

In [2]:

    
t1 = regression_test()
t1.main()
t1.plot_target_function()
t1.plot_learned_function()









    



epoch 0, learning rate 0.001
mean_squared_error at step 0: 0.93096






    












    












    



mean_squared_error at step 10: 0.513602
mean_squared_error at step 20: 0.724879
mean_squared_error at step 30: 0.538038
mean_squared_error at step 40: 0.488257
mean_squared_error at step 50: 0.829993
mean_squared_error at step 60: 0.817744
mean_squared_error at step 70: 0.758409
mean_squared_error at step 80: 0.512449
mean_squared_error at step 90: 0.686509
Adding run metadata for 99
mean_squared_error at step 100: 0.575598
mean_squared_error at step 110: 0.655199
mean_squared_error at step 120: 0.500416
mean_squared_error at step 130: 0.431964
mean_squared_error at step 140: 0.46208
mean_squared_error at step 150: 0.448454
mean_squared_error at step 160: 0.403187
mean_squared_error at step 170: 0.310823
mean_squared_error at step 180: 0.478239
mean_squared_error at step 190: 0.300645
Adding run metadata for 199
mean_squared_error at step 200: 0.39098
mean_squared_error at step 210: 0.387448
mean_squared_error at step 220: 0.277
mean_squared_error at step 230: 0.325153
mean_squared_error at step 240: 0.406707
mean_squared_error at step 250: 0.215618
mean_squared_error at step 260: 0.346688
mean_squared_error at step 270: 0.309476
mean_squared_error at step 280: 0.341576
mean_squared_error at step 290: 0.341816
Adding run metadata for 299
mean_squared_error at step 300: 0.205261
mean_squared_error at step 310: 0.301954
mean_squared_error at step 320: 0.171609
mean_squared_error at step 330: 0.339538
mean_squared_error at step 340: 0.421771
mean_squared_error at step 350: 0.229932
mean_squared_error at step 360: 0.295231
mean_squared_error at step 370: 0.385098
mean_squared_error at step 380: 0.287757
mean_squared_error at step 390: 0.187043
Adding run metadata for 399
mean_squared_error at step 400: 0.381218
mean_squared_error at step 410: 0.307151
mean_squared_error at step 420: 0.340749
mean_squared_error at step 430: 0.326814
mean_squared_error at step 440: 0.201297
mean_squared_error at step 450: 0.361614
mean_squared_error at step 460: 0.373912
mean_squared_error at step 470: 0.280378
mean_squared_error at step 480: 0.28364
mean_squared_error at step 490: 0.37707
Adding run metadata for 499
mean_squared_error at step 500: 0.391538
mean_squared_error at step 510: 0.286749
mean_squared_error at step 520: 0.296542
mean_squared_error at step 530: 0.304832
mean_squared_error at step 540: 0.346294
mean_squared_error at step 550: 0.210388
mean_squared_error at step 560: 0.158737
mean_squared_error at step 570: 0.166001
mean_squared_error at step 580: 0.258276
mean_squared_error at step 590: 0.3204
Adding run metadata for 599
mean_squared_error at step 600: 0.264778
mean_squared_error at step 610: 0.191161
mean_squared_error at step 620: 0.378212
mean_squared_error at step 630: 0.195428
mean_squared_error at step 640: 0.204435
mean_squared_error at step 650: 0.269477
mean_squared_error at step 660: 0.385042
mean_squared_error at step 670: 0.253094
mean_squared_error at step 680: 0.328341
mean_squared_error at step 690: 0.318946
Adding run metadata for 699
mean_squared_error at step 700: 0.315397
mean_squared_error at step 710: 0.194027
mean_squared_error at step 720: 0.230374
mean_squared_error at step 730: 0.216251
mean_squared_error at step 740: 0.220079
mean_squared_error at step 750: 0.184908
mean_squared_error at step 760: 0.231001
mean_squared_error at step 770: 0.193219
mean_squared_error at step 780: 0.183513
mean_squared_error at step 790: 0.333675
Adding run metadata for 799
mean_squared_error at step 800: 0.17586
mean_squared_error at step 810: 0.207563
mean_squared_error at step 820: 0.205347
mean_squared_error at step 830: 0.310354
mean_squared_error at step 840: 0.211592
mean_squared_error at step 850: 0.170641
mean_squared_error at step 860: 0.183409
mean_squared_error at step 870: 0.250022
mean_squared_error at step 880: 0.226234
mean_squared_error at step 890: 0.249131
Adding run metadata for 899
mean_squared_error at step 900: 0.183679
mean_squared_error at step 910: 0.291811
mean_squared_error at step 920: 0.251731
mean_squared_error at step 930: 0.187654
mean_squared_error at step 940: 0.245309
mean_squared_error at step 950: 0.169191
mean_squared_error at step 960: 0.170191
mean_squared_error at step 970: 0.174376
mean_squared_error at step 980: 0.222295
mean_squared_error at step 990: 0.259514
Adding run metadata for 999
epoch 1, learning rate 0.000599484
mean_squared_error at step 1000: 0.233353






