質問編集履歴

2

normalのスペルミスを訂正し、加筆しましたあ。

2017/07/24 04:36

投稿

daisuke_factex
daisuke_factex

スコア42

test CHANGED
File without changes
test CHANGED
@@ -92,19 +92,15 @@
92
92
 
93
93
 
94
94
 
95
- 再度動かしたところ、
95
+ nomal→normalに訂正したところ、
96
96
 
97
97
 
98
98
 
99
99
  ```
100
100
 
101
- >>> import tensorflow as tf
102
-
103
- >>>
104
-
105
101
  >>> def weight(shape = []):
106
102
 
107
- ... initial = tf.truncated_nomal(shape, stddev = 0.01)
103
+ ... initial = tf.truncated_normal(shape, stddev = 0.01)
108
104
 
109
105
  ... return tf.Variable(initial)
110
106
 
@@ -126,26 +122,76 @@
126
122
 
127
123
  >>> W = weight(shape = [Q,P])
128
124
 
129
- Traceback (most recent call last):
130
-
131
- File "<stdin>", line 1, in <module>
132
-
133
- File "<stdin>", line 2, in weight
134
-
135
- AttributeError: module 'tensorflow' has no attribute 'truncated_nomal'
136
-
137
125
  >>> b = bias(shape = [P])
138
126
 
139
127
  >>> f = tf.matmul(X ,W)+b
140
128
 
141
129
  Traceback (most recent call last):
142
130
 
131
+ File "/home/daisuke/.local/lib/python3.5/site-packages/tensorflow/python/framework/common_shapes.py", line 671, in _call_cpp_shape_fn_impl
132
+
133
+ input_tensors_as_shapes, status)
134
+
135
+ File "/usr/lib/python3.5/contextlib.py", line 66, in __exit__
136
+
137
+ next(self.gen)
138
+
139
+ File "/home/daisuke/.local/lib/python3.5/site-packages/tensorflow/python/framework/errors_impl.py", line 466, in raise_exception_on_not_ok_status
140
+
141
+ pywrap_tensorflow.TF_GetCode(status))
142
+
143
+ tensorflow.python.framework.errors_impl.InvalidArgumentError: Dimensions must be equal, but are 0 and 2 for 'MatMul_1' (op: 'MatMul') with input shapes: [?,0], [2,5].
144
+
145
+
146
+
147
+ During handling of the above exception, another exception occurred:
148
+
149
+
150
+
151
+ Traceback (most recent call last):
152
+
143
153
  File "<stdin>", line 1, in <module>
144
154
 
155
+ File "/home/daisuke/.local/lib/python3.5/site-packages/tensorflow/python/ops/math_ops.py", line 1816, in matmul
156
+
157
+ a, b, transpose_a=transpose_a, transpose_b=transpose_b, name=name)
158
+
159
+ File "/home/daisuke/.local/lib/python3.5/site-packages/tensorflow/python/ops/gen_math_ops.py", line 1217, in _mat_mul
160
+
161
+ transpose_b=transpose_b, name=name)
162
+
163
+ File "/home/daisuke/.local/lib/python3.5/site-packages/tensorflow/python/framework/op_def_library.py", line 767, in apply_op
164
+
165
+ op_def=op_def)
166
+
167
+ File "/home/daisuke/.local/lib/python3.5/site-packages/tensorflow/python/framework/ops.py", line 2508, in create_op
168
+
169
+ set_shapes_for_outputs(ret)
170
+
171
+ File "/home/daisuke/.local/lib/python3.5/site-packages/tensorflow/python/framework/ops.py", line 1873, in set_shapes_for_outputs
172
+
173
+ shapes = shape_func(op)
174
+
175
+ File "/home/daisuke/.local/lib/python3.5/site-packages/tensorflow/python/framework/ops.py", line 1823, in call_with_requiring
176
+
177
+ return call_cpp_shape_fn(op, require_shape_fn=True)
178
+
179
+ File "/home/daisuke/.local/lib/python3.5/site-packages/tensorflow/python/framework/common_shapes.py", line 610, in call_cpp_shape_fn
180
+
145
- NameError: name 'W' is not defined
181
+ debug_python_shape_fn, require_shape_fn)
182
+
183
+ File "/home/daisuke/.local/lib/python3.5/site-packages/tensorflow/python/framework/common_shapes.py", line 676, in _call_cpp_shape_fn_impl
184
+
185
+ raise ValueError(err.message)
186
+
187
+ ValueError: Dimensions must be equal, but are 0 and 2 for 'MatMul_1' (op: 'MatMul') with input shapes: [?,0], [2,5].
188
+
189
+ >>>
190
+
191
+
146
192
 
147
193
 
148
194
 
149
195
  ```
150
196
 
151
- 違うエラーが出てきてしいました
197
+ より複雑なエラーが出てきました・・・

1

エラーなどの全文を追加しました。

2017/07/24 04:36

投稿

daisuke_factex
daisuke_factex

スコア42

test CHANGED
File without changes
test CHANGED
@@ -83,3 +83,69 @@
83
83
  python3.x
84
84
 
85
85
  tensorflow 0.7.1
86
+
87
+
88
+
89
+
90
+
91
+ ####追記
92
+
93
+
94
+
95
+ 再度動かしたところ、
96
+
97
+
98
+
99
+ ```
100
+
101
+ >>> import tensorflow as tf
102
+
103
+ >>>
104
+
105
+ >>> def weight(shape = []):
106
+
107
+ ... initial = tf.truncated_nomal(shape, stddev = 0.01)
108
+
109
+ ... return tf.Variable(initial)
110
+
111
+ ...
112
+
113
+ >>> def bias(dtype =tf.float32, shape = []):
114
+
115
+ ... initial = tf.zeros(shape, dtype= dtype)
116
+
117
+ ... return tf.Variable(initial)
118
+
119
+ ...
120
+
121
+ >>> P,Q=5,2
122
+
123
+ >>>
124
+
125
+ >>> X = tf.placeholder(dtype = tf.float32, shape =[None, 0])
126
+
127
+ >>> W = weight(shape = [Q,P])
128
+
129
+ Traceback (most recent call last):
130
+
131
+ File "<stdin>", line 1, in <module>
132
+
133
+ File "<stdin>", line 2, in weight
134
+
135
+ AttributeError: module 'tensorflow' has no attribute 'truncated_nomal'
136
+
137
+ >>> b = bias(shape = [P])
138
+
139
+ >>> f = tf.matmul(X ,W)+b
140
+
141
+ Traceback (most recent call last):
142
+
143
+ File "<stdin>", line 1, in <module>
144
+
145
+ NameError: name 'W' is not defined
146
+
147
+
148
+
149
+ ```
150
+
151
+ と違うエラーが出てきてしまいました。