Using all organ points!
Starting epoch 1...
100% 324/324 [02:31<00:00,  2.14it/s, Batch loss=103]
100% 68/68 [00:05<00:00, 12.98it/s]
The accuracy on the non masked validation set is 1.16
100% 68/68 [00:06<00:00, 10.47it/s]
The accuracy on the masked validation set is 0.89
======================
Found new best with avg accuracy: 1.02 on epoch 1. Saving model!!!
======================
Saving intermediate checkpoint...
Starting epoch 2...
100% 324/324 [02:31<00:00,  2.13it/s, Batch loss=91.8]
100% 68/68 [00:05<00:00, 12.77it/s]
The accuracy on the non masked validation set is 2.45
100% 68/68 [00:06<00:00, 10.25it/s]
The accuracy on the masked validation set is 0.94
======================
Found new best with avg accuracy: 1.7 on epoch 2. Saving model!!!
======================
Saving intermediate checkpoint...
Starting epoch 3...
100% 324/324 [02:31<00:00,  2.13it/s, Batch loss=83.3]
100% 68/68 [00:05<00:00, 12.95it/s]
The accuracy on the non masked validation set is 5.43
100% 68/68 [00:06<00:00, 10.48it/s]
The accuracy on the masked validation set is 2.26
======================
Found new best with avg accuracy: 3.84 on epoch 3. Saving model!!!
======================
Saving intermediate checkpoint...
Starting epoch 4...
100% 324/324 [02:32<00:00,  2.13it/s, Batch loss=103]
100% 68/68 [00:05<00:00, 13.05it/s]
The accuracy on the non masked validation set is 13.33
100% 68/68 [00:06<00:00, 10.31it/s]
The accuracy on the masked validation set is 4.85
======================
Found new best with avg accuracy: 9.09 on epoch 4. Saving model!!!
======================
Saving intermediate checkpoint...
Starting epoch 5...
100% 324/324 [02:30<00:00,  2.16it/s, Batch loss=73.7]
100% 68/68 [00:05<00:00, 13.07it/s]
The accuracy on the non masked validation set is 13.74
100% 68/68 [00:06<00:00, 10.41it/s]
The accuracy on the masked validation set is 7.62
======================
Found new best with avg accuracy: 10.68 on epoch 5. Saving model!!!
======================
Saving intermediate checkpoint...
Starting epoch 6...
100% 324/324 [02:34<00:00,  2.09it/s, Batch loss=86.5]
100% 68/68 [00:05<00:00, 12.77it/s]
The accuracy on the non masked validation set is 10.9
100% 68/68 [00:06<00:00, 10.58it/s]
The accuracy on the masked validation set is 6.85
Avg accuracy on epoch 6 is: 8.87
Saving intermediate checkpoint...
Starting epoch 7...
100% 324/324 [02:33<00:00,  2.11it/s, Batch loss=66.7]
100% 68/68 [00:05<00:00, 12.91it/s]
The accuracy on the non masked validation set is 10.4
100% 68/68 [00:06<00:00, 10.60it/s]
The accuracy on the masked validation set is 9.08
Avg accuracy on epoch 7 is: 9.74
Saving intermediate checkpoint...
Starting epoch 8...
100% 324/324 [02:30<00:00,  2.15it/s, Batch loss=66]
100% 68/68 [00:05<00:00, 13.14it/s]
The accuracy on the non masked validation set is 25.9
100% 68/68 [00:06<00:00, 10.51it/s]
The accuracy on the masked validation set is 10.46
======================
Found new best with avg accuracy: 18.18 on epoch 8. Saving model!!!
======================
Saving intermediate checkpoint...
Starting epoch 9...
100% 324/324 [02:33<00:00,  2.12it/s, Batch loss=58]
100% 68/68 [00:05<00:00, 13.00it/s]
The accuracy on the non masked validation set is 20.62
100% 68/68 [00:06<00:00, 10.58it/s]
The accuracy on the masked validation set is 12.11
Avg accuracy on epoch 9 is: 16.37
Saving intermediate checkpoint...
Starting epoch 10...
100% 324/324 [02:33<00:00,  2.11it/s, Batch loss=104]
100% 68/68 [00:05<00:00, 13.09it/s]
The accuracy on the non masked validation set is 23.63
100% 68/68 [00:06<00:00, 10.42it/s]
The accuracy on the masked validation set is 12.29
Avg accuracy on epoch 10 is: 17.96
Saving intermediate checkpoint...
