You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
With a few code changes, VCM can use Python3 (in the virtualenv named 'divime' installed at provision time, configured by environment.yml) specifically the Torch package installed already there for Yunitator & Noisemes to use... thus eliminating the need to install Torch twice (once for Python2, once for Python3).
This will make the VM slightly smaller & faster to set up. Slightly.
Things that'd have to happen:
source activate divime in the VCM run script
don't change PATH in VCM run script
switch python2 to python3 in the VCM run script
add the gerbils, I mean pandas package to environment.yml (needed by vcm)
eliminate the lines in bootstrap.sh that install torch and cuda
The text was updated successfully, but these errors were encountered:
With a few code changes, VCM can use Python3 (in the virtualenv named 'divime' installed at provision time, configured by environment.yml) specifically the Torch package installed already there for Yunitator & Noisemes to use... thus eliminating the need to install Torch twice (once for Python2, once for Python3).
This will make the VM slightly smaller & faster to set up. Slightly.
Things that'd have to happen:
source activate divime
in the VCM run scriptThe text was updated successfully, but these errors were encountered: