Prediction results are different in app.roboflow vs serverless Hosted API call

*I am experimenting with a simple model in preparation for a bigger project. This simple project I trained using roboflow instant. See the Project below. When I use the β€œ* Sign in to Roboflow *and upload a picture, it does a pretty good job at predicting the hats but when I use β€œ*https://serverless.roboflow.com/project1-initw/hat-detector-fvaq7-instant-3” then I am getting way more predictions (close to 100+). It’s not able to do good prediction using serverless hosted API method.

  • Project Type:
  • Operating System & Browser: Chrome Browser
  • Project Universe Link or Workspace/Project ID: see above

Can you please help me figure out why I am running into this issue. Need some quick help please.

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.