Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Results in ios and android is different for a same model #7

Open
viuts opened this issue May 29, 2019 · 1 comment
Open

Results in ios and android is different for a same model #7

viuts opened this issue May 29, 2019 · 1 comment

Comments

@viuts
Copy link

viuts commented May 29, 2019

I am running example app on ios/android devices, using same model, same input image in both side, but the results is not same.

Generally the sequence of topN for Mobilenet is same, but the confidence percentage is a bit off, like testing a train image on android it says 69% but in ios it says 73%

I am also testing on some custom model on image classification as well, we result is more off in that case. In android it says 51%/49% and on ios it says 90%/10%

Is it the code of transforming the image to bitmap is a bit different on ios/android so the result is different?

@SJCaldwell
Copy link

Seeing similar things on my end, iOS prediction is significantly better than Android prediction for the same images. Find anymore info on this @viuts?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants