-
Notifications
You must be signed in to change notification settings - Fork 54
Running on gpatlas
dguest edited this page Sep 20, 2016
·
2 revisions
I've setup a virtual environment that should have everything you need to run on gpatlas.
- Open a clean session
- run
> source /home/dguest/public/lwtnn-gpatlas/setup-and-build.sh
You should see a few lines printed out, with *** Success! *** somewhere in the middle if everything works.
Internally this script checks out lwtnn, compiles, and then sets up a virtual environment for the required python packages.
To set up the virtual environment, run the following:
> source /home/dguest/public/venv/lwtnn/bin/activate
this should set up a virtual environment with the right version of python and the libraries you need to run the lwtnn converters. To leave the virtual environment just run type deactivate.
Once you've activated the virtual environment, you can run the tests in lwtnn/tests.
To test an arbitrary network, see Testing Your NN.