Post

Replies

Boosts

Views

Activity

Max 16k images for Image Classifier training????
I'm hitting a limit when trying to train an Image Classifier. It's at about 16k images (in line with the error info) - and it gives the error: IOSurface creation failed: e00002be parentID: 00000000 properties: { IOSurfaceAllocSize = 529984; IOSurfaceBytesPerElement = 4; IOSurfaceBytesPerRow = 1472; IOSurfaceElementHeight = 1; IOSurfaceElementWidth = 1; IOSurfaceHeight = 360; IOSurfaceName = CoreVideo; IOSurfaceOffset = 0; IOSurfacePixelFormat = 1111970369; IOSurfacePlaneComponentBitDepths = ( 8, 8, 8, 8 ); IOSurfacePlaneComponentNames = ( 4, 3, 2, 1 ); IOSurfacePlaneComponentRanges = ( 1, 1, 1, 1 ); IOSurfacePurgeWhenNotInUse = 1; IOSurfaceSubsampling = 1; IOSurfaceWidth = 360; } (likely per client IOSurface limit of 16384 reached) I feel like I was able to use more images than this before upgrading to Sonoma - but I don't have the receipts.... Is there a way around this? I have oodles of spare memory on my machine - it's using about 16gb of 64 when it crashes... code to create the model is let parameters = MLImageClassifier.ModelParameters(validation: .dataSource(validationDataSource), maxIterations: 25, augmentation: [], algorithm: .transferLearning( featureExtractor: .scenePrint(revision: 2), classifier: .logisticRegressor )) let model = try MLImageClassifier(trainingData: .labeledDirectories(at: trainingDir.url), parameters: parameters) I have also tried the same training source in CreateML, it runs through 'extracting features', and crashes at about 16k images processed. Thank you
1
0
176
Oct ’24
New Vison api - CoreML - "The VNDetectorProcessOption_ScenePrints required option was not found"
I'm trying to run a coreML model. This is an image classifier generated using: let parameters = MLImageClassifier.ModelParameters(validation: .dataSource(validationDataSource), maxIterations: 25, augmentation: [], algorithm: .transferLearning( featureExtractor: .scenePrint(revision: 2), classifier: .logisticRegressor )) let model = try MLImageClassifier(trainingData: .labeledDirectories(at: trainingDir.url), parameters: parameters) I'm trying to run it with the new async Vision api let model = try MLModel(contentsOf: modelUrl) guard let modelContainer = try? CoreMLModelContainer(model: model) else { fatalError("The model is missing") } let request = CoreMLRequest(model: modelContainer) let image = NSImage(named:"testImage")! let cgImage = image.toCGImage()! let handler = ImageRequestHandler(cgImage) do { let results = try await handler.perform(request) print(results) } catch { print("Failed: \(error)") } This gives me Failed: internalError("Error Domain=com.apple.Vision Code=7 "The VNDetectorProcessOption_ScenePrints required option was not found" UserInfo={NSLocalizedDescription=The VNDetectorProcessOption_ScenePrints required option was not found}") Please help! Am I missing something?
1
0
223
Oct ’24
Can you specify port for NWListener.Service without creating an NWListener on that port
I'm running a webserver for a specific service on port 8900 I'm using telegraph to run the webserver, so that opens and claims the port. I also want to advertise the service on bonjour - ideally with the correct port. This is trivial with NetService - but that's deprecated, so I should probably move to the Network framework. I can advertise without specifying a port listener = try NWListener(service: service, using: .tcp) but, then my service broadcasts addresses with port:61443 I can advertise using listener = try NWListener(using: .tcp, on: <myport>) however, that fails in my use case because (unsurprisingly) the Listener isn't able to get the port (my server already has it) Is this just a gap in the new API, or am I missing something?
1
0
656
Nov ’23
How to disable position animation in SwiftUI when button style is animated
I have a simple button style which animates the size of the button when it is pressed. Unfortunately, this causes the position of the button to animate (which I don't want). Is there a way to limit the animation only to the scale? I have tried surrounding the offending animation with .animation(nil) and .animation(.none) modifiers - but that didn't work. import SwiftUI struct ExampleBackgroundStyle: ButtonStyle {     func makeBody(configuration: Self.Configuration) -> some View {     configuration.label       .padding()       .foregroundColor(.white)       .background(Color.red)       .cornerRadius(40)       .padding(.horizontal, 20)       .animation(nil)       .scaleEffect(configuration.isPressed ? 0.6 : 1.0)       .animation(.easeIn(duration: 0.3))       .animation(nil)   } } when I show a button in a sheet, it animates in from the top left (0,0) to the position in the centre of the sheet after the sheet appears. If I remove the animation on the scaleEffect, then this doesn't happen.
3
0
3.8k
Jun ’20
Why have 2011 videos been delisted?
The 2020 session 'Author fragmented MPEG-4 content with AVAssetWriter' refers to the 2011 session 'Working with media in AVFoundation' (and when I say 'refers', I mean it says 'I'm not going to cover this important part - just look at the 2011 video) However - that session seems to be deleted/gone Frustrating... #wwdc20-10011
0
0
509
Oct ’21
What should we do when App Review just makes rules up?
I was just rejected for the following: Guideline 2.1 - Performance We found that while you have submitted in-app purchase products for your app, it does not show the expiration date in the app after a subscription purchase is made. However searching section 2.1 I can't even find the letters 'exp' (looking for expires, expiry) As far as I can tell - this requirement isn't anywhere in the rules. It looks like App Review (or a reviewer) are just making things up. How should we handle this?
0
0
481
Jun ’20
Can you take screenshots of Xcode Previews?
To be clear - I'm talking about the magically generated and updated PreviewProvider previews in Xcode Let's say I have a handful set up for different devices. iPhone X, iPad Pro, iPhone SEIs there an easy way for me to take screenshots to send for review?Obviously I can scroll down the screen and do it manually - but I'm thinking of something that would output a screenshot for each device.It seems like it _should_ be an obvious feature, but I can't find it!
2
0
5.2k
May ’20
IP addresses from NSNetService - why does my code fail in release builds?
I have some code that I have cobbled togther without really understanding the details.It works great in debug builds, but fails if you build for release.I found this out when I released an update 😟can anyone explain what is going on?extension NetService { var debugOnlyIPAddresses : [String] { var ipAddresses:[String] = [] guard let addresses = addresses else { return [] } for address in addresses { let data = address as NSData let inetAddress: sockaddr_in = data.castToCPointer() if inetAddress.sin_family == __uint8_t(AF_INET) { if let ip = String(cString: inet_ntoa(inetAddress.sin_addr), encoding: .ascii) { ipAddresses.append(ip) } } else if inetAddress.sin_family == __uint8_t(AF_INET6) { let inetAddress6: sockaddr_in6 = data.castToCPointer() let ipStringBuffer = UnsafeMutablePointer.allocate(capacity: Int(INET6_ADDRSTRLEN)) var addr = inetAddress6.sin6_addr if let ipString = inet_ntop(Int32(inetAddress6.sin6_family), &amp;addr, ipStringBuffer, __uint32_t(INET6_ADDRSTRLEN)) { if let ip = String(cString: ipString, encoding: .ascii) { // IPv6 ipAddresses.append(ip) } } ipStringBuffer.deallocate() } } print("returning \(ipAddresses)") return ipAddresses } }in release builds, inetAddress.sin_family is always 0
2
0
1.1k
Apr ’20