Hi We're on tensorflow 2.20 that has support now for python 3.13 (finally!). tensorflow-metal is still only supporting 2.18 which is over a year old. When can we expect to see support in tensorflow-metal for tf 2.20 (or later!) ? I bought a mac thinking I would be able to get great performance from the M processors but here I am using my CPU for my ML projects. If it's taking so long to release it, why not open source it so the community can keep it more up to date? cheers Matt

I'm downloading a fine-tuned model from HuggingFace which is then cached on my Mac when the app first starts. However, I wanted to test adding a progress bar to show the download progress. To test this I need to delete the cached model. From what I've seen online this is cached at /Users/userName/.cache/huggingface/hub However, if I delete the files from here, using Terminal, the app still seems to be able to access the model. Is the model cached somewhere else? On my iPhone it seems deleting the app also deletes the cached model (app data) so that is useful.

I have used mlx_lm.lora to fine tune a mistral-7b-v0.3-4bit model with my data. I fused the mistral model with my adapters and upload the fused model to my directory on huggingface. I was able to use mlx_lm.generate to use the fused model in Terminal. However, I don't know how to load the model in Swift. I've used Imports import SwiftUI import MLX import MLXLMCommon import MLXLLM let modelFactory = LLMModelFactory.shared let configuration = ModelConfiguration( id: "pharmpk/pk-mistral-7b-v0.3-4bit" ) // Load the model off the main actor, then assign on the main actor let loaded = try await modelFactory.loadContainer(configuration: configuration) { progress in print("Downloading progress: \(progress.fractionCompleted * 100)%") } await MainActor.run { self.model = loaded } I'm getting an error runModel error: downloadError("A server with the specified hostname could not be found.") Any suggestions? Thanks, David PS, I can load the model from the app bundle // directory: Bundle.main.resourceURL! but it's too big to upload for Testflight

Using Tensorflow for Silicon gives inaccurate results when compared to Google Colab GPU (9-15% differences). Here are my install versions for 4 anaconda env's. I understand the Floating point precision can be an issue, batch size, activation functions but how do you rectify this issue for the past 3 years? 1.) Version TF: 2.12.0, Python 3.10.13, tensorflow-deps: 2.9.0, tensorflow-metal: 1.2.0, h5py: 3.6.0, keras: 2.12.0 2.) Version TF: 2.19.0, Python 3.11.0, tensorflow-metal: 1.2.0, h5py: 3.13.0, keras: 3.9.2, jax: 0.6.0, jax-metal: 0.1.1,jaxlib: 0.6.0, ml_dtypes: 0.5.1 3.) python: 3.10.13,tensorflow: 2.19.0,tensorflow-metal: 1.2.0, h5py: 3.13.0, keras: 3.9.2, ml_dtypes: 0.5.1 4.) Version TF: 2.16.2, tensorflow-deps:2.9.0,Python: 3.10.16, tensorflow-macos 2.16.2, tensorflow-metal: 1.2.0, h5py:3.13.0, keras: 3.9.2, ml_dtypes: 0.3.2 Install of Each ENV with common example: Create ENV: conda create --name TF_Env_V2 --no-default-packages start env: source TF_Env_Name ENV_1.) conda install -c apple tensorflow-deps , conda install tensorflow,pip install tensorflow-metal,conda install ipykernel ENV_2.) conda install pip python==3.11, pip install tensorflow,pip install tensorflow-metal,conda install ipykernel ENV_3) conda install pip python 3.10.13,pip install tensorflow, pip install tensorflow-metal,conda install ipykernel ENV_4) conda install -c apple tensorflow-deps, pip install tensorflow-macos, pip install tensor-metal, conda install ipykernel Example used on all 4 env: import tensorflow as tf cifar = tf.keras.datasets.cifar100 (x_train, y_train), (x_test, y_test) = cifar.load_data() model = tf.keras.applications.ResNet50( include_top=True, weights=None, input_shape=(32, 32, 3), classes=100,) loss_fn = tf.keras.losses.SparseCategoricalCrossentropy(from_logits=False) model.compile(optimizer="adam", loss=loss_fn, metrics=["accuracy"]) model.fit(x_train, y_train, epochs=5, batch_size=64)

The WWDC25: Explore large language models on Apple silicon with MLX video talks about using your own data to fine-tune a large language model. But the video doesn't explain what kind of data can be used. The video just shows the command to use and how to point to the data folder. Can I use PDFs, Word documents, Markdown files to train the model? Are there any code examples on GitHub that demonstrate how to do this?

