# From tensorflow/models/research/
1. protoc object_detection/protos/*.proto --python_out=.
2. export PYTHONPATH=$PYTHONPATH:`pwd`:`pwd`/slim
# test
3. python3 object_detection/builders/model_builder_test.py

1. cd models/research/object_detection
2. python3 dataset_tools/create_pascal_tf_record.py --data_dir voc/VOCdevkit/ --year=VOC2012 --set=train --output_path=voc/pascal_train.record
3. python3 dataset_tools/create_pascal_tf_record.py --data_dir voc/VOCdevkit/ --year=VOC2012 --set=val --output_path=voc/pascal_val.record
4. cp data/pascal_label_map.pbtxt voc/
# 下载模型文件http://download.tensorflow.org/models/object_detection/faster_rcnn_inception_resnet_v2_atrous_coco_11_06_2017.tar.gz
并解压，解压后得到frozen_inference_graph.pb 、graph.pbtxt 、model.ckpt.data-00000-of-00001 、model.ckpt.index、model.ckpt.meta 5 个文件。在voc文件夹中新建一个 pretrained 文件夹，并将这5个文件复制进去。 
5. cp samples/configs/faster_rcnn_inception_resnet_v2_atrous_pets.config \
  voc/voc.config
6. 修改voc.config
7. python3 train.py --train_dir voc/train_dir/ --pipeline_config_path voc/voc.config
8. tensorboard --logdir voc/train_dir/
9. python3 export_inference_graph.py \ 
--input_type image_tensor \ 
--pipeline_config_path voc/voc.config \
--trained_checkpoint_prefix voc/train_dir/model.ckpt-7622 \
--output_directory voc/export/

I believe it is the same Python3 incompatibility that has crept up before (see #3443 ). The issue is with models/research/object_detection/utils/learning_schedules.py lines 167-169. Currently it is

rate_index = tf.reduce_max(tf.where(tf.greater_equal(global_step, boundaries),
                                      range(num_boundaries),
                                      [0] * num_boundaries))
Wrap list() around the range() like this:

rate_index = tf.reduce_max(tf.where(tf.greater_equal(global_step, boundaries),
                                     list(range(num_boundaries)),
                                      [0] * num_boundaries))
and you should be good to go. Mine is off and training.

@jwnsu make sure you change your pipeline.config from:

        manual_step_learning_rate {
          initial_learning_rate: 0.00001
          schedule {
            step: 0
            learning_rate: .0003
          }
          schedule {
            step: 900000
            learning_rate: .00003
          }
          schedule {
            step: 1200000
            learning_rate: .000003
          }
        }
      }
      momentum_optimizer_value: 0.9
    }
    use_moving_average: false
  }
to the new slightly different:

        manual_step_learning_rate {
          initial_learning_rate: 0.0003
          schedule {
            step: 900000
            learning_rate: .00003
          }
          schedule {
            step: 1200000
            learning_rate: .000003
          }
        }
      }
      momentum_optimizer_value: 0.9
    }
    use_moving_average: false
  }

1. sudo apt-get install python3-tk 
2. sudo apt-get install -f 

In python3 range is a generator object - it does not return a list. Convert it to a list before shuffling.

python setup.py build_ext --inplace
Traceback (most recent call last):
  File "setup.py", line 1, in <module>
    from setuptools import setup, Extension
ImportError: No module named setuptools
Makefile:3: recipe for target 'all' failed

sudo apt-get install python-setuptools

git clone https://github.com/pdollar/coco.git

cd coco/PythonAPI
make
sudo make install
sudo python setup.py install

before doing above steps install cython