Dockerfile for creating a docker image with default configurations for Deep Learning tasks, including:
- Ubuntu 16.04
- Nvidia CUDA 8.0 + cuDNN 5.0 (hub)
- Basic Tools
- cmake
- curl, wget
- git
- nano, vim
- unzip, tar
- ffmpeg
- Languages compilers
- C++ (g++)
- Python 2.7
- Python 3.5
- Fortran (required by OpenCV)
- Python Machine Learning Libraries
- scipy
- nose
- scikit-image
- matplotlib
- pandas
- sklearn
- sympy
- OpenCV 3.x (compiled with)
- CUDA
- CUBLAS Support
- V4L
- TBB
- PTHREADS
- OpenGL
- OpenCL
- TIFF
- GDAL
- VTK
- IPP
- cstripes
- eigen
- Python 3.5 bindings
- Tensorflow
- Version: 1.2.1
- GPU Support: CUDA 8.0
- Python: 3.5
- PyTorch:
- Version: 0.1.12
- GPU Support: CUDA 8.0
- Python: 3.5
- Keras
- Version: 2.0.6
- Theano: YES (v0.9)
- Tensorflow: YES (1.2.1)