I watched this year WWDC25 "Read Documents using the Vision framework". At the end of video there is mention of new DetectHandPoseRequest model for hand pose detection in Vision API. I looked Apple documentation and I don't see new revision. Moreover probably typo in video because there is only DetectHumanPoseRequst (swift based) and VNDetectHumanHandPoseRequest (obj-c based) (notice lack of Human prefix in WWDC video) First one have revision only added in iOS 18+: https://developer.apple.com/documentation/vision/detecthumanhandposerequest/revision-swift.enum/revision1 Second one have revision only added in iOS14+: https://developer.apple.com/documentation/vision/vndetecthumanhandposerequestrevision1 I don't see any new revision targeting iOS26+

Bear with me, please. Please make sure a highly skilled technical person reads and understands this. I want to describe my vision for (AI/Algorithmically) Optimised Operating Systems. To explain it properly, I will describe the process to build it (pseudo). Required Knowledge (no particular order): Processor Logic Circuits, LLM models, LLM tool usage, Python OO coding, Procedural vs OO, NLP fuzzy matching, benchmarking, canvas/artefacts/dynamic HTML interfaces, concepts of how AI models are vastly compressed and miniaturised forms of full data, Algorithmic vs AI. First, take all OO Python code (example) on GitHub (example), then separate each function from each object into its own procedure (procedural logic) by making a logical procedural list of actions to perform only that function based on its entire dependency chain (i.e. all other objects it relies on). Relate all compiled functions using (for example) fuzzy matching on the name, or AI-based functional profiling to get multiple instances of each function. Starting with the most used function, test each one against the others that perform the same task for bugs and completeness. Determine the fastest, most optimal version of that function (and every function). Add a single instance of each most optimal function to the centralised tool codebase, which will later be utilised by the language models. This ensures we rely only on the most optimised function for each and every use case — with every program using one shared instance of that function instead of compiling it separately. In this way, we optimise all existing (Python OO) code based on maximum availability of certain resources. Use the optimised code to teach the AI how to map each user task directly to optimised code (i.e. each user-facing function in unoptimised program code to its optimised version). Later, teach the codebase how to optimise these functions directly to the logic circuits on the processor — as an Operating System — using essentially the same process. Eventually, use the most optimised form for tasks via a dynamic canvas/artefacts UI interface generated by the Start Menu using the language model. Allow custom interfaces to emerge dynamically from the available functions based on language understanding. Grow and optimise the algorithmic codebase until the entire operating system — with versioning — evolves based on the AI-generated user interfaces (compressing them into new versions of the tools codebase). Please don’t give up until you understand this concept completely and nothing in the description is beyond your full comprehension. It’s a highly significant step forward.

Hi, recently i tried to fine-tune Gemma-2-2b mlx model on my macbook (24 GB UMA). The code started running, after few seconds i saw swap size reaching 50GB and ram around 23 GB and then it stopped. I ran the Gemma-2-2b (cuda) on colab, it ran and occupied 27 GB on A100 gpu and worked fine. Here i didn't experienced swap issue. Now my question is if my UMA was more than 27 GB, i also would not have experienced swap disk issue. Thanks.

I'm playing with the new Vision API for iOS18, specifically with the new CalculateImageAestheticsScoresRequest API. When I try to perform the image observation request I get this error: internalError("Error Domain=NSOSStatusErrorDomain Code=-1 \"Failed to create espresso context.\" UserInfo={NSLocalizedDescription=Failed to create espresso context.}") The code is pretty straightforward: if let image = image { let request = CalculateImageAestheticsScoresRequest() Task { do { let cgImg = image.cgImage! let observations = try await request.perform(on: cgImg) let description = observations.description let score = observations.overallScore print(description) print(score) } catch { print(error) } } } I'm running it on a M2 using the simulator. Is it a bug? What's wrong?

