-
Notifications
You must be signed in to change notification settings - Fork 38
/
Copy pathNASNetMobile.log
65 lines (58 loc) · 4.46 KB
/
NASNetMobile.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
63
64
2020-03-06 19:51:48,020 __main__ INFO Getting dataframe with alt_label: False.
2020-03-06 19:51:48,087 __main__ INFO Running with test enabled.
2020-03-06 19:51:48,092 __main__ INFO Found 5 classes in data set.
2020-03-06 19:51:48,092 __main__ INFO Creating model with cnn base: NASNetMobile
2020-03-06 19:51:48,092 __main__ INFO batch size: 32, dense units 128, dropout: 0.2
2020-03-06 19:51:48,092 __main__ INFO learning rate: 0.0001, l2 penalty: 0.0001, freeze 200
2020-03-06 19:52:01,664 __main__ INFO Generating validation dataset.
2020-03-06 19:52:01,681 __main__ INFO Generating train dataset.
2020-03-06 19:52:01,691 __main__ INFO Class dict: {'Unknown': 0, 'eva_st_angel': 1, 'lindo_st_angel': 2, 'nico_st_angel': 3, 'nikki_st_angel': 4}
2020-03-06 19:52:01,691 __main__ INFO Number of training samples: 5839
2020-03-06 19:52:01,692 __main__ INFO Number of validation samples: 1459
2020-03-06 19:52:01,692 __main__ INFO Class weights: {4: 1.9443207126948776, 0: 1.0186697782963827, 2: 1.0, 1: 2.954314720812183, 3: 1.9617977528089887}
2020-03-06 19:52:01,692 __main__ INFO Steps per epoch: 182
2020-03-06 19:52:01,692 __main__ INFO Validation steps: 45
2020-03-06 19:52:01,692 __main__ INFO Starting pass 1.
2020-03-06 20:28:02,736 __main__ INFO Finished pass 1.
2020-03-06 20:28:15,957 __main__ INFO Starting pass 2 with learning rate: 1e-05
2020-03-06 22:04:52,825 __main__ INFO Finished pass 2.
2020-03-06 22:06:12,040 tensorflow WARNING From /home/lindo/.virtualenvs/szm/lib/python3.6/site-packages/tensorflow_core/python/ops/resource_variable_ops.py:1786: calling BaseResourceVariable.__init__ (from tensorflow.python.ops.resource_variable_ops) with constraint is deprecated and will be removed in a future version.
Instructions for updating:
If using Keras pass *_constraint arguments to layers.
2020-03-06 22:06:21,741 tensorflow INFO Assets written to: /home/lindo/develop/smart-zoneminder/person-class/train-results/NASNetMobile/1/assets
2020-03-06 22:06:28,802 __main__ INFO Exported SavedModel to /home/lindo/develop/smart-zoneminder/person-class/train-results/NASNetMobile
2020-03-06 22:06:28,802 __main__ INFO Evaluating model on 1825 test samples.
2020-03-06 22:06:50,890 __main__ INFO Classification report:
precision recall f1-score support
Unknown 0.86 0.80 0.83 534
eva_st_angel 0.85 0.80 0.82 166
lindo_st_angel 0.91 0.95 0.93 516
nico_st_angel 0.90 0.88 0.89 299
nikki_st_angel 0.85 0.93 0.88 310
accuracy 0.88 1825
macro avg 0.87 0.87 0.87 1825
weighted avg 0.88 0.88 0.88 1825
2020-03-06 22:06:50,897 __main__ INFO Confusion matrix:
[[427 17 39 19 32]
[ 17 132 2 4 11]
[ 19 0 491 4 2]
[ 16 3 9 264 7]
[ 16 4 0 3 287]]
2020-03-06 22:07:36,866 __main__ INFO Quantized tflite model saved to: /home/lindo/develop/smart-zoneminder/person-class/train-results/NASNetMobile-person-classifier-quant.tflite
2020-03-06 22:07:40,498 __main__ INFO Edge TPU model compilation results:
Edge TPU Compiler version 2.0.291256449
Model compiled successfully in 3414 ms.
Input model: /home/lindo/develop/smart-zoneminder/person-class/train-results/NASNetMobile-person-classifier-quant.tflite
Input size: 5.32MiB
Output model: /home/lindo/develop/smart-zoneminder/person-class/train-results/NASNetMobile-person-classifier-quant_edgetpu.tflite
Output size: 7.45MiB
On-chip memory available for caching model parameters: 6.31MiB
On-chip memory used for caching model parameters: 6.14MiB
Off-chip memory used for streaming uncached model parameters: 10.38KiB
Number of Edge TPU subgraphs: 1
Total number of operations: 756
Operation log: /home/lindo/develop/smart-zoneminder/person-class/train-results/NASNetMobile-person-classifier-quant_edgetpu.log
Model successfully compiled but not all operations are supported by the Edge TPU. A percentage of the model will instead run on the CPU, which is slower. If possible, consider updating your model to use only operations supported by the Edge TPU. For details, visit g.co/coral/model-reqs.
Number of operations that will run on Edge TPU: 754
Number of operations that will run on CPU: 2
See the operation log file for individual operation details.