Starting epoch 11...
100% 324/324 [02:33<00:00,  2.11it/s, Batch loss=86.1]
100% 68/68 [00:05<00:00, 12.79it/s]
The accuracy on the non masked validation set is 25.01
100% 68/68 [00:06<00:00, 10.54it/s]
The accuracy on the masked validation set is 12.22
======================
Found new best with avg accuracy: 18.61 on epoch 11. Saving model!!!
======================
Saving intermediate checkpoint...
Starting epoch 12...
100% 324/324 [02:30<00:00,  2.15it/s, Batch loss=90.2]
100% 68/68 [00:05<00:00, 12.91it/s]
The accuracy on the non masked validation set is 24.66
100% 68/68 [00:06<00:00, 10.43it/s]
The accuracy on the masked validation set is 11.82
Avg accuracy on epoch 12 is: 18.24
Saving intermediate checkpoint...
Starting epoch 13...
100% 324/324 [02:28<00:00,  2.18it/s, Batch loss=63.7]
100% 68/68 [00:05<00:00, 12.99it/s]
The accuracy on the non masked validation set is 26.47
100% 68/68 [00:06<00:00, 10.59it/s]
The accuracy on the masked validation set is 13.72
======================
Found new best with avg accuracy: 20.1 on epoch 13. Saving model!!!
======================
Saving intermediate checkpoint...
Starting epoch 14...
100% 324/324 [02:31<00:00,  2.14it/s, Batch loss=50.6]
100% 68/68 [00:05<00:00, 12.89it/s]
The accuracy on the non masked validation set is 20.17
100% 68/68 [00:06<00:00, 10.57it/s]
The accuracy on the masked validation set is 12.45
Avg accuracy on epoch 14 is: 16.31
Saving intermediate checkpoint...
Starting epoch 15...
100% 324/324 [02:31<00:00,  2.14it/s, Batch loss=38.8]
100% 68/68 [00:05<00:00, 12.92it/s]
The accuracy on the non masked validation set is 27.29
100% 68/68 [00:06<00:00, 10.50it/s]
The accuracy on the masked validation set is 11.67
Avg accuracy on epoch 15 is: 19.48
Saving intermediate checkpoint...
Starting epoch 16...
100% 324/324 [02:30<00:00,  2.16it/s, Batch loss=111]
100% 68/68 [00:05<00:00, 13.14it/s]
The accuracy on the non masked validation set is 27.33
100% 68/68 [00:06<00:00, 10.58it/s]
The accuracy on the masked validation set is 10.54
Avg accuracy on epoch 16 is: 18.94
Saving intermediate checkpoint...
Starting epoch 17...
100% 324/324 [02:31<00:00,  2.14it/s, Batch loss=70.5]
100% 68/68 [00:05<00:00, 13.09it/s]
The accuracy on the non masked validation set is 27.98
100% 68/68 [00:06<00:00, 10.62it/s]
The accuracy on the masked validation set is 9.44
Avg accuracy on epoch 17 is: 18.71
Saving intermediate checkpoint...
Starting epoch 18...
100% 324/324 [02:33<00:00,  2.11it/s, Batch loss=40]
100% 68/68 [00:05<00:00, 13.18it/s]
The accuracy on the non masked validation set is 38.28
100% 68/68 [00:06<00:00, 10.63it/s]
The accuracy on the masked validation set is 14.3
======================
Found new best with avg accuracy: 26.29 on epoch 18. Saving model!!!
======================
Saving intermediate checkpoint...
Starting epoch 19...
100% 324/324 [02:34<00:00,  2.10it/s, Batch loss=64.3]
100% 68/68 [00:05<00:00, 12.95it/s]
The accuracy on the non masked validation set is 36.77
100% 68/68 [00:06<00:00, 10.49it/s]
The accuracy on the masked validation set is 13.79
Avg accuracy on epoch 19 is: 25.28
Saving intermediate checkpoint...
Starting epoch 20...
100% 324/324 [02:32<00:00,  2.12it/s, Batch loss=62.8]
100% 68/68 [00:05<00:00, 13.21it/s]
The accuracy on the non masked validation set is 39.39
100% 68/68 [00:06<00:00, 10.44it/s]
The accuracy on the masked validation set is 12.32
Avg accuracy on epoch 20 is: 25.86
Saving intermediate checkpoint...
Starting epoch 21...
