Keras installation

The reason that I have installed CUDA is to use Keras.
Keras is Theano-based Deep Learning library.


Installation process of Keras is simple compared to that of CUDA

1. Check your graphic card is available for GPU

2. Install CUDA
CUDA installation(korean) : http://junya906.blogspot.kr/2015/10/test.html

3. Install NumPy & SciPy

$ sudo apt-get install python-numpy python-scipy

4. Install PyYAML & Cython (pip is installed first)

$ pip install --user PyYAML
$ pip install cython

5. Install BLAS & HDF5 & Git

(blas)
$ sudo apt-get install libblas-dev checkinstall
$ sudo apt-get install libblas-doc checkinstall
$ sudo apt-get install liblapacke-dev checkinstall
$ sudo apt-get install liblapack-doc checkinstall
(hdf5)
$ sudo apt-get install libhdf5-serial-dev
(git)
$ sudo apt-get install git

6. Install Theano

(For Ubuntu 11.10 through 14.04)
$ sudo apt-get install python-dev python-nose g++ libopenblas-dev
$ sudo pip install Theano
(For Ubuntu 11.04)
$ sudo apt-get install python-dev python-nose g++ libatlas3gf-base libatlas-dev
$ sudo pip install Theano

7. Install Keras

$ sudo pip install keras

8. Download example codes & dataset and Test


ref:
http://keras.io/#installation
http://deeplearning.net/software/theano/install.html#install