coreml

It would be great to have FBeta, F2, or F0.5 metrics to be implemented without the need for a custom metric class defined by user.

catboost version: 0.26

🐞Describe the bug

Converting Frontend cannot be 100% converted when pytorch model is converted to MLmodel. I get a warning

Trace

`TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other

Link

https://github.com/tucan9389/MLProblem-MLFramework-for-iOS-template-project

Is it possible to implement an overlay to show where the qrcode is in the camera feed? just like this https://www.appcoda.com/barcode-reader-swift/

Was wondering how can I get the videoPreviewLayer of the camera so I can use transformedMetadataObject