Following WWDC24 video "Discover Swift enhancements in the Vision framework" recommendations (cfr video at 10'41"), I used the following code to perform multiple new iOS 18 `RecognizedTextRequest' in parallel. Problem: if more than 2 request are run in parallel, the request will hang, leaving the app in a state where no more requests can be started. -> deadlock I tried other ways to run the requests, but no matter the method employed, or what device I use: no more than 2 requests can ever be run in parallel. func triggerDeadlock() {} try await withThrowingTaskGroup(of: Void.self) { group in // See: WWDC 2024 Discover Siwft enhancements in the Vision framework at 10:41 // ############## THIS IS KEY let maxOCRTasks = 5 // On a real-device, if more than 2 RecognizeTextRequest are launched in parallel using tasks, the request hangs // ############## THIS IS KEY for idx in 0..<maxOCRTasks { let url = ... // URL to some image group.addTask { // Perform OCR let _ = await performOCRRequest(on: url: url) } } var nextIndex = maxOCRTasks for try await _ in group { // Wait for the result of the next child task that finished if nextIndex < pageCount { group.addTask { let url = ... // URL to some image // Perform OCR let _ = await performOCRRequest(on: url: url) } nextIndex += 1 } } } } // MARK: - ASYNC/AWAIT version with iOS 18 @available(iOS 18, *) func performOCRRequest(on url: URL) async throws -> [RecognizedText] { // Create request var request = RecognizeTextRequest() // Single request: no need for ImageRequestHandler // Configure request request.recognitionLevel = .accurate request.automaticallyDetectsLanguage = true request.usesLanguageCorrection = true request.minimumTextHeightFraction = 0.016 // Perform request let textObservations: [RecognizedTextObservation] = try await request.perform(on: url) // Convert [RecognizedTextObservation] to [RecognizedText] return textObservations.compactMap { observation in observation.topCandidates(1).first } } I also found this Swift forums post mentioning something very similar. I also opened a feedback: FB17240843

During testing the “Bringing advanced speech-to-text capabilities to your app” sample app demonstrating the use of iOS 26 SpeechAnalyzer, I noticed that the language model for the English locale was presumably already downloaded. Upon checking the documentation of AssetInventory, I found out that indeed, the language model can be preinstalled on the system. Can someone from the dev team share more info about what assets are preinstalled by the system? For example, can we safely assume that the English language model will almost certainly be already preinstalled by the OS if the phone has the English locale?

