Skip to content

beratkurar/textline_segmentation_using_fcn

Repository files navigation

Installations for running keras-theano on GPU

  1. Upgrade pip and install opencv2
cd ~
pip install --upgrade pip
pip install opencv-python
  1. Upgrade keras and set theano background
pip uninstall keras
pip install keras
vi ~/.keras/keras.json
  {
    "backend": "theano",
    "image_data_format": "channels_first",
    "floatx": "float32",
    "epsilon": 1e-07
  }
  1. Upgrade theano
pip uninstall theano
pip install theano==1.0.1
  1. Install pygpu which is necessary to run theano on GPU
git clone https://github.com/Theano/libgpuarray.git
cd libgpuarray
mkdir Build
cd Build
cmake .. -DCMAKE_BUILD_TYPE=Release
make
make install
cd ..
python3 setup.py build_ext -L /root/libgpuarray/lib -I /root/libgpuarray/include
python3 setup.py install
sudo ldconfig
  1. Run python on a specific GPU
THEANO_FLAGS=device=cuda0 python3

How to run the project

  1. Download challenging text line dataset

  2. Preprocess

python3 preprocess.py
  1. Train
THEANO_FLAGS=device=cuda0 python3
  1. Predict
python3 SlidingPredictor.py
  1. Postprocess
python3 directional_morphing.py

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages