Tensorflow安装object detection API（新）

1、
安装tensorflow1.14：
pip install tensorflow==1.14 -i https://pypi.tuna.tsinghua.edu.cn/simple
2、
安装对应版本的keras：
pip
install    keras==2.2.5
3、
安装protobuf：
https://github.com/google/protobuf/releases/tag/v3.4.0
下载得到的文件：将bin文件里面的protoc.exe复制到D:\Program
Files (x86)\anaconda\Lib\site-packages\tensorflow\models\research目录下;
执行以下命令：

 

protoc object_detection/protos/anchor_generator.proto --python_out=.
protoc object_detection/protos/argmax_matcher.proto --python_out=.
protoc object_detection/protos/bipartite_matcher.proto --python_out=.
protoc object_detection/protos/box_coder.proto --python_out=.
protoc object_detection/protos/box_predictor.proto --python_out=.
protoc object_detection/protos/calibration.proto --python_out=.
protoc object_detection/protos/eval.proto --python_out=.
protoc object_detection/protos/faster_rcnn.proto --python_out=.
protoc object_detection/protos/faster_rcnn_box_coder.proto --python_out=.
protoc object_detection/protos/flexible_grid_anchor_generator.proto --python_out=.
protoc object_detection/protos/graph_rewriter.proto --python_out=.
protoc object_detection/protos/grid_anchor_generator.proto --python_out=.
protoc object_detection/protos/hyperparams.proto --python_out=.
protoc object_detection/protos/image_resizer.proto --python_out=.
protoc object_detection/protos/input_reader.proto --python_out=.
protoc object_detection/protos/keypoint_box_coder.proto --python_out=.
protoc object_detection/protos/losses.proto --python_out=.
protoc object_detection/protos/matcher.proto --python_out=.
protoc object_detection/protos/mean_stddev_box_coder.proto --python_out=.
protoc object_detection/protos/model.proto --python_out=.
protoc object_detection/protos/multiscale_anchor_generator.proto --python_out=.
protoc object_detection/protos/optimizer.proto --python_out=.
protoc object_detection/protos/pipeline.proto --python_out=.
protoc object_detection/protos/post_processing.proto --python_out=.
protoc object_detection/protos/preprocessor.proto --python_out=.
protoc object_detection/protos/region_similarity_calculator.proto --python_out=.
protoc object_detection/protos/square_box_coder.proto --python_out=.
protoc object_detection/protos/ssd.proto --python_out=.
protoc object_detection/protos/ssd_anchor_generator.proto --python_out=.
protoc object_detection/protos/string_int_label_map.proto --python_out=.
protoc object_detection/protos/target_assigner.proto --python_out=.
protoc object_detection/protos/train.proto --python_out=.

4、
在用户变量里面添加PYTHONPATH：
D:\Program Files
(x86)\anaconda\Lib\site-packages\tensorflow\models\research
D:\Program Files
(x86)\anaconda\Lib\site-packages\tensorflow\models\research\slim
D:\Program Files
(x86)\anaconda\Lib\site-packages\tensorflow\models\research\object_detection
在系统变量Path同样添加以上路径
5、 在路径D:\Program Files
(x86)\anaconda\Lib\site-packages\tensorflow\models\research下添加测试代码：python object_detection/builders/model_builder_test.py
若出现报错：ModuleNotFoundError: No module named ‘nets’
将slim文件下的nets复制到object_detection/builders文件下
运行成功：

注意：对于GPU版本的tensorflow，亲测的tensorflow-gpu1.14可以用，1.14以下的试过用不了。

tensorflow-gpu1.14对应CUDA10和cuDNN7.4