    












    












    



mean_squared_error at step 1010: 0.129791
mean_squared_error at step 1020: 0.165568
mean_squared_error at step 1030: 0.23444
mean_squared_error at step 1040: 0.22728
mean_squared_error at step 1050: 0.209837
mean_squared_error at step 1060: 0.285362
mean_squared_error at step 1070: 0.260559
mean_squared_error at step 1080: 0.321544
mean_squared_error at step 1090: 0.196922
Adding run metadata for 1099
mean_squared_error at step 1100: 0.316578
mean_squared_error at step 1110: 0.158721
mean_squared_error at step 1120: 0.142061
mean_squared_error at step 1130: 0.197337
mean_squared_error at step 1140: 0.332029
mean_squared_error at step 1150: 0.171725
mean_squared_error at step 1160: 0.225019
mean_squared_error at step 1170: 0.184228
mean_squared_error at step 1180: 0.202005
mean_squared_error at step 1190: 0.173889
Adding run metadata for 1199
mean_squared_error at step 1200: 0.212512
mean_squared_error at step 1210: 0.312262
mean_squared_error at step 1220: 0.220522
mean_squared_error at step 1230: 0.362792
mean_squared_error at step 1240: 0.134721
mean_squared_error at step 1250: 0.200121
mean_squared_error at step 1260: 0.249025
mean_squared_error at step 1270: 0.296927
mean_squared_error at step 1280: 0.190962
mean_squared_error at step 1290: 0.193724
Adding run metadata for 1299
mean_squared_error at step 1300: 0.270279
mean_squared_error at step 1310: 0.3435
mean_squared_error at step 1320: 0.226368
mean_squared_error at step 1330: 0.264402
mean_squared_error at step 1340: 0.194958
mean_squared_error at step 1350: 0.293118
mean_squared_error at step 1360: 0.145725
mean_squared_error at step 1370: 0.394868
mean_squared_error at step 1380: 0.436502
mean_squared_error at step 1390: 0.287923
Adding run metadata for 1399
mean_squared_error at step 1400: 0.176074
mean_squared_error at step 1410: 0.167937
mean_squared_error at step 1420: 0.258976
mean_squared_error at step 1430: 0.322887
mean_squared_error at step 1440: 0.242763
mean_squared_error at step 1450: 0.247921
mean_squared_error at step 1460: 0.224026
mean_squared_error at step 1470: 0.28173
mean_squared_error at step 1480: 0.211074
mean_squared_error at step 1490: 0.258194
Adding run metadata for 1499
mean_squared_error at step 1500: 0.15244
mean_squared_error at step 1510: 0.150273
mean_squared_error at step 1520: 0.214954
mean_squared_error at step 1530: 0.286997
mean_squared_error at step 1540: 0.159063
mean_squared_error at step 1550: 0.113519
mean_squared_error at step 1560: 0.187151
mean_squared_error at step 1570: 0.127575
mean_squared_error at step 1580: 0.225302
mean_squared_error at step 1590: 0.0766547
Adding run metadata for 1599
mean_squared_error at step 1600: 0.216783
mean_squared_error at step 1610: 0.182173
mean_squared_error at step 1620: 0.18313
mean_squared_error at step 1630: 0.224778
mean_squared_error at step 1640: 0.204281
mean_squared_error at step 1650: 0.229493
mean_squared_error at step 1660: 0.160298
mean_squared_error at step 1670: 0.252267
mean_squared_error at step 1680: 0.230523
mean_squared_error at step 1690: 0.118422
Adding run metadata for 1699
mean_squared_error at step 1700: 0.230623
mean_squared_error at step 1710: 0.103794
mean_squared_error at step 1720: 0.187948
mean_squared_error at step 1730: 0.189931
mean_squared_error at step 1740: 0.220563
mean_squared_error at step 1750: 0.183084
mean_squared_error at step 1760: 0.327697
mean_squared_error at step 1770: 0.246432
mean_squared_error at step 1780: 0.185527
mean_squared_error at step 1790: 0.134462
Adding run metadata for 1799
mean_squared_error at step 1800: 0.167199
mean_squared_error at step 1810: 0.148338
mean_squared_error at step 1820: 0.128966
mean_squared_error at step 1830: 0.150861
mean_squared_error at step 1840: 0.195555
mean_squared_error at step 1850: 0.167745
mean_squared_error at step 1860: 0.127934
mean_squared_error at step 1870: 0.26611
mean_squared_error at step 1880: 0.107829
mean_squared_error at step 1890: 0.262748
Adding run metadata for 1899
mean_squared_error at step 1900: 0.21449
mean_squared_error at step 1910: 0.257845
mean_squared_error at step 1920: 0.146662
mean_squared_error at step 1930: 0.306521
mean_squared_error at step 1940: 0.150945
mean_squared_error at step 1950: 0.211914
mean_squared_error at step 1960: 0.187322
mean_squared_error at step 1970: 0.15962
mean_squared_error at step 1980: 0.1297
mean_squared_error at step 1990: 0.248273
Adding run metadata for 1999
epoch 2, learning rate 0.000359381
mean_squared_error at step 2000: 0.138196






