# Create venv directory for essential tensorflow support
# This only needs to be run by once
./bootstrap.sh
source venv/bin/activate # Enable python environment with tensorflow
Now the unit tests can be started with ./Rod/test.py. There is no visual results by far, however.
bootstrap.sh will install AVX2-enabled tensorflow package if supported. Please use systems with AVX2 support for best performance.
Check AVX2 support: run grep avx2 /proc/cpuinfo in command line. If there
are outputs from this command, then AVX2 is supported by you CPU. Otherwise
nothing will be printed.
For visualization of the motion data generated into /tmp/tftest# directories
(# is a number) by test_M.py, please run .BlenderVis.py --simdir=/tmp/tftest# under Rod/ directory.
Note: You don't need to call test_M.py explicitly if test.py has been
executed, since test.py runs test_M.py as part of unit tests.