The model predicts test image data correctly but when implementing with anvil I don't get the output

You actually shared a link to run your app rather than to clone it. (You’re not alone.) Please try again to share the clone link. It would help us better diagnose your issue.