)(nn) # unstack in channel axis |
| 90 | >>> len(nn) # 5 |
| 91 | >>> nn[0].shape # (4,) |
| 92 | |
| 93 | """ |
| 94 | |
| 95 | def __init__(self, num=None, axis=0, name=None): #'unstack'): |
| 96 | super().__init__(name) |
| 97 | self.num = num |
| 98 | self.axis = axis |
| 99 | |
| 100 | self.build(None) |
| 101 | self._built = True |
| 102 | logging.info("UnStack %s: num: %s axis: %d" % (self.name, self.num, self.axis)) |
| 103 | |
| 104 | def __repr__(self): |
| 105 | s = '{classname}(num={num}, axis={axis}' |
| 106 | if self.name is not None: |
| 107 | s += ', name=\'{name}\'' |
| 108 | s += ')' |
| 109 | return s.format(classname=self.__class__.__name__, **self.__dict__) |
| 110 | |
| 111 | def build(self, inputs_shape): |
| 112 | pass |
| 113 | |
| 114 | def forward(self, inputs): |
| 115 | outputs = tf.unstack(inputs, num=self.num, axis=self.axis, name=self.name) |
| 116 | return outputs |
no outgoing calls
searching dependent graphs…