    












    












    



mean_squared_error at step 2010: 0.240463
mean_squared_error at step 2020: 0.260626
mean_squared_error at step 2030: 0.254858
mean_squared_error at step 2040: 0.135836
mean_squared_error at step 2050: 0.185329
mean_squared_error at step 2060: 0.185363
mean_squared_error at step 2070: 0.136643
mean_squared_error at step 2080: 0.0955505
mean_squared_error at step 2090: 0.22686
Adding run metadata for 2099
mean_squared_error at step 2100: 0.229203
mean_squared_error at step 2110: 0.177493
mean_squared_error at step 2120: 0.086746
mean_squared_error at step 2130: 0.204007
mean_squared_error at step 2140: 0.184626
mean_squared_error at step 2150: 0.116024
mean_squared_error at step 2160: 0.119143
mean_squared_error at step 2170: 0.133705
mean_squared_error at step 2180: 0.246761
mean_squared_error at step 2190: 0.187696
Adding run metadata for 2199
mean_squared_error at step 2200: 0.202236
mean_squared_error at step 2210: 0.113362
mean_squared_error at step 2220: 0.186842
mean_squared_error at step 2230: 0.142962
mean_squared_error at step 2240: 0.129912
mean_squared_error at step 2250: 0.252297
mean_squared_error at step 2260: 0.11741
mean_squared_error at step 2270: 0.257936
mean_squared_error at step 2280: 0.129069
mean_squared_error at step 2290: 0.198846
Adding run metadata for 2299
mean_squared_error at step 2300: 0.1536
mean_squared_error at step 2310: 0.168881
mean_squared_error at step 2320: 0.199762
mean_squared_error at step 2330: 0.154419
mean_squared_error at step 2340: 0.155898
mean_squared_error at step 2350: 0.191441
mean_squared_error at step 2360: 0.216058
mean_squared_error at step 2370: 0.220462
mean_squared_error at step 2380: 0.259913
mean_squared_error at step 2390: 0.170237
Adding run metadata for 2399
mean_squared_error at step 2400: 0.197277
mean_squared_error at step 2410: 0.222046
mean_squared_error at step 2420: 0.124559
mean_squared_error at step 2430: 0.214211
mean_squared_error at step 2440: 0.127654
mean_squared_error at step 2450: 0.274112