100% 324/324 [02:29<00:00,  2.16it/s, Batch loss=50.1]
100% 68/68 [00:05<00:00, 13.02it/s]
The accuracy on the non masked validation set is 39.53
100% 68/68 [00:06<00:00, 10.62it/s]
The accuracy on the masked validation set is 12.36
Avg accuracy on epoch 21 is: 25.94
Saving intermediate checkpoint...
Starting epoch 22...
100% 324/324 [02:33<00:00,  2.12it/s, Batch loss=40.1]
100% 68/68 [00:05<00:00, 13.05it/s]
The accuracy on the non masked validation set is 39.46
100% 68/68 [00:06<00:00, 10.52it/s]
The accuracy on the masked validation set is 11.85
Avg accuracy on epoch 22 is: 25.66
Saving intermediate checkpoint...
Starting epoch 23...
100% 324/324 [02:29<00:00,  2.17it/s, Batch loss=73.5]
100% 68/68 [00:05<00:00, 12.93it/s]
The accuracy on the non masked validation set is 40.2
100% 68/68 [00:06<00:00, 10.27it/s]
The accuracy on the masked validation set is 12.44
======================
Found new best with avg accuracy: 26.32 on epoch 23. Saving model!!!
======================
Saving intermediate checkpoint...
Starting epoch 24...
100% 324/324 [02:32<00:00,  2.13it/s, Batch loss=71]
100% 68/68 [00:05<00:00, 12.83it/s]
The accuracy on the non masked validation set is 40.94
100% 68/68 [00:06<00:00, 10.47it/s]
The accuracy on the masked validation set is 12.61
======================
Found new best with avg accuracy: 26.78 on epoch 24. Saving model!!!
======================
Saving intermediate checkpoint...
Starting epoch 25...
100% 324/324 [02:34<00:00,  2.09it/s, Batch loss=36.5]
100% 68/68 [00:05<00:00, 13.07it/s]
The accuracy on the non masked validation set is 42.33
100% 68/68 [00:06<00:00, 10.25it/s]
The accuracy on the masked validation set is 12.48
======================
Found new best with avg accuracy: 27.41 on epoch 25. Saving model!!!
======================
Saving intermediate checkpoint...
Starting epoch 26...
100% 324/324 [02:33<00:00,  2.11it/s, Batch loss=45.1]
100% 68/68 [00:05<00:00, 12.99it/s]
The accuracy on the non masked validation set is 47.74
100% 68/68 [00:06<00:00, 10.48it/s]
The accuracy on the masked validation set is 15.79
======================
Found new best with avg accuracy: 31.76 on epoch 26. Saving model!!!
======================
Saving intermediate checkpoint...
Starting epoch 27...
100% 324/324 [02:34<00:00,  2.09it/s, Batch loss=38.9]
100% 68/68 [00:05<00:00, 13.08it/s]
The accuracy on the non masked validation set is 55.0
100% 68/68 [00:06<00:00, 10.52it/s]
The accuracy on the masked validation set is 18.4
======================
Found new best with avg accuracy: 36.7 on epoch 27. Saving model!!!
======================
Saving intermediate checkpoint...
Starting epoch 28...
100% 324/324 [02:30<00:00,  2.15it/s, Batch loss=50.7]
100% 68/68 [00:05<00:00, 13.08it/s]
The accuracy on the non masked validation set is 59.62
100% 68/68 [00:06<00:00, 10.57it/s]
The accuracy on the masked validation set is 19.85
======================
Found new best with avg accuracy: 39.74 on epoch 28. Saving model!!!
======================
Saving intermediate checkpoint...
Starting epoch 29...
100% 324/324 [02:31<00:00,  2.13it/s, Batch loss=21]
100% 68/68 [00:05<00:00, 13.22it/s]
The accuracy on the non masked validation set is 52.99
100% 68/68 [00:06<00:00, 10.34it/s]
The accuracy on the masked validation set is 20.43
Avg accuracy on epoch 29 is: 36.71
Saving intermediate checkpoint...
Starting epoch 30...
100% 324/324 [02:31<00:00,  2.14it/s, Batch loss=35.1]
100% 68/68 [00:05<00:00, 13.14it/s]
The accuracy on the non masked validation set is 58.48
100% 68/68 [00:06<00:00, 10.28it/s]
The accuracy on the masked validation set is 23.17
======================
Found new best with avg accuracy: 40.83 on epoch 30. Saving model!!!
======================
Saving intermediate checkpoint...
Starting epoch 31...
