rflow - Pipeline Interface for TensorFlow in R
This is an experimental and work-in-progress R library for easily using Google TensorFlow, inspired by Google skflow package. This package aims to help R users to easily build popular deep learning models as well as customized models with flexible and desired deep architectures. Note that the style of the way the package has been developed is new and experimental. Any feedback would be appreciated.
Dependencies
Python Packages (get pip from here first)
- sklearn (install via
pip install -U scikit-learn
) - pandas (install via
pip install pandas
) - tensorflow (instruction here)
- skflow (install via
pip install git+git://github.com/google/skflow.git
)
R Packages
- rPython
Installation (via Github after successfully installed the above dependencies)
devtools::install_github('terrytangyuan/rflow')
Check out demos for available usages. Issues and pull requests are more than welcomed.
Contact terrytangyuan@gmail.com for questions/suggestions/collaborations.