-
Notifications
You must be signed in to change notification settings - Fork 38
/
Copy pathResNet50.log
65 lines (58 loc) · 4.43 KB
/
ResNet50.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
2020-03-07 07:04:48,979 __main__ INFO Getting dataframe with alt_label: False.
2020-03-07 07:04:49,045 __main__ INFO Running with test enabled.
2020-03-07 07:04:49,048 __main__ INFO Found 5 classes in data set.
2020-03-07 07:04:49,048 __main__ INFO Creating model with cnn base: ResNet50
2020-03-07 07:04:49,048 __main__ INFO batch size: 32, dense units 128, dropout: 0.2
2020-03-07 07:04:49,048 __main__ INFO learning rate: 0.0001, l2 penalty: 0.0001, freeze 40
2020-03-07 07:04:53,355 __main__ INFO Generating validation dataset.
2020-03-07 07:04:53,371 __main__ INFO Generating train dataset.
2020-03-07 07:04:53,381 __main__ INFO Class dict: {'Unknown': 0, 'eva_st_angel': 1, 'lindo_st_angel': 2, 'nico_st_angel': 3, 'nikki_st_angel': 4}
2020-03-07 07:04:53,381 __main__ INFO Number of training samples: 5839
2020-03-07 07:04:53,381 __main__ INFO Number of validation samples: 1459
2020-03-07 07:04:53,381 __main__ INFO Class weights: {4: 1.9443207126948776, 0: 1.0186697782963827, 2: 1.0, 1: 2.954314720812183, 3: 1.9617977528089887}
2020-03-07 07:04:53,381 __main__ INFO Steps per epoch: 182
2020-03-07 07:04:53,382 __main__ INFO Validation steps: 45
2020-03-07 07:04:53,382 __main__ INFO Starting pass 1.
2020-03-07 07:31:10,629 __main__ INFO Finished pass 1.
2020-03-07 07:31:13,737 __main__ INFO Starting pass 2 with learning rate: 1e-05
2020-03-07 09:33:12,492 __main__ INFO Finished pass 2.
2020-03-07 09:33:32,609 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.
2020-03-07 09:33:35,278 tensorflow INFO Assets written to: /home/lindo/develop/smart-zoneminder/person-class/train-results/ResNet50/1/assets
2020-03-07 09:33:36,991 __main__ INFO Exported SavedModel to /home/lindo/develop/smart-zoneminder/person-class/train-results/ResNet50
2020-03-07 09:33:36,992 __main__ INFO Evaluating model on 1825 test samples.
2020-03-07 09:33:44,811 __main__ INFO Classification report:
precision recall f1-score support
Unknown 0.86 0.87 0.86 534
eva_st_angel 0.85 0.81 0.83 166
lindo_st_angel 0.94 0.94 0.94 516
nico_st_angel 0.94 0.87 0.90 299
nikki_st_angel 0.84 0.91 0.87 310
accuracy 0.89 1825
macro avg 0.89 0.88 0.88 1825
weighted avg 0.89 0.89 0.89 1825
2020-03-07 09:33:44,818 __main__ INFO Confusion matrix:
[[464 14 20 10 26]
[ 15 134 3 3 11]
[ 22 1 486 4 3]
[ 16 4 5 261 13]
[ 23 4 1 1 281]]
2020-03-07 09:34:14,691 __main__ INFO Quantized tflite model saved to: /home/lindo/develop/smart-zoneminder/person-class/train-results/ResNet50-person-classifier-quant.tflite
2020-03-07 09:34:15,895 __main__ INFO Edge TPU model compilation results:
Edge TPU Compiler version 2.0.291256449
Model compiled successfully in 985 ms.
Input model: /home/lindo/develop/smart-zoneminder/person-class/train-results/ResNet50-person-classifier-quant.tflite
Input size: 23.56MiB
Output model: /home/lindo/develop/smart-zoneminder/person-class/train-results/ResNet50-person-classifier-quant_edgetpu.tflite
Output size: 23.11MiB
On-chip memory available for caching model parameters: 6.25MiB
On-chip memory used for caching model parameters: 6.25MiB
Off-chip memory used for streaming uncached model parameters: 16.69MiB
Number of Edge TPU subgraphs: 1
Total number of operations: 78
Operation log: /home/lindo/develop/smart-zoneminder/person-class/train-results/ResNet50-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: 76
Number of operations that will run on CPU: 2
See the operation log file for individual operation details.