I'm using Filemaker, with Monkey Bread Software plugin's CoreML features, to find that it can only write to .mlmodelc.
Are these (.mlmodel = .mlmodelc) the same? If not, how do you generate a .mlmodelc using XCode.
Please let me know, thanks.
Differences in .mlmodel and .mlmodelc
What is an MLMODELC file? An MLMODELC file is a compiled machine learning model created by Apple Xcode. Xcode creates an MLMODELC file when a developer packages an app development project that includes an .MLMODEL file. The MLMODELC file is stored within the packaged app (an .APP or .IPA file).
More Information Using the Create ML component, Apple Xcode users can create, train, and embed machine learning models in macOS and iOS apps. During development, these models are saved as MLMODEL files.
When a developer packages an app that includes an MLMODEL file for distribution, XCode compiles the MLMODEL file to the MLMODELC format. This optimizes the model for use on an Apple device. MLMODELC files are typically stored in an app's Resources directory.
NOTE: In some cases, an app may instead download and compile a Create ML model dynamically when a user runs that app.
.mlmodel is an open model format (See https://apple.github.io/coremltools/mlmodel/Format/Model.html) that one can compose and edit.
.mlmodelc is compiled from .mlmodel or .mlpackage and contains Core ML specific representation of the model.
When the developer adds .mlmodel or .mlpackage to an application's project, Xcode automatically compiles it and includes the .mlmodelc in the application's bundle.
One can also manually compile a model from command line:
xcrun coremlcompiler compile MyModel.mlmodel /path/to/output/
Lastly, .mlpackage is a better version of .mlmodel. The key difference is that it can store the model architecture and weights in different files so that one can inspect and edit the model architecture and associated metadata without loading the weight files.