I think you should include your pre-trained model(s) for download (e.g. include a Google Drive link). Most people would want to play around with your app without having to train a model from scratch.
It would be useful to provide the option to run the model locally as well. The dependency on GCS might deter people from using this repo unless there are clear instructions (or at least reference to links) on setting up GCS and uploading the model to it. It's the same thing for the Android component; adding either detailed instructions on getting the app up and running or links to resources that show how to do that would be helpful.
I think you should include your pre-trained model(s) for download (e.g. include a Google Drive link). Most people would want to play around with your app without having to train a model from scratch.
It would be useful to provide the option to run the model locally as well. The dependency on GCS might deter people from using this repo unless there are clear instructions (or at least reference to links) on setting up GCS and uploading the model to it. It's the same thing for the Android component; adding either detailed instructions on getting the app up and running or links to resources that show how to do that would be helpful.
It'd make more sense to pass these variables as arguments to this script using something like
argparsehttps://github.com/hyang0129/food-class-app/blob/efb369a1bf28e564fe0cb0de96e3fa2ede0eb908/main.py#L45-L46