Error when trying to train about ocrsegment

[watersink]

this wrong is simple, just because you don't have this file label.txt in the folder uw3，perhaps you shuould make one label.txt yourself

[watersink]

you should write data label in this file,one line by one line

@watersink I copied K47LYBN.framed.png & K47LYBN.lines.png to the uw3 folder.
Then I created label.txt file, containing:

./K47LYBN.framed.png ./K47LYBN.lines.png

But when running train_test.py an error message appears:

home@home-lnx:~/ocrsegment$ python3 train_test.py 
2018-09-19 03:05:57.689610: I tensorflow/core/platform/cpu_feature_guard.cc:140] Your CPU supports instructions that this TensorFlow binary was not compiled to use: SSE4.1 SSE4.2 AVX
2018-09-19 03:06:00.357113: W tensorflow/core/framework/allocator.cc:101] Allocation of 1113262080 exceeds 10% of system memory.
2018-09-19 03:06:02.264820: W tensorflow/core/framework/allocator.cc:101] Allocation of 1113262080 exceeds 10% of system memory.
2018-09-19 03:06:17.398615: W tensorflow/core/framework/allocator.cc:101] Allocation of 278236160 exceeds 10% of system memory.
2018-09-19 03:06:27.941151: W tensorflow/core/framework/allocator.cc:101] Allocation of 556472320 exceeds 10% of system memory.
terminate called after throwing an instance of 'std::bad_alloc'
  what():  std::bad_alloc
Aborted (core dumped)

Is this error caused by tensorflow itself, or is it because I don't have a GPU.
Can I train using only CPU?

Here are the commands that I used to install tensorflow and the other requirements:

conda create -n python3 python=3.6 pillow tensorflow=1.8
conda install -n python3 -c menpo opencv3

[watersink]

pip3 install tensorflow-gpu==1.8.0
pip3 install tensorflow==1.8.0
pip3 install opencv-python

@watersink I tested training on my other system with 30GB ram and Quadro M4000 graphics card and it worked there.
Thanks!

[watersink]

@mrocr I have one qq group(742072640)，hope you can join，so that the discussion will be easier and faster.