Hey Devs, I'm trying to create my own Real Time Text detection like this Apple project. https://developer.apple.com/documentation/vision/extracting-phone-numbers-from-text-in-images I want to use the new iOS18 RecognizeTextRequest instead of the old VNRecognizeTextRequest in my SwiftUI project. This is my delegate code with the camera setup. I removed region of interest for debugging but I'm trying to scan English words in books. The idea is to get one word in the ROI in the future. But I can't even get proper words so testing without ROI incase my math is wrong. @Observable class CameraManager: NSObject, AVCapturePhotoCaptureDelegate ... override init() { super.init() setUpVisionRequest() } private func setUpVisionRequest() { textRequest = RecognizeTextRequest(.revision3) } ... func setup() -> Bool { captureSession.beginConfiguration() guard let captureDevice = AVCaptureDevice.default( .builtInWideAngleCamera, for: .video, position: .back) else { return false } self.captureDevice = captureDevice guard let deviceInput = try? AVCaptureDeviceInput(device: captureDevice) else { return false } /// Check whether the session can add input. guard captureSession.canAddInput(deviceInput) else { print("Unable to add device input to the capture session.") return false } /// Add the input and output to session captureSession.addInput(deviceInput) /// Configure the video data output videoDataOutput.setSampleBufferDelegate( self, queue: videoDataOutputQueue) if captureSession.canAddOutput(videoDataOutput) { captureSession.addOutput(videoDataOutput) videoDataOutput.connection(with: .video)? .preferredVideoStabilizationMode = .off } else { return false } // Set zoom and autofocus to help focus on very small text do { try captureDevice.lockForConfiguration() captureDevice.videoZoomFactor = 2 captureDevice.autoFocusRangeRestriction = .near captureDevice.unlockForConfiguration() } catch { print("Could not set zoom level due to error: \(error)") return false } captureSession.commitConfiguration() // potential issue with background vs dispatchqueue ?? Task(priority: .background) { captureSession.startRunning() } return true } } // Issue here ??? extension CameraManager: AVCaptureVideoDataOutputSampleBufferDelegate { func captureOutput( _ output: AVCaptureOutput, didOutput sampleBuffer: CMSampleBuffer, from connection: AVCaptureConnection ) { guard let pixelBuffer = CMSampleBufferGetImageBuffer(sampleBuffer) else { return } Task { textRequest.recognitionLevel = .fast textRequest.recognitionLanguages = [Locale.Language(identifier: "en-US")] do { let observations = try await textRequest.perform(on: pixelBuffer) for observation in observations { let recognizedText = observation.topCandidates(1).first print("recognized text \(recognizedText)") } } catch { print("Recognition error: \(error.localizedDescription)") } } } } The results I get look like this ( full page of English from a any book) recognized text Optional(RecognizedText(string: e bnUI W4, confidence: 0.5)) recognized text Optional(RecognizedText(string: ?'U, confidence: 0.3)) recognized text Optional(RecognizedText(string: traQt4, confidence: 0.3)) recognized text Optional(RecognizedText(string: li, confidence: 0.3)) recognized text Optional(RecognizedText(string: 15,1,#, confidence: 0.3)) recognized text Optional(RecognizedText(string: jllÈ, confidence: 0.3)) recognized text Optional(RecognizedText(string: vtrll, confidence: 0.3)) recognized text Optional(RecognizedText(string: 5,1,: 11, confidence: 0.5)) recognized text Optional(RecognizedText(string: 1141, confidence: 0.3)) recognized text Optional(RecognizedText(string: jllll ljiiilij41, confidence: 0.3)) recognized text Optional(RecognizedText(string: 2f4, confidence: 0.3)) recognized text Optional(RecognizedText(string: ktril, confidence: 0.3)) recognized text Optional(RecognizedText(string: ¥LLI, confidence: 0.3)) recognized text Optional(RecognizedText(string: 11[Itl,, confidence: 0.3)) recognized text Optional(RecognizedText(string: 'rtlÈ131, confidence: 0.3)) Even with ROI set to a specific rectangle Normalized to Vision, I get the same results with single characters returning gibberish. Any help would be amazing thank you. Am I using the buffer right ? Am I using the new perform(on: CVPixelBuffer) right ? Maybe I didn't set up my camera properly? I can provide code

I have a question. In China, long pressing a picture in the album can segment the target. Is this model a local model? Is there any information? Can developers use it?

In this WWDC25 session, it is explictely mentioned that apps should support AttributedString for text parameters to their App Intents. However, I have not gotten this to work. Whenever I pass rich text (either generated by the new "Use Model" intent or generated manually for example using "Make Rich Text from Markdown"), my Intent gets an AttributedString with the correct characters, but with all attributes stripped (so in effect just plain text). struct TestIntent: AppIntent { static var title = LocalizedStringResource(stringLiteral: "Test Intent") static var description = IntentDescription("Tests Attributed Strings in Intent Parameters.") @Parameter var text: AttributedString func perform() async throws -> some IntentResult & ReturnsValue<AttributedString> { return .result(value: text) } } Is there anything else I am missing?

In WWDC25 Metal 4 released quite excited new features for machine learning optimization, but as we all know the pytorch based on metal shader performance (mps) is the one of most important tools for Mac machine learning area.but on mps introduced website we cannot see any support information for metal4.

WWDC25: Combine Metal 4 machine learning and graphics Demonstrated a way to combine neural network in the graphics pipeline directly through the shaders, using an example of Texture Compression. However there is no mention of using which ML technique texture is compressed. Can anyone point me to some well known model/s for this particular use case shown in WWDC25.

