4 ms·

Any idea if this can be used with Jupyter? I have a bunch of notebooks that take 4-8 hours to run. This could potentially make my life much easier.

• kzrdude 5y ago
  No reason not to try it
  • zibzab 5y ago
    I did, but I see absolutely no performance improvements. Maybe notebooks require something more?
    • kzrdude 5y ago
      I don't know your setup but regular jupyterlab notebook execution is really exactly like python in almost everything, there's not much that should be different.
• prionassembly 5y ago
  You can always export jupyter notebooks to straight .py files. Also: how in th
• napetrov 5y ago
  yes, notebooks have no difference for sklernex You likely not exporting algorithms from scikit learn again after patching - i.e. patch call should be made prior to import. Note there are kaggle notebooks that showcase same optimizations - https://www.kaggle.com/napetrov/tps04-svm-with-intel-extension-for-scikit-learn https://www.kaggle.com/napetrov/tps04-svm-with-intel-extensi... They are also basically notebooks