forked from EdjeElectronics/TensorFlow-Object-Detection-API-Tutorial-Train-Multiple-Objects-Windows-10
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Note.txt
176 lines (120 loc) · 7.7 KB
/
Note.txt
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
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
Docker用法
docker create --name smith-tensorflow -it --runtime=nvidia --network=host -v $HOME:/host usi/tensorflow
docker ps -a
docker start smith-tensorflow
docker exec -it smith-tensorflow bash
tutorial
#Python + Tensorflow
Setup:
> https://www.tensorflow.org/install/pip or https://www.tensorflow.org/install/
sudo apt update
sudo apt install python3-dev python3-pip
sudo pip3 install -U virtualenv # system-wide install
mkdir virtualenvs
cd virtualenvs
virtualenv --system-site-packages -p python3 tensorflow
cd ..
source ./virtualenvs/tensorflow/bin/activate
(tensorflow)$ pip install --upgrade pip
(tensorflow)$ pip install --upgrade tensorflow
# verify with
(tensorflow)$ python -c "import tensorflow as tf; tf.enable_eager_execution(); print(tf.reduce_sum(tf.random_normal([1000, 1000])))"
# Object Detection : https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/installation.md
(tensorflow)$ git clone https://github.com/tensorflow/tensorflow.git
(tensorflow)$ cd tensorflow
(tensorflow)$ git clone https://github.com/tensorflow/models.git
cd
(tensorflow)$ sudo apt-get install protobuf-compiler python-pil python-lxml python-tk
(tensorflow)$ pip install Cython
(tensorflow)$ pip install contextlib2
(tensorflow)$ pip install jupyter
(tensorflow)$ pip install matplotlib
##cocoapi
git clone https://github.com/cocodataset/cocoapi.git
cd cocoapi/PythonAPI
make
cp -r pycocotools tensorflow/models/research/
##protobuf
cd tensorflow/models/research/
protoc object_detection/protos/*.proto --python_out=.
#https://github.com/EdjeElectronics/TensorFlow-Object-Detection-API-Tutorial-Train-Multiple-Objects-Windows-10#3-gather-and-label-pictures
#(Note: TensorFlow occassionally adds new .proto files to the \protos folder. If you get an error saying ImportError: cannot import name 'something_something_pb2' , you may need to update the protoc command to include the new .proto files.)
#https://github.com/EdjeElectronics/TensorFlow-Object-Detection-API-Tutorial-Train-Multiple-Objects-Windows-10#3-gather-and-label-pictures
(tensorflow1) C:\tensorflow1\models\research> python setup.py build
(tensorflow1) C:\tensorflow1\models\research> python setup.py install
export PYTHONPATH=$PYTHONPATH:~/tensorflow/tensorflow-official/models/research/:~/tensorflow/tensorflow-official/models/research/slim:~/tensorflow/tensorflow-official/models
and add to .bashrc
#verify
(tensorflow)$ python object_detection/builders/model_builder_test.py
#in tensorflow1/models/research/object_detection
(tensorflow)$ jupyter notebook object_detection_tutorial.ipynb
#Customize object detection
How To Train an Object Detection Classifier for Multiple Objects Using TensorFlow (GPU) on Windows 10
https://github.com/EdjeElectronics/TensorFlow-Object-Detection-API-Tutorial-Train-Multiple-Objects-Windows-10
<~/tensorflow>
1.mkdir tensorflow-custom
2.cd tensorflow-custom
3.git clone https://github.com/EdjeElectronics/TensorFlow-Object-Detection-API-Tutorial-Train-Multiple-Objects-Windows-10.git
4.git clone https://github.com/tensorflow/models.git
5.copy code from TensorFlow-Object-Detection-API-Tutorial-Train-Multiple-Objects-Windows-10/ to model/research/object_detection
<~/tensorflow/tensorflow1/>
cp TensorFlow-Object-Detection-API-Tutorial model/research/object_detection
6.export PYTHONPATH="~/tensorflow/tensorflow1/models:~/tensorflow/tensorflow1/models/research/:~/tensorflow/tensorflow1/models/research/slim:$PYTHONPATH"
<~/tensorflow/tensorflow1/models/research/object_detection/TensorFlow-Object-Detection-API-Tutorial>
7.modify the training/labelmap.pbtxt
(tensorflow)$ vi training/labelmap.pbtxt
8.generate csv (tensorflow1)
#Because we will call generate_tfrecord.py later, which need csv
#parse the minx,miny,filename..... to a csv file
10.<~/tensorflow/tensorflow1/models/research/object_detection/TensorFlow-Object-Detection-API-Tutorial>
`(tensorflow)$ python xml_to_csv.py`
11.<~/tensorflow/tensorflow1/models/research/object_detection/TensorFlow-Object-Detection-API-Tutorial>
generate tfrecord (required by object detection)
# Modify label mapping in generate_tfrecord.py
(tensorflow)$ python generate_tfrecord.py
`(tensorflow)$ python generate_tfrecord.py --csv_input=images/train_labels.csv --image_dir=images/train --output_path=train.record`
`(tensorflow)$ python generate_tfrecord.py --csv_input=images/test_labels.csv --image_dir=images/test --output_path=test.record`
12. download model and modify pipeline config.
(this can also be download from https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/detection_model_zoo.md)
for example : faster_rcnn_inception_v2_coco_2018_01_28
Configure training
<~/tensorflow/tensorflow1/models/research/object_detection/>
`vi ./faster_rcnn_inception_v2_coco_2018_01_28/pipeline.config`
```
num_classes=6
....
....
fine_tune_checkpoint : "/home/smith/tensorflow/tensorflow2/models/research/object_detection/faster_rcnn_inception_v2_coco_2018_01_28/model.ckpt"
.....
....
input_path : "/home/smith/tensorflow/tensorflow2/models/research/object_detection/TensorFlow-Object-Detection-API-Tutorial/train.record"
label_map_path: "/home/smith/tensorflow/tensorflow2/models/research/object_detection/TensorFlow-Object-Detection-API-Tutorial/training/labelmap.pbtxt"
....
....
input_path : "/home/smith/tensorflow/tensorflow2/models/research/object_detection/TensorFlow-Object-Detection-API-Tutorial/test.record"
label_map_path: "/home/smith/tensorflow/tensorflow2/models/research/object_detection/TensorFlow-Object-Detection-API-Tutorial/training/labelmap.pbtxt"
```
> 注意一定要用絕對路徑，而且不支援 ~/ 的表示法
5.Run!
> UPDATE 9/26/18: As of version 1.9, TensorFlow has deprecated the "train.py" file and replaced it with "model_main.py" file.
> I haven't been able to get model_main.py to work correctly yet (I run in to errors related to pycocotools).
> Fortunately, the train.py file is still available in the /object_detection/legacy folder.
> Simply move train.py from /object_detection/legacy into the /object_detection folder and then continue following the steps below.
This creates a frozen_inference_graph.pb file in the \object_detection\inference_graph folder. The .pb file contains the object detection classifier.
(tensorflow)$ python ~/tensorflow/tensorflow-official/models/research/object_detection/train.py --logtostderr --train_dir=training/ --pipeline_config_path=training/faster_rcnn_inception_v2_pets.config
or
python ~/tensorflow/tensorflow-official/models/research/object_detection/legacy/train.py --logtostderr --train_dir=training/ --pipeline_config_path=training/faster_rcnn_inception_v2_pets.config
#Note
如果遇到
> File "object_detection/train.py", line 49, in <module>
> from object_detection.builders import dataset_builder
> ImportError: cannot import name 'dataset_builder'
或是
> ModuleNotFoundError: No module named 'nets'
之類的
請檢查PYTHONPATH
如果遇到
> File "/home/smith/tensorflow/tensorflow2/models/research/object_detection/utils/learning_schedules.py", line 160, in manual_stepping
> raise ValueError('First step cannot be zero.')
請參照 https://github.com/tensorflow/models/issues/3794
(tensorflow)$ python ~/tensorflow/tensorflow-official/models/research/object_detection/export_inference_graph.py --input_type image_tensor --pipeline_config_path training/faster_rcnn_inception_v2_pets.config --trained_checkpoint_prefix training/model.ckpt-XXXX --output_directory inference_graph