Hello, I've been dealing with a puzzling issue for some time now, and I’m hoping someone here might have insights or suggestions. The Problem: We’re observing an occasional crash in our app that seems to originate from the Vision framework. Frequency: It happens randomly, after many successful executions of the same code, hard to tell how long the app was working, but in some cases app could run for like a month without any issues. Devices: The issue doesn't seem device-dependent (we’ve seen it on various iPad models). OS Versions: The crashes started occurring with iOS 18.0.1 and are still present in 18.1 and 18.1.1. What I suspected: The crash logs point to a potential data race within the Vision framework. The relevant section of the code where the crash happens: guard let cgImage = image.cgImage else { throw ... } let request = VNCoreMLRequest(model: visionModel) try VNImageRequestHandler(cgImage: cgImage).perform([request]) // <- the line causing the crash Since the code is rather simple, I'm not sure what else there could be missing here. The images sent here are uniform (fixed size). Model is loaded and working, the crash occurs random after a period of time and the call worked correctly many times. Also, the model variable is not an optional. Here is the crash log: libobjc.A objc_exception_throw CoreFoundation -[NSMutableArray removeObjectsAtIndexes:] Vision -[VNWeakTypeWrapperCollection _enumerateObjectsDroppingWeakZeroedObjects:usingBlock:] Vision -[VNWeakTypeWrapperCollection addObject:droppingWeakZeroedObjects:] Vision -[VNSession initWithCachingBehavior:] Vision -[VNCoreMLTransformer initWithOptions:model:error:] Vision -[VNCoreMLRequest internalPerformRevision:inContext:error:] Vision -[VNRequest performInContext:error:] Vision -[VNRequestPerformer _performOrderedRequests:inContext:error:] Vision -[VNRequestPerformer _performRequests:onBehalfOfRequest:inContext:error:] Vision -[VNImageRequestHandler performRequests:gatheredForensics:error:] OurApp ModelWrapper.perform And I'm a bit lost at this point, I've tried everything I could image so far. I've tried to putting a symbolic breakpoint in the removeObjectsAtIndexes to check if some library (e.g. crash reporter) we use didn't do some implementation swap. There was none, and if anything did some method swizzling, I'd expect that to show in the stack trace before the original code would be called. I did peek into the previous functions and I've noticed a lock used in one of the Vision methods, so in my understanding any data race in this code shouldn't be possible at all. I've also put breakpoints in the NSLock variants, to check for swizzling/override with a category and possibly messing the locking - again, nothing was there. There is also another model that is running on a separate queue, but after seeing the line with the locking in the debugger, it doesn't seem to me like this could cause a problem, at least not in this specific spot. Is there something I'm missing here, or something I'm doing wrong? Thanks in advance for your help!

Does CoreML object detection only support AABB (Axis-Aligned Bounding Boxes) or also OBB (Oriented Bounded Boxes)? If not, any way to do it using Apple frameworks?

I am using gemini2.5-flash with SwiftUI. How can I receive a response in JSON?

