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

Metadata-populated model test and more #2

Closed
sayakpaul opened this issue Dec 5, 2020 · 3 comments
Closed

Metadata-populated model test and more #2

sayakpaul opened this issue Dec 5, 2020 · 3 comments

Comments

@sayakpaul
Copy link
Owner

@khanhlvg,

I have populated the fixed-shape models with metadata (notebook). Would you be able to help verify if those work in Android Studio? If they do then I will go ahead and publish with TensorFlow Hub.

Here's where you can find the metadata populated models.

Note: I was not able to complete the inference with integer quantized fixed-shape models (notebook).

@sayakpaul
Copy link
Owner Author

Also, I am able to do the full-integer quantization with tf-nightly (2.5.0-dev20201205) but it's not yet supported in TensorFlow 2.3.0.

@khanhlvg
Copy link

Yes it can be imported correctly. Please proceed with publishing to TF Hub :)

For the integer quantized model, it looks like there's a limitation in the TFLite integer quantization scheme. I'll share with the TFLite team for further investigation but I don't think it can be fixed soon. Please feel free to proceed without it.

@sayakpaul
Copy link
Owner Author

Thanks @khanhlvg.

Were you able to run inference in Android Studio as well? That way we could verify if the metadata population was indeed correct.

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