質問編集履歴

5

追記

2018/05/21 04:54

投稿

6123_sadaharu-.
6123_sadaharu-.

スコア6

test CHANGED
File without changes
test CHANGED
@@ -40,7 +40,9 @@
40
40
 
41
41
  必要があればコードやスペックなども記述しますので、宜しくお願い致します。
42
42
 
43
-
43
+ 以下、nvidia-smiを実行。
44
+
45
+ GPU-Util = 0%なのが気がかりです。
44
46
 
45
47
 
46
48
 

4

追記

2018/05/21 04:54

投稿

6123_sadaharu-.
6123_sadaharu-.

スコア6

test CHANGED
File without changes
test CHANGED
@@ -81,3 +81,213 @@
81
81
 
82
82
 
83
83
  ```
84
+
85
+
86
+
87
+ 以下、上手く実行された例
88
+
89
+ ```ここに言語を入力
90
+
91
+ # -*- coding: utf-8 -*-
92
+
93
+
94
+
95
+ import tensorflow as tf
96
+
97
+ import numpy as np
98
+
99
+
100
+
101
+ # Create 100 phony x, y data points in NumPy, y = x * 0.1 + 0.3
102
+
103
+ x_data = np.random.rand(1000,100).astype(np.float32)
104
+
105
+ y_data = x_data * 0.1 + 0.3
106
+
107
+ print(x_data)
108
+
109
+
110
+
111
+ X = tf.placeholder(dtype = tf.float32, shape = [None, x_data.shape[1]])
112
+
113
+ Y = tf.placeholder(dtype = tf.float32, shape = [None, y_data.shape[1]])
114
+
115
+ # Try to find values for W and b that compute y_data = W * x_data + b
116
+
117
+ # (We know that W should be 0.1 and b 0.3, but TensorFlow will
118
+
119
+ # figure that out for us.)
120
+
121
+ W = tf.Variable(tf.random_uniform([100,100], -1.0, 1.0))
122
+
123
+ b = tf.Variable(tf.zeros([1]))
124
+
125
+
126
+
127
+ ######相違点#######
128
+
129
+ y = W * X + b
130
+
131
+ #y=tf.matmul(W,X)+b
132
+
133
+
134
+
135
+ # Minimize the mean squared errors.
136
+
137
+ loss = tf.reduce_mean(tf.square(y -Y))
138
+
139
+ optimizer = tf.train.GradientDescentOptimizer(0.5)
140
+
141
+ train = optimizer.minimize(loss)
142
+
143
+
144
+
145
+ # Before starting, initialize the variables. We will 'run' this first.
146
+
147
+ init = tf.global_variables_initializer()
148
+
149
+
150
+
151
+ # Launch the graph.
152
+
153
+ sess = tf.Session()
154
+
155
+ sess.run(init)
156
+
157
+ #print(x_data)
158
+
159
+ #print(y_data)
160
+
161
+
162
+
163
+ # Fit the line.
164
+
165
+ for step in range(201):
166
+
167
+ #sess.run(train)
168
+
169
+ sess.run(train,feed_dict={X:x_data[0:100],Y:y_data[0:100]})
170
+
171
+ if step % 20 == 0:
172
+
173
+ print(step, sess.run(W), sess.run(b))
174
+
175
+
176
+
177
+ # Learns best fit is W: [0.1], b: [0.3]
178
+
179
+
180
+
181
+ # Close the Session when we're done.
182
+
183
+ sess.close()
184
+
185
+
186
+
187
+ ```
188
+
189
+
190
+
191
+ 以下、上手くいかずフリーズした例
192
+
193
+ ```ここに言語を入力
194
+
195
+ # -*- coding: utf-8 -*-
196
+
197
+
198
+
199
+ import tensorflow as tf
200
+
201
+ import numpy as np
202
+
203
+
204
+
205
+ # Create 100 phony x, y data points in NumPy, y = x * 0.1 + 0.3
206
+
207
+ x_data = np.random.rand(1000,100).astype(np.float32)
208
+
209
+ y_data = x_data * 0.1 + 0.3
210
+
211
+ print(x_data)
212
+
213
+
214
+
215
+ X = tf.placeholder(dtype = tf.float32, shape = [None, x_data.shape[1]])
216
+
217
+ Y = tf.placeholder(dtype = tf.float32, shape = [None, y_data.shape[1]])
218
+
219
+ # Try to find values for W and b that compute y_data = W * x_data + b
220
+
221
+ # (We know that W should be 0.1 and b 0.3, but TensorFlow will