At WWDC25 we launched a new type of Lab event for the developer community - Group Labs. A Group Lab is a panel Q&A designed for a large audience of developers. Group Labs are a unique opportunity for the community to submit questions directly to a panel of Apple engineers and designers. Here are the highlights from the WWDC25 Group Lab for Machine Learning and AI Frameworks. What are you most excited about in the Foundation Models framework? The Foundation Models framework provides access to an on-device Large Language Model (LLM), enabling entirely on-device processing for intelligent features. This allows you to build features such as personalized search suggestions and dynamic NPC generation in games. The combination of guided generation and streaming capabilities is particularly exciting for creating delightful animations and features with reliable output. The seamless integration with SwiftUI and the new design material Liquid Glass is also a major advantage. When should I still bring my own LLM via CoreML? It's generally recommended to first explore Apple's built-in system models and APIs, including the Foundation Models framework, as they are highly optimized for Apple devices and cover a wide range of use cases. However, Core ML is still valuable if you need more control or choice over the specific model being deployed, such as customizing existing system models or augmenting prompts. Core ML provides the tools to get these models on-device, but you are responsible for model distribution and updates. Should I migrate PyTorch code to MLX? MLX is an open-source, general-purpose machine learning framework designed for Apple Silicon from the ground up. It offers a familiar API, similar to PyTorch, and supports C, C++, Python, and Swift. MLX emphasizes unified memory, a key feature of Apple Silicon hardware, which can improve performance. It's recommended to try MLX and see if its programming model and features better suit your application's needs. MLX shines when working with state-of-the-art, larger models. Can I test Foundation Models in Xcode simulator or device? Yes, you can use the Xcode simulator to test Foundation Models use cases. However, your Mac must be running macOS Tahoe. You can test on a physical iPhone running iOS 18 by connecting it to your Mac and running Playgrounds or live previews directly on the device. Which on-device models will be supported? any open source models? The Foundation Models framework currently supports Apple's first-party models only. This allows for platform-wide optimizations, improving battery life and reducing latency. While Core ML can be used to integrate open-source models, it's generally recommended to first explore the built-in system models and APIs provided by Apple, including those in the Vision, Natural Language, and Speech frameworks, as they are highly optimized for Apple devices. For frontier models, MLX can run very large models. How often will the Foundational Model be updated? How do we test for stability when the model is updated? The Foundation Model will be updated in sync with operating system updates. You can test your app against new model versions during the beta period by downloading the beta OS and running your app. It is highly recommended to create an "eval set" of golden prompts and responses to evaluate the performance of your features as the model changes or as you tweak your prompts. Report any unsatisfactory or satisfactory cases using Feedback Assistant. Which on-device model/API can I use to extract text data from images such as: nutrition labels, ingredient lists, cashier receipts, etc? Thank you. The Vision framework offers the RecognizeDocumentRequest which is specifically designed for these use cases. It not only recognizes text in images but also provides the structure of the document, such as rows in a receipt or the layout of a nutrition label. It can also identify data like phone numbers, addresses, and prices. What is the context window for the model? What are max tokens in and max tokens out? The context window for the Foundation Model is 4,096 tokens. The split between input and output tokens is flexible. For example, if you input 4,000 tokens, you'll have 96 tokens remaining for the output. The API takes in text, converting it to tokens under the hood. When estimating token count, a good rule of thumb is 3-4 characters per token for languages like English, and 1 character per token for languages like Japanese or Chinese. Handle potential errors gracefully by asking for shorter prompts or starting a new session if the token limit is exceeded. Is there a rate limit for Foundation Models API that is limited by power or temperature condition on the iPhone? Yes, there are rate limits, particularly when your app is in the background. A budget is allocated for background app usage, but exceeding it will result in rate-limiting errors. In the foreground, there is no rate limit unless the device is under heavy load (e.g., camera open, game mode). The system dynamically balances performance, battery life, and thermal conditions, which can affect the token throughput. Use appropriate quality of service settings for your tasks (e.g., background priority for background work) to help the system manage resources effectively. Do the foundation models support languages other than English? Yes, the on-device Foundation Model is multilingual and supports all languages supported by Apple Intelligence. To get the model to output in a specific language, prompt it with instructions indicating the user's preferred language using the locale API (e.g., "The user's preferred language is en-US"). Putting the instructions in English, but then putting the user prompt in the desired output language is a recommended practice. Are larger server-based models available through Foundation Models? No, the Foundation Models API currently only provides access to the on-device Large Language Model at the core of Apple Intelligence. It does not support server-side models. On-device models are preferred for privacy and for performance reasons. Is it possible to run Retrieval-Augmented Generation (RAG) using the Foundation Models framework? Yes, it is possible to run RAG on-device, but the Foundation Models framework does not include a built-in embedding model. You'll need to use a separate database to store vectors and implement nearest neighbor or cosine distance searches. The Natural Language framework offers simple word and sentence embeddings that can be used. Consider using a combination of Foundation Models and Core ML, using Core ML for your embedding model.