FilesExpand file tree

 master

/

tutorial_dynamic_rnn.py

Copy path

Blame

More file actions

Blame

More file actions

Latest commit

 

History

History

History

164 lines (114 loc) · 4.56 KB

 master

/

tutorial_dynamic_rnn.py

Copy path

Top

File metadata and controls

• Code
• Blame

164 lines (114 loc) · 4.56 KB

Raw

Copy raw file

Download raw file

Open symbols panel

Edit and raw actions

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

#! /usr/bin/python

# -*- coding: utf8 -*-

import tensorflow as tf

import tensorlayer as tl

import numpy as np

"""

You will learn:

1. How to use Dynamic RNN with TensorFlow only.

2. For TensorLayer, please see DynamicRNNLayer.

Reference

----------

1. `Wild-ML Blog <http://www.wildml.com/2016/08/rnns-in-tensorflow-a-practical-guide-and-undocumented-features/>`

2. `dynamic_rnn.ipynb <https://github.com/dennybritz/tf-rnn/blob/master/dynamic_rnn.ipynb>`_

"""

## 1. Simple example

# Create input data of 2 example

X = np.random.randn(2, 10, 8) # [batch_size, n_step, n_features]

# The second example is of length 6, the first example is of length 10

X[1,6:] = 0

X_lengths = [10, 6]

cell = tf.nn.rnn_cell.LSTMCell(num_units=64, state_is_tuple=True)

outputs, last_states = tf.nn.dynamic_rnn(

cell=cell,

inputs=X,

sequence_length=X_lengths,

dtype=tf.float64,

)

result = tf.contrib.learn.run_n(

{"outputs": outputs, "last_states": last_states}, n=1, feed_dict=None)

assert result[0]["outputs"].shape == (2, 10, 64) # [batch_size, n_step, n_hidden]

# Outputs for the second example past past length 6 should be 0

assert (result[0]["outputs"][1,7,:] == np.zeros(cell.output_size)).all()

## 2. How to control the initial state

batch_size = X.shape[0]

with tf.variable_scope('name') as vs:

cell = tf.nn.rnn_cell.LSTMCell(num_units=64, state_is_tuple=True)

initial_state = cell.zero_state(batch_size, dtype=tf.float64)#"float")

outputs, last_states = tf.nn.dynamic_rnn(

cell=cell,

inputs=X,

sequence_length=X_lengths,

initial_state=initial_state,

dtype=tf.float64,

)

result = tf.contrib.learn.run_n(

{"outputs": outputs, "last_states": last_states}, n=1, feed_dict=None)

assert result[0]["outputs"].shape == (2, 10, 64)

assert (result[0]["outputs"][1,7,:] == np.zeros(cell.output_size)).all()

## 3. How to automatically get the last outputs by automatically compute X_lengths

def advanced_indexing_op(input, index):

""" Advanced Indexing for Sequences. """

batch_size = tf.shape(input)[0]

max_length = int(input.get_shape()[1])

dim_size = int(input.get_shape()[2])

index = tf.range(0, batch_size) * max_length + (index - 1)

flat = tf.reshape(input, [-1, dim_size])

relevant = tf.gather(flat, index)

return relevant

def retrieve_seq_length_op(data):

""" An op to compute the length of a sequence. 0 are masked. """

with tf.name_scope('GetLength'):

used = tf.sign(tf.reduce_max(tf.abs(data), reduction_indices=2))

length = tf.reduce_sum(used, reduction_indices=1)

length = tf.cast(length, tf.int32)

return length

sequence_length = retrieve_seq_length_op(

incoming if isinstance(X, tf.Tensor) else tf.pack(X))

batch_size = X.shape[0]

with tf.variable_scope('name2') as vs: #, initializer=tf.constant_initializer(value=0.1)) as vs:

cell = tf.nn.rnn_cell.LSTMCell(num_units=64, state_is_tuple=True)

initial_state = cell.zero_state(batch_size, dtype=tf.float64)#"float")

outputs, last_states = tf.nn.dynamic_rnn(

cell=cell,

inputs=X,

# sequence_length=X_lengths,

sequence_length= sequence_length,

initial_state=initial_state,

dtype=tf.float64,

)

result = tf.contrib.learn.run_n(

{"outputs": outputs, "last_states": last_states}, n=1, feed_dict=None)

print('all outputs', result[0]["outputs"])

# print(' outputs 2nd', result[0]["outputs"][1,5])

# exit()

# print(sequence_length)

# exit()

# automatically get the last output

outputs = tf.transpose(tf.pack(outputs), [1, 0, 2])

last_outputs = advanced_indexing_op(outputs, sequence_length)

last_states = result[0]["last_states"]

sess = tf.Session()

sess.run(tf.initialize_all_variables())

# print('last outputs',sess.run(last_outputs)) # (2, 64) # TO DO

# print('last lstm states',last_states, last_states.c.shape, last_states.h.shape)

# 4. How to use DynamicRNNLayer

# data = [[1,2,3,4,5,6,7,8],

# [4,5,6,7],

# [2,5,7]]

# # [batch_size, n_step, n_features]

# # x = tf.placeholder(tf.int32, [3, -1, 1])

#

#

# batched_data = tf.train.batch(

# tensors=[data],

# batch_size=3,

# dynamic_pad=True,

# name="x_batch"

# )

# res = tf.contrib.learn.run_n({"y": batched_data}, n=1, feed_dict=None)

#

# # Print the result

# print("Batch shape: {}".format(res[0]["y"].shape))

# print(res[0]["y"])

#