質問編集履歴
3
test
CHANGED
File without changes
|
test
CHANGED
File without changes
|
2
タイトル
test
CHANGED
@@ -1 +1 @@
|
|
1
|
-
保存したモデルの再利用時のエラーについて
|
1
|
+
[pycharm]保存したモデルの再利用時のエラーについて[tensorflow]
|
test
CHANGED
File without changes
|
1
コードの修正
test
CHANGED
File without changes
|
test
CHANGED
@@ -28,15 +28,13 @@
|
|
28
28
|
|
29
29
|
|
30
30
|
|
31
|
-
|
31
|
+
|
32
32
|
|
33
33
|
dir_dev = r"data"
|
34
34
|
|
35
35
|
n_im_dev = 90 # How many images to load
|
36
36
|
|
37
|
-
|
37
|
+
|
38
|
-
|
39
|
-
# No normalization or rotations:
|
40
38
|
|
41
39
|
X_dev, Y_dev = load_images_from_folder( # Load images for evaluating model
|
42
40
|
|
@@ -180,8 +178,6 @@
|
|
180
178
|
|
181
179
|
reuse=tf.AUTO_REUSE,
|
182
180
|
|
183
|
-
#reuse=False,
|
184
|
-
|
185
181
|
variables_collections=None,
|
186
182
|
|
187
183
|
outputs_collections=None,
|
@@ -214,8 +210,6 @@
|
|
214
210
|
|
215
211
|
reuse=tf.AUTO_REUSE,
|
216
212
|
|
217
|
-
# reuse=False,
|
218
|
-
|
219
213
|
variables_collections=None,
|
220
214
|
|
221
215
|
outputs_collections=None,
|
@@ -226,13 +220,13 @@
|
|
226
220
|
|
227
221
|
|
228
222
|
|
229
|
-
|
223
|
+
|
230
224
|
|
231
225
|
FC_M = tf.reshape(FC2, [tf.shape(x)[0], tf.shape(x)[1], tf.shape(x)[2], 1])
|
232
226
|
|
233
227
|
|
234
228
|
|
235
|
-
|
229
|
+
|
236
230
|
|
237
231
|
W1 = parameters['W1']
|
238
232
|
|
@@ -242,33 +236,29 @@
|
|
242
236
|
|
243
237
|
|
244
238
|
|
245
|
-
|
239
|
+
|
246
|
-
|
247
|
-
# Input size (n_im, n_H0, n_W0, 1), output size (n_im, n_H0, n_W0, 64)
|
248
240
|
|
249
241
|
Z1 = tf.nn.conv2d(FC_M, W1, strides=[1, 1, 1, 1], padding='SAME')
|
250
242
|
|
251
|
-
|
243
|
+
|
252
244
|
|
253
245
|
CONV1 = tf.nn.relu(Z1)
|
254
246
|
|
255
247
|
|
256
248
|
|
257
|
-
|
249
|
+
|
258
|
-
|
259
|
-
# Input size (n_im, n_H0, n_W0, 64), output size (n_im, n_H0, n_W0, 64)
|
260
250
|
|
261
251
|
Z2 = tf.nn.conv2d(CONV1, W2, strides=[1, 1, 1, 1], padding='SAME')
|
262
252
|
|
263
|
-
|
253
|
+
|
264
254
|
|
265
255
|
CONV2 = tf.nn.relu(Z2)
|
266
256
|
|
267
257
|
|
268
258
|
|
269
|
-
|
259
|
+
|
270
|
-
|
271
|
-
|
260
|
+
|
261
|
+
|
272
262
|
|
273
263
|
batch_size = tf.shape(x)[0]
|
274
264
|
|
@@ -306,31 +296,31 @@
|
|
306
296
|
|
307
297
|
|
308
298
|
|
309
|
-
|
299
|
+
|
310
300
|
|
311
301
|
X, Y = create_placeholders(n_H0, n_W0)
|
312
302
|
|
313
303
|
|
314
304
|
|
315
|
-
|
305
|
+
|
316
306
|
|
317
307
|
parameters = initialize_parameters()
|
318
308
|
|
319
309
|
|
320
310
|
|
321
|
-
|
311
|
+
|
322
312
|
|
323
313
|
forward_propagation(X, parameters)
|
324
314
|
|
325
315
|
|
326
316
|
|
327
|
-
|
317
|
+
|
328
318
|
|
329
319
|
saver = tf.train.Saver()
|
330
320
|
|
331
321
|
|
332
322
|
|
333
|
-
|
323
|
+
|
334
324
|
|
335
325
|
with tf.Session() as sess:
|
336
326
|
|
@@ -356,8 +346,6 @@
|
|
356
346
|
|
357
347
|
|
358
348
|
|
359
|
-
# Reconstruct the image using trained model
|
360
|
-
|
361
349
|
_, Y_recon = model(X_dev)
|
362
350
|
|
363
351
|
print('Y_recon.shape = ', Y_recon.shape)
|
@@ -368,9 +356,7 @@
|
|
368
356
|
|
369
357
|
|
370
358
|
|
371
|
-
|
359
|
+
|
372
|
-
|
373
|
-
# 4 images to visualize:
|
374
360
|
|
375
361
|
im1 = 32
|
376
362
|
|
@@ -382,15 +368,13 @@
|
|
382
368
|
|
383
369
|
|
384
370
|
|
385
|
-
|
371
|
+
|
386
|
-
|
387
|
-
# Complex image from real and imaginary part
|
388
372
|
|
389
373
|
X_dev_compl = X_dev[:, :, :, 0] + X_dev[:, :, :, 1] * 1j
|
390
374
|
|
391
375
|
|
392
376
|
|
393
|
-
|
377
|
+
|
394
378
|
|
395
379
|
X_iFFT0 = np.fft.ifft2(X_dev_compl[im1, :, :])
|
396
380
|
|
@@ -402,7 +386,7 @@
|
|
402
386
|
|
403
387
|
|
404
388
|
|
405
|
-
|
389
|
+
|
406
390
|
|
407
391
|
X_iFFT_M1 = np.sqrt(np.power(X_iFFT0.real, 2)
|
408
392
|
|
@@ -422,10 +406,6 @@
|
|
422
406
|
|
423
407
|
|
424
408
|
|
425
|
-
# SHOW
|
426
|
-
|
427
|
-
# Show Y - input images
|
428
|
-
|
429
409
|
plt.subplot(341), plt.imshow(Y_dev[im1, :, :], cmap='gray')
|
430
410
|
|
431
411
|
plt.title('Y_dev1'), plt.xticks([]), plt.yticks([])
|
@@ -444,7 +424,7 @@
|
|
444
424
|
|
445
425
|
|
446
426
|
|
447
|
-
|
427
|
+
|
448
428
|
|
449
429
|
plt.subplot(345), plt.imshow(X_iFFT_M1, cmap='gray')
|
450
430
|
|
@@ -464,7 +444,7 @@
|
|
464
444
|
|
465
445
|
|
466
446
|
|
467
|
-
|
447
|
+
|
468
448
|
|
469
449
|
plt.subplot(349), plt.imshow(Y_recon[im1, :, :], cmap='gray')
|
470
450
|
|