pycaret / deployment-heroku Goto Github PK

Built a pipeline and catboost model using pycaret 2.2.3 and streamlit 0.73.1. Followed the example here, paring down modules to what is shown in app.py. Added a .slugignore etc. to further try and reduce the slug size which is 547MB and exceeds the 500MB limit for Free tier dynos in Heroku. On the streamlit repo, the advice was to use an earlier version of PyCaret.

Have you tried to run this repo with 2.2.x using streamlit and heroku? Has the footprint of PyCaret increased dramatically over the past 6-7 months? Any tips/tricks to slim down the slug?

I have tried now for +15 hours of deploying a pycaret to Heroku and it feels simple impossible. I went to this repo, cloned it, connected to Heroku and pressed deploy branch and got the same exact errors. Is this out dated or is there no way of deploying pycaret..

The error I get is with Numba and llvmlite:

Building wheel for numba (setup.py): finished with status 'error'
Building wheel for llvmlite (setup.py): finished with status 'error'

I'm so lost for hope and don't know what to do...