Coverage for /pythoncovmergedfiles/medio/medio/usr/local/lib/python3.8/site-packages/keras/src/layers/pooling/base_pooling3d.py: 23%
47 statements
« prev ^ index » next coverage.py v7.4.0, created at 2024-01-03 07:57 +0000
« prev ^ index » next coverage.py v7.4.0, created at 2024-01-03 07:57 +0000
1# Copyright 2015 The TensorFlow Authors. All Rights Reserved.
2#
3# Licensed under the Apache License, Version 2.0 (the "License");
4# you may not use this file except in compliance with the License.
5# You may obtain a copy of the License at
6#
7# http://www.apache.org/licenses/LICENSE-2.0
8#
9# Unless required by applicable law or agreed to in writing, software
10# distributed under the License is distributed on an "AS IS" BASIS,
11# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
12# See the License for the specific language governing permissions and
13# limitations under the License.
14# ==============================================================================
15"""Private base class for pooling 3D layers."""
18import tensorflow.compat.v2 as tf
20from keras.src import backend
21from keras.src.engine.base_layer import Layer
22from keras.src.engine.input_spec import InputSpec
23from keras.src.utils import conv_utils
26class Pooling3D(Layer):
27 """Pooling layer for arbitrary pooling functions, for 3D inputs.
29 This class only exists for code reuse. It will never be an exposed API.
31 Args:
32 pool_function: The pooling function to apply, e.g. `tf.nn.max_pool2d`.
33 pool_size: An integer or tuple/list of 3 integers:
34 (pool_depth, pool_height, pool_width)
35 specifying the size of the pooling window.
36 Can be a single integer to specify the same value for
37 all spatial dimensions.
38 strides: An integer or tuple/list of 3 integers,
39 specifying the strides of the pooling operation.
40 Can be a single integer to specify the same value for
41 all spatial dimensions.
42 padding: A string. The padding method, either 'valid' or 'same'.
43 Case-insensitive.
44 data_format: A string, one of `channels_last` (default) or
45 `channels_first`.
46 The ordering of the dimensions in the inputs.
47 `channels_last` corresponds to inputs with shape
48 `(batch, depth, height, width, channels)`
49 while `channels_first` corresponds to
50 inputs with shape `(batch, channels, depth, height, width)`.
51 name: A string, the name of the layer.
52 """
54 def __init__(
55 self,
56 pool_function,
57 pool_size,
58 strides,
59 padding="valid",
60 data_format="channels_last",
61 name=None,
62 **kwargs
63 ):
64 super().__init__(name=name, **kwargs)
65 if data_format is None:
66 data_format = backend.image_data_format()
67 if strides is None:
68 strides = pool_size
69 self.pool_function = pool_function
70 self.pool_size = conv_utils.normalize_tuple(pool_size, 3, "pool_size")
71 self.strides = conv_utils.normalize_tuple(
72 strides, 3, "strides", allow_zero=True
73 )
74 self.padding = conv_utils.normalize_padding(padding)
75 self.data_format = conv_utils.normalize_data_format(data_format)
76 self.input_spec = InputSpec(ndim=5)
78 def call(self, inputs):
79 pool_shape = (1,) + self.pool_size + (1,)
80 strides = (1,) + self.strides + (1,)
82 if self.data_format == "channels_first":
83 # TF does not support `channels_first` with 3D pooling operations,
84 # so we must handle this case manually.
85 # TODO(fchollet): remove this when TF pooling is feature-complete.
86 inputs = tf.transpose(inputs, (0, 2, 3, 4, 1))
88 outputs = self.pool_function(
89 inputs,
90 ksize=pool_shape,
91 strides=strides,
92 padding=self.padding.upper(),
93 )
95 if self.data_format == "channels_first":
96 outputs = tf.transpose(outputs, (0, 4, 1, 2, 3))
97 return outputs
99 def compute_output_shape(self, input_shape):
100 input_shape = tf.TensorShape(input_shape).as_list()
101 if self.data_format == "channels_first":
102 len_dim1 = input_shape[2]
103 len_dim2 = input_shape[3]
104 len_dim3 = input_shape[4]
105 else:
106 len_dim1 = input_shape[1]
107 len_dim2 = input_shape[2]
108 len_dim3 = input_shape[3]
109 len_dim1 = conv_utils.conv_output_length(
110 len_dim1, self.pool_size[0], self.padding, self.strides[0]
111 )
112 len_dim2 = conv_utils.conv_output_length(
113 len_dim2, self.pool_size[1], self.padding, self.strides[1]
114 )
115 len_dim3 = conv_utils.conv_output_length(
116 len_dim3, self.pool_size[2], self.padding, self.strides[2]
117 )
118 if self.data_format == "channels_first":
119 return tf.TensorShape(
120 [input_shape[0], input_shape[1], len_dim1, len_dim2, len_dim3]
121 )
122 else:
123 return tf.TensorShape(
124 [input_shape[0], len_dim1, len_dim2, len_dim3, input_shape[4]]
125 )
127 def get_config(self):
128 config = {
129 "pool_size": self.pool_size,
130 "padding": self.padding,
131 "strides": self.strides,
132 "data_format": self.data_format,
133 }
134 base_config = super().get_config()
135 return dict(list(base_config.items()) + list(config.items()))