Efficient Distributed SDCA
- Jeff Hajewski
- Mingrui Liu
The project requires the following to build:
- cmake
- C++ 14
Optionally, you may use the Ninja build system if you have it installed.
To run the project, first move into the build
directory (e.g., cd build
from project root
). From within the build directory, run the command
cmake ..
make
To run the executable, run:
./edsdca_main
If you have the Ninja bulid system installed, you can select Ninja via:
cmake -GNinja ..
ninja
Running the executable is the same as above.
There are also test executables that are created (depending on available
hardware). These are run in the same manner as edsdca_main
To use a different dataset in edsdca_main
, modify the path to the specified
dataset in the file src/edsdca/run_models.cc
. Make sure the new dataset has
a format similar to the other datasets. The main supported formats are:
- CSV
- LibSVM sparse representation