    



---------------------------------------------------------------------------
KeyboardInterrupt                         Traceback (most recent call last)
<ipython-input-2-5eb08ea7cdfb> in <module>()
      1 
      2 t1 = regression_test()
----> 3 t1.main()
      4 t1.plot_target_function()
      5 t1.plot_learned_function()

/home/frederik/Dokumente/DeepRL/Tensorflow/regression_test.py in main(self)
    300         tf.gfile.MakeDirs(FLAGS.summaries_dir)
    301 
--> 302         self.train(self.sess)
    303 
    304 

/home/frederik/Dokumente/DeepRL/Tensorflow/regression_test.py in train(self, sess)
    286                         tmpdict[learning_rate] = l_r
    287                         summary, _ = sess.run([merged, train_step],feed_dict=tmpdict)
--> 288                         train_writer.add_summary(summary, step + epoch*steps_per_epoch)
    289 
    290                 if (step + epoch*steps_per_epoch) %1000 == 0:

/home/frederik/anaconda2/lib/python2.7/site-packages/tensorflow/python/training/summary_io.pyc in add_summary(self, summary, global_step)
    133     if isinstance(summary, bytes):
    134       summ = summary_pb2.Summary()
--> 135       summ.ParseFromString(summary)
    136       summary = summ
    137     event = event_pb2.Event(wall_time=time.time(), summary=summary)

/home/frederik/anaconda2/lib/python2.7/site-packages/google/protobuf/message.pyc in ParseFromString(self, serialized)
    183     """
    184     self.Clear()
--> 185     self.MergeFromString(serialized)
    186 
    187   def SerializeToString(self):

/home/frederik/anaconda2/lib/python2.7/site-packages/google/protobuf/internal/python_message.pyc in MergeFromString(self, serialized)
   1089     length = len(serialized)
   1090     try:
-> 1091       if self._InternalParse(serialized, 0, length) != length:
   1092         # The only reason _InternalParse would return early is if it
   1093         # encountered an end-group tag.

/home/frederik/anaconda2/lib/python2.7/site-packages/google/protobuf/internal/python_message.pyc in InternalParse(self, buffer, pos, end)
   1125         pos = new_pos
   1126       else:
-> 1127         pos = field_decoder(buffer, new_pos, end, self, field_dict)
   1128         if field_desc:
   1129           self._UpdateOneofState(field_desc)

/home/frederik/anaconda2/lib/python2.7/site-packages/google/protobuf/internal/decoder.pyc in DecodeRepeatedField(buffer, pos, end, message, field_dict)
    610           raise _DecodeError('Truncated message.')
    611         # Read sub-message.
--> 612         if value.add()._InternalParse(buffer, pos, new_pos) != new_pos:
    613           # The only reason _InternalParse would return early is if it
    614           # encountered an end-group tag.

/home/frederik/anaconda2/lib/python2.7/site-packages/google/protobuf/internal/python_message.pyc in InternalParse(self, buffer, pos, end)
   1107 
   1108   def InternalParse(self, buffer, pos, end):
-> 1109     self._Modified()
   1110     field_dict = self._fields
   1111     unknown_field_list = self._unknown_fields

/home/frederik/anaconda2/lib/python2.7/site-packages/google/protobuf/internal/python_message.pyc in Modified(self)
   1328       self._listener_for_children.dirty = True
   1329       self._is_present_in_parent = True
-> 1330       self._listener.Modified()
   1331 
   1332   def _UpdateOneofState(self, field):

KeyboardInterrupt: