Coverage for /pythoncovmergedfiles/medio/medio/usr/local/lib/python3.8/site-packages/keras/src/layers/pooling/base_pooling1d.py: 28%

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 1D layers.""" 

16 

17 

18import tensorflow.compat.v2 as tf 

19 

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 

24 

25 

26class Pooling1D(Layer): 

27 """Pooling layer for arbitrary pooling functions, for 1D inputs. 

28 

29 This class only exists for code reuse. It will never be an exposed API. 

30 

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 a single integer, 

34 representing the size of the pooling window. 

35 strides: An integer or tuple/list of a single integer, specifying the 

36 strides of the pooling operation. 

37 padding: A string. The padding method, either 'valid' or 'same'. 

38 Case-insensitive. 

39 data_format: A string, 

40 one of `channels_last` (default) or `channels_first`. 

41 The ordering of the dimensions in the inputs. 

42 `channels_last` corresponds to inputs with shape 

43 `(batch, steps, features)` while `channels_first` 

44 corresponds to inputs with shape 

45 `(batch, features, steps)`. 

46 name: A string, the name of the layer. 

47 """ 

48 

49 def __init__( 

50 self, 

51 pool_function, 

52 pool_size, 

53 strides, 

54 padding="valid", 

55 data_format="channels_last", 

56 name=None, 

57 **kwargs 

58 ): 

59 super().__init__(name=name, **kwargs) 

60 if data_format is None: 

61 data_format = backend.image_data_format() 

62 if strides is None: 

63 strides = pool_size 

64 self.pool_function = pool_function 

65 self.pool_size = conv_utils.normalize_tuple(pool_size, 1, "pool_size") 

66 self.strides = conv_utils.normalize_tuple( 

67 strides, 1, "strides", allow_zero=True 

68 ) 

69 self.padding = conv_utils.normalize_padding(padding) 

70 self.data_format = conv_utils.normalize_data_format(data_format) 

71 self.input_spec = InputSpec(ndim=3) 

72 

73 def call(self, inputs): 

74 pad_axis = 2 if self.data_format == "channels_last" else 3 

75 inputs = tf.expand_dims(inputs, pad_axis) 

76 outputs = self.pool_function( 

77 inputs, 

78 self.pool_size + (1,), 

79 strides=self.strides + (1,), 

80 padding=self.padding, 

81 data_format=self.data_format, 

82 ) 

83 return tf.squeeze(outputs, pad_axis) 

84 

85 def compute_output_shape(self, input_shape): 

86 input_shape = tf.TensorShape(input_shape).as_list() 

87 if self.data_format == "channels_first": 

88 steps = input_shape[2] 

89 features = input_shape[1] 

90 else: 

91 steps = input_shape[1] 

92 features = input_shape[2] 

93 length = conv_utils.conv_output_length( 

94 steps, self.pool_size[0], self.padding, self.strides[0] 

95 ) 

96 if self.data_format == "channels_first": 

97 return tf.TensorShape([input_shape[0], features, length]) 

98 else: 

99 return tf.TensorShape([input_shape[0], length, features]) 

100 

101 def get_config(self): 

102 config = { 

103 "strides": self.strides, 

104 "pool_size": self.pool_size, 

105 "padding": self.padding, 

106 "data_format": self.data_format, 

107 } 

108 base_config = super().get_config() 

109 return dict(list(base_config.items()) + list(config.items())) 

110