I converted a two-way classification model (uses a probability from 0 to 1) with Keras TensorFlow 2.3.1 and used coremltools to convert to mlmodel. I added this to my app in Xcode to use and I can't get a result out of the output. I am using ios 14.3 with Mac big sur 11.1.
I attached my app to GitHub to be debugged at https://github.com/jbkey730/SeeFood-iOS13-Completed.git . Its originally from a class but I was testing out my models first with this app. I'm running this on an iPhone 8 with ios 14.3.
I looked up some other information about this saying it was because of the upgrade to ios 14 from ios 13. Is there a fix to this bug yet and how do I implement it in my app?
The error code in total is...
2021-01-09 19:22:32.231074-0600 SeeFood-CoreML[818:53598] [Camera] Failed to read exposureBiasesByMode dictionary: Error Domain=NSCocoaErrorDomain Code=4864 "*** -[NSKeyedUnarchiver _initForReadingFromData:error:throwLegacyExceptions:]: data is NULL" UserInfo={NSDebugDescription=*** -[NSKeyedUnarchiver _initForReadingFromData:error:throwLegacyExceptions:]: data is NULL}
2021-01-09 19:22:35.192738-0600 SeeFood-CoreML[818:53598] Metal API Validation Enabled
0.016954016
I would like to get a from this line of code output from 0 to 1 of my output here...it currently is id:0x0
let topResult = results.first
so the topResult would be something from 0 to 1 so I can make a threshold at 0.5 to classify if it's one of two types of seed.
question from:
https://stackoverflow.com/questions/65649336/debugging-keras-mlmodel-in-xcode-camera-failed-to-read-exposurebiasesbymode-d 与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…