DeepVision can load any image classification Core ML model and execute it in real-time on video from the camera of your device.
You can start by trying the built-in model, SqueezeNet, which is only 5 MB in size and can detect 1000 categories of objects.
DeepVision will load the model and show you the most probable result predicted by the model. It can also read the information embedded in the model like the creator, license or description.
Note: You can precompile your models to load them faster in DeepVision. Use the command “xcrun coremlcompiler” in the Terminal of your Mac with Xcode 9 installed to compile Core ML models (.mlmodel) into compiled packages (.mlmodelc).
DeepVision is also available for iOS, check it out on the App Store!