android-bootstrap VS optical-breacher

This is interesting. The closest I can compare it to is lobe.ai.

lobe.ai says object detection is coming soon

I need urgent help please!!! I've just installed the new Version of lobe.ai on my MAC and now, after it has finished, the prediction rate has decreased from more than 90% to 50% :-( :-(

Using lobe.ai 0.10.1130.5 I successfully trained using my Webcam Logitech C920. The camera turned live, and I could take individual and rapid-snap photos. But after proceeding to 'Use', the camera button does show, but nothing happens when I press it, not does hovering raise a floating menu. What am I doing wrong?

Found a fix here? There is some bugs in the lobe.ai code:

There is this android-bootstrap https://github.com/lobe/android-bootstrap. In that "getting started" I did everything but I feel like there are steps missing between 4 and 5. The "Run" button is grayed-out.

I can run the lobe.ai web version locally.

lobe.ai

Are you copying both the saved_model.tflite model and signature.json files? https://github.com/lobe/android-bootstrap#get-started

I have no plans on open sourcing my app, but here are bits and pieces I used to build upon :)

This one is better, it has OCR, you can upload screenshots or use your phone's camera.

Get a cyberdeck with big enough buffer, the Stephenson Tech Mk.2 has 7 and only requires 12 street cred, best early game cyberdeck imo. If you have problems finding solutions you can use this to help, just need to point you phone's camera at the screen.

You can use this to solve breach protocol for you, just need to point your phone camera at the screen.

I got every piece of legendary quickhack from access point drops. Get the perk that automatically highlights access points. (Note there is a bug with drop chance get this mod to fix it on PC) If you are tired of solving the puzzle manually you can use this to help you. Also there are two side quests that can give you quickhacks: KOLD MIRAGE and Spellbound.