forked from shirleyatgithub/joker-network
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathtraining.log
62 lines (59 loc) · 5.77 KB
/
training.log
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
2020-11-03 06:02:22,566-INFO-Tensorflow version 2.3.1
2020-11-03 06:02:22,566-INFO-dataset path: TRAIN_DATA_FOLDER=/home/tobi/Downloads/trixsyDataset/training_dataset
2020-11-03 06:02:28,004-INFO-initializing model from joker_net_20201102-2031/
2020-11-03 06:02:28,540-INFO-Model: "sequential_1"
2020-11-03 06:02:28,540-INFO-_________________________________________________________________
2020-11-03 06:02:28,541-INFO-Layer (type) Output Shape Param #
2020-11-03 06:02:28,541-INFO-=================================================================
2020-11-03 06:02:28,541-INFO-conv1 (Conv2D) (None, 54, 54, 64) 7808
2020-11-03 06:02:28,541-INFO-_________________________________________________________________
2020-11-03 06:02:28,541-INFO-batch_normalization_2 (Batch (None, 54, 54, 64) 256
2020-11-03 06:02:28,541-INFO-_________________________________________________________________
2020-11-03 06:02:28,541-INFO-max_pooling2d_3 (MaxPooling2 (None, 26, 26, 64) 0
2020-11-03 06:02:28,542-INFO-_________________________________________________________________
2020-11-03 06:02:28,542-INFO-conv2 (Conv2D) (None, 26, 26, 64) 102464
2020-11-03 06:02:28,542-INFO-_________________________________________________________________
2020-11-03 06:02:28,542-INFO-batch_normalization_3 (Batch (None, 26, 26, 64) 256
2020-11-03 06:02:28,542-INFO-_________________________________________________________________
2020-11-03 06:02:28,543-INFO-max_pooling2d_4 (MaxPooling2 (None, 12, 12, 64) 0
2020-11-03 06:02:28,543-INFO-_________________________________________________________________
2020-11-03 06:02:28,543-INFO-conv3 (Conv2D) (None, 12, 12, 128) 73856
2020-11-03 06:02:28,543-INFO-_________________________________________________________________
2020-11-03 06:02:28,543-INFO-conv4 (Conv2D) (None, 12, 12, 128) 147584
2020-11-03 06:02:28,543-INFO-_________________________________________________________________
2020-11-03 06:02:28,543-INFO-max_pooling2d_5 (MaxPooling2 (None, 5, 5, 128) 0
2020-11-03 06:02:28,543-INFO-_________________________________________________________________
2020-11-03 06:02:28,544-INFO-flatten_1 (Flatten) (None, 3200) 0
2020-11-03 06:02:28,544-INFO-_________________________________________________________________
2020-11-03 06:02:28,544-INFO-fc8 (Dense) (None, 100) 320100
2020-11-03 06:02:28,544-INFO-_________________________________________________________________
2020-11-03 06:02:28,544-INFO-dropout_1 (Dropout) (None, 100) 0
2020-11-03 06:02:28,544-INFO-_________________________________________________________________
2020-11-03 06:02:28,544-INFO-output (Dense) (None, 2) 202
2020-11-03 06:02:28,544-INFO-=================================================================
2020-11-03 06:02:28,546-INFO-Total params: 652,526
2020-11-03 06:02:28,546-INFO-Trainable params: 652,270
2020-11-03 06:02:28,546-INFO-Non-trainable params: 256
2020-11-03 06:02:28,546-INFO-_________________________________________________________________
2020-11-03 06:02:28,546-INFO-making training generator
2020-11-03 06:02:28,546-INFO-making training generator
2020-11-03 06:02:31,229-INFO-making validation generator
2020-11-03 06:02:31,583-INFO-making test generator
2020-11-03 06:02:31,953-INFO-summary of /home/tobi/Downloads/trixsyDataset/training_dataset/train/: 180689 samples: 34.4% nonjoker, 65.6% joker
2020-11-03 06:02:31,953-INFO-summary of /home/tobi/Downloads/trixsyDataset/training_dataset/valid/: 22586 samples: 34.4% nonjoker, 65.6% joker
2020-11-03 06:02:31,953-INFO-summary of /home/tobi/Downloads/trixsyDataset/training_dataset/test/: 22587 samples: 34.4% nonjoker, 65.6% joker
2020-11-03 06:02:32,075-INFO-starting training
2020-11-03 07:51:37,607-INFO-Done with model.fit; history is
{'loss': [0.012133830226957798, 0.012278683483600616, 0.012239210307598114, 0.01161526795476675, 0.011500281281769276, 0.010900965891778469, 0.011205693706870079, 0.010979349724948406, 0.011105594225227833, 0.011781805194914341], 'categorical_accuracy': [0.9937627911567688, 0.9940007328987122, 0.9936686754226685, 0.9942830204963684, 0.9943327903747559, 0.9944766759872437, 0.994415819644928, 0.9944656491279602, 0.9943659901618958, 0.9943217635154724], 'val_loss': [0.03155764937400818, 0.013815019279718399, 0.014698613435029984, 0.013142377138137817, 0.02054186724126339, 0.01722804084420204, 0.013894373551011086, 0.02121536247432232, 0.01542014442384243, 0.016030065715312958], 'val_categorical_accuracy': [0.99039226770401, 0.9952625632286072, 0.9946869611740112, 0.9956167340278625, 0.9930930733680725, 0.9942885041236877, 0.9950854778289795, 0.9928717017173767, 0.9944213032722473, 0.9944213032722473]} and is saved as training_history.npy
2020-11-03 07:51:37,607-INFO-history.history.keys()=dict_keys(['loss', 'categorical_accuracy', 'val_loss', 'val_categorical_accuracy'])
2020-11-03 07:51:37,842-INFO-saving model to folder joker_net_20201103-0602
2020-11-03 07:51:38,961-INFO-converting model to tensorflow lite model
2020-11-03 07:51:39,417-INFO-saving tflite model as joker_net_20201103-0602.tflite
2020-11-03 07:51:39,419-INFO-evaluating accuracy
2020-11-03 07:51:53,523-INFO-On test set /home/tobi/Downloads/trixsyDataset/training_dataset/test/ loss=0.323, acc=0.9188
2020-11-03 07:52:07,029-INFO-**** final test set balanced accuracy: 93.790\% (chance would be 50\%)
Confusion matrix nonjoker/joker:
[[12982 1830]
[ 5 7770]]
2020-11-03 07:52:07,029-INFO-**** done training after 109.7m; model saved in joker_net_20201103-0602 and joker_net_20201103-0602.tflite.
See training.log for logging output for this run.