222
+
223
+ # figure that out for us.)
224
+
225
+ W = tf.Variable(tf.random_uniform([100,100], -1.0, 1.0))
226
+
227
+ b = tf.Variable(tf.zeros([1]))
228
+
229
+
230
+
231
+ ######相違点#######
232
+
233
+ #y = W * X + b
234
+
235
+ y=tf.matmul(W,X)+b
236
+
237
+
238
+
239
+
240
+
241
+ # Minimize the mean squared errors.
242
+
243
+ loss = tf.reduce_mean(tf.square(y -Y))
244
+
245
+ optimizer = tf.train.GradientDescentOptimizer(0.5)
246
+
247
+ train = optimizer.minimize(loss)
248
+
249
+
250
+
251
+ # Before starting, initialize the variables. We will 'run' this first.
252
+
253
+ init = tf.global_variables_initializer()
254
+
255
+
256
+
257
+ # Launch the graph.
258
+
259
+ sess = tf.Session()
260
+
261
+ sess.run(init)
262
+
263
+ #print(x_data)
264
+
265
+ #print(y_data)
266
+
267
+
268
+
269
+ # Fit the line.
270
+
271
+ for step in range(201):
272
+
273
+ #sess.run(train)
274
+
275
+ sess.run(train,feed_dict={X:x_data[0:100],Y:y_data[0:100]})
276
+
277
+ if step % 20 == 0:
278
+
279
+ print(step, sess.run(W), sess.run(b))
280
+
281
+
282
+
283
+ # Learns best fit is W: [0.1], b: [0.3]
284
+
285
+
286
+
287
+ # Close the Session when we're done.
288
+
289
+ sess.close()
290
+
291
+
292
+
293
+ ```

3

nvidia-smiの追記

2018/05/21 04:13

投稿

6123_sadaharu-.
6123_sadaharu-.

スコア6

test CHANGED
File without changes
test CHANGED
@@ -39,3 +39,45 @@
39
39
  ご意見よろしくお願い致します。
40
40
 
41
41
  必要があればコードやスペックなども記述しますので、宜しくお願い致します。
42
+
43
+
44
+
45
+
46
+
47
+ ```
48
+
49
+ +-----------------------------------------------------------------------------+
50
+
51
+ | NVIDIA-SMI 384.111 Driver Version: 384.111 |
52
+
53
+ |-------------------------------+----------------------+----------------------+
54
+
55
+ | GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
56
+
57
+ | Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
58
+
59
+ |===============================+======================+======================|
60
+
61
+ | 0 GeForce GTX TIT... Off | 00000000:02:00.0 Off | N/A |
62
+
63
+ | 22% 48C P2 67W / 250W | 11645MiB / 12205MiB | 0% Default |
64
+
65
+ +-------------------------------+----------------------+----------------------+
66
+
67
+
68
+
69
+ +-----------------------------------------------------------------------------+
70
+
71
+ | Processes: GPU Memory |
72
+
73
+ | GPU PID Type Process name Usage |
74
+
75
+ |=============================================================================|
76
+
77
+ | 0 7449 C python 11634MiB |
78
+
79
+ +-----------------------------------------------------------------------------+
80
+
81
+
82
+
83
+ ```

2

誤字

2018/05/20 13:43

投稿

6123_sadaharu-.
6123_sadaharu-.

スコア6

test CHANGED
File without changes
test CHANGED
@@ -35,3 +35,7 @@
35
35
 
36
36
 
37
37
  feed_dictを使うと停止してしまうみたいなのですが、どうすれば改善できると思いますか?
38
+
39
+ ご意見よろしくお願い致します。
40
+
41
+ 必要があればコードやスペックなども記述しますので、宜しくお願い致します。

1

誤字

2018/05/20 11:14

投稿

6123_sadaharu-.
6123_sadaharu-.

スコア6

test CHANGED
File without changes
test CHANGED
@@ -1,8 +1,6 @@
1
1
  ### 前提・実現したいこと
2
2
 
3
3
 
4
-
5
- ここに質問の内容を詳しく書いてください。
6
4
 
7
5
  tensorflowを使ってGPUで学習をしようとしています。
8
6