0

(Google Cloud Platform) I am creating a training model to distinguish between dog and cat, and hope that if I use a people image to test the model, instead of returning "cat" or "dog" it will return "other" label as the result.

But in fact, whatever Images I used (people, stars, tree, etc) it always return "cat" or "dog" as the results.

Has anyone experienced the same situation like me? Would you please kindly give me some advice?

Thank you!

Phuong
  • 1
  • 2

1 Answers1

0

It depends how you train the model.

if you want to get the result for "other", you need to train a label to predict other.

I recommend you to use Cloud Vision API from google. Which they have trained millions of labels already. https://cloud.google.com/vision/docs/quickstart-client-libraries

Bira
  • 4,531
  • 2
  • 27
  • 42