# 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
二、训练新的模型
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/
- ValueError: Tried to convert 't' to a tensor and failed. Error: Argument must be a dense tensor: range(0, 3) - got shape [3], but wanted [].
pytho3兼容性问题 参见
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
}
-
错误：tensorflow.python.framework.errors_impl.InternalError: Dst tensor is not initialized. 分析：出现这个错误一般是GPU内存耗尽。 解决办法：运行程序之前，先运行export CUDA_VISIBLE_DEVICES=0
-
ImportError: No module named '_tkinter', please install the python3-tk package 解决方法：
1. sudo apt-get install python3-tk
2. sudo apt-get install -f
- TypeError: 'range' object does not support item assignment
In python3 range is a generator object - it does not return a list. Convert it to a list before shuffling.
- ImportError: No module named setuptools
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
解决方法：
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