Missing modules when deploying the model

Community Cloud
,
1

hello i hope you are doing great !
i have been facing an issue when deploying my web application in the streamlit cloud service. The app uses the multiple ML libraries in order to perform the predictions. the core issue was the incompatibility between the scikit-learn and imbalanced-learn libraries. even though the code was executed successfuly in the local.

  • i specified the right version in requirements.txt file
  • i even specified the right python version in runtime.txt
  • however, somehow the streamlit compiler still uses the latest version of python with incompatible libraries.

idk if its a bug in the steamlit cloud service or it’s an issue from my end.

Note : i used some alternative libs of imbalanced_learn in order to make the process work. however i could not know the cause of the issue i just mentioned

2

Hi @hamzabenai06. Community Cloud doesn’t use runtime.txt. You need to choose the version of Python from your app’s settings from within the Community Cloud interface.

If choosing the correct version of Python from within Community Cloud doesn’t fix it, please can you share the logs which show the processing of dependencies and any errors raised during package processing and installation?