100% 324/324 [02:32<00:00,  2.13it/s, Batch loss=54.1]
100% 68/68 [00:05<00:00, 13.05it/s]
The accuracy on the non masked validation set is 63.24
100% 68/68 [00:06<00:00, 10.62it/s]
The accuracy on the masked validation set is 24.33
======================
Found new best with avg accuracy: 43.79 on epoch 31. Saving model!!!
======================
Saving intermediate checkpoint...
Starting epoch 32...
100% 324/324 [02:31<00:00,  2.14it/s, Batch loss=41.7]
100% 68/68 [00:05<00:00, 13.15it/s]
The accuracy on the non masked validation set is 57.78
100% 68/68 [00:06<00:00, 10.61it/s]
The accuracy on the masked validation set is 24.1
Avg accuracy on epoch 32 is: 40.94
Saving intermediate checkpoint...
Starting epoch 33...
100% 324/324 [02:31<00:00,  2.14it/s, Batch loss=78.3]
100% 68/68 [00:05<00:00, 13.05it/s]
The accuracy on the non masked validation set is 67.86
100% 68/68 [00:06<00:00, 10.65it/s]
The accuracy on the masked validation set is 25.85
======================
Found new best with avg accuracy: 46.86 on epoch 33. Saving model!!!
======================
Saving intermediate checkpoint...
Starting epoch 34...
100% 324/324 [02:30<00:00,  2.15it/s, Batch loss=39.7]
100% 68/68 [00:05<00:00, 13.18it/s]
The accuracy on the non masked validation set is 68.37
100% 68/68 [00:06<00:00, 10.52it/s]
The accuracy on the masked validation set is 25.54
======================
Found new best with avg accuracy: 46.95 on epoch 34. Saving model!!!
======================
Saving intermediate checkpoint...
Starting epoch 35...
100% 324/324 [02:31<00:00,  2.13it/s, Batch loss=41.5]
100% 68/68 [00:05<00:00, 13.14it/s]
The accuracy on the non masked validation set is 67.95
100% 68/68 [00:06<00:00, 10.59it/s]
The accuracy on the masked validation set is 27.03
======================
Found new best with avg accuracy: 47.49 on epoch 35. Saving model!!!
======================
Saving intermediate checkpoint...
Starting epoch 36...
100% 324/324 [02:28<00:00,  2.17it/s, Batch loss=23.7]
100% 68/68 [00:05<00:00, 13.39it/s]
The accuracy on the non masked validation set is 69.21
100% 68/68 [00:06<00:00, 10.64it/s]
The accuracy on the masked validation set is 26.98
======================
Found new best with avg accuracy: 48.09 on epoch 36. Saving model!!!
======================
Saving intermediate checkpoint...
Starting epoch 37...
100% 324/324 [02:27<00:00,  2.19it/s, Batch loss=29.6]
100% 68/68 [00:05<00:00, 12.99it/s]
The accuracy on the non masked validation set is 62.37
100% 68/68 [00:06<00:00, 10.67it/s]
The accuracy on the masked validation set is 25.08
Avg accuracy on epoch 37 is: 43.72
Saving intermediate checkpoint...
Starting epoch 38...
100% 324/324 [02:29<00:00,  2.17it/s, Batch loss=31.8]
100% 68/68 [00:05<00:00, 13.21it/s]
The accuracy on the non masked validation set is 71.43
100% 68/68 [00:06<00:00, 10.49it/s]
The accuracy on the masked validation set is 28.53
======================
Found new best with avg accuracy: 49.98 on epoch 38. Saving model!!!
======================
Saving intermediate checkpoint...
Starting epoch 39...
100% 324/324 [02:31<00:00,  2.14it/s, Batch loss=35.4]
100% 68/68 [00:05<00:00, 13.05it/s]
The accuracy on the non masked validation set is 71.61
100% 68/68 [00:06<00:00, 10.45it/s]
The accuracy on the masked validation set is 28.0
Avg accuracy on epoch 39 is: 49.8
Saving intermediate checkpoint...
Starting epoch 40...
100% 324/324 [02:32<00:00,  2.13it/s, Batch loss=30.7]
100% 68/68 [00:05<00:00, 12.99it/s]
The accuracy on the non masked validation set is 74.38
100% 68/68 [00:06<00:00, 10.43it/s]
The accuracy on the masked validation set is 30.34
======================
Found new best with avg accuracy: 52.36 on epoch 40. Saving model!!!
======================
Saving intermediate checkpoint...