Cam4DL vs TFLite LiteRT TensorFlow Test

Cam4DL is an app for creating image and video datasets that are typically needed in deep learning projects. RECORD VIDEO CLIPS READY FOR ACTION CLASSIFICATION With Cam4DL you can now record small video clips that you can use for training an action classifier in CreateML. You can choose to record these video clips with the same frame rate and timing options available during CreateML training phase. You can record multiple video clips automatically and hands free. First select a category then tap on the shutter button to start the recording. When a video clip recording ends you have 3 seconds to get ready for the next clip recording. This automatically starts until you stop the recording by tapping again on the shutter button. The video clips are recorded without sound because this is hardly used for action classification. TAKE PHOTOS FOR OBJECT AND FACE ANALYSIS Using Cam4DL you can also take photos organized in categories and optionally generate depth maps and face metadata with ease at the moment of capture. Just create a project of type Face or Object, and choose if you want to include a depth map. Then choose if you want to optionally crop and resize the image at the moment of capture. You can then start taking photos that are automatically cropped, resized and organized in folders along with their metadata. The best part of Cam4DL is that all of the generated data is open and accessible from the Apple Files app, so that transferring this data to your workstation is easier and faster than ever. FACE DATA AND DEPTHMAP AUTOMATICALLY COMPUTED When you take a photo inside a project of type Face, the app automatically computes face features and save them in json files, so that you can later use them for your data science projects. When you select the depth map option during project creation, the app automatically computes the depth map while you take the photo and stores it in a png file. Cam4DL keeps the dataset pictures separate from Apple's Photos app. OPEN ACCESS TO ALL THE GENERATED DATA All the photos you take with Cam4DL and the relative metadata such as the face data are categorized and saved inside the app's document directory so that you can easily share them with your workstation for training. To access the data generated by Cam4DL, just open Apple's Files app and then navigate to "iPhone -> Cam4DL" folder. Made with Love and SwiftUI WARNING: All the datasets are stored inside the app's sandbox. For this reason if you delete the app you also delete the datasets, including all the videos, photos and the associated metadata. If you for any reason need to delete the app REMEMBER TO BACKUP your datasets to another location using the Files app.

How fast does your .tflite model actually run on this iPhone? Pick CPU, GPU or Neural Engine. Get real latency and QPS. No Mac, no Xcode project, no Bazel build, no cloud, no account. Open a .tflite file from the Files app, pick an accelerator, run it. A real measurement on the exact device you care about, in about a minute — no Bazel workspace, no benchmark tool built from source. PICK THE ACCELERATOR - Neural Engine — attaches the Core ML delegate, so supported operators run on the Apple Neural Engine - GPU — attaches the GPU delegate, backed by Metal - CPU — the plain interpreter, with a thread count you choose: 1, 2, 4, 6 or 8 One configuration per run: run it, change the delegate or thread count, load it again, compare the numbers yourself. WHAT YOU GET - Latency: mean, min and max, in milliseconds - Queries per second - Total queries completed and total run duration - A query count you set per run, from 50 to 5,000 invocations Mean, min and max time the model inference call only — no pre-processing, no post-processing. Duration covers the whole loop. The screen stays awake during a run. WHAT YOU SEE ABOUT THE MODEL - Every input and output tensor: index, name, shape and data type (uInt8, int32, float16, float32 and the rest) - Model file size and framework - Live app memory usage against the memory available to the app, refreshed while the model runs - A device tab with model identifier, system version, disk space, and the exact TensorFlow Lite runtime version this build links against A MobileNet model ships in the app and loads on launch, so you see a real measurement before importing your own. READ THIS BEFORE YOU TRUST A NUMBER Inference runs on a zeroed dummy input. The app fills every input tensor with zeros and invokes the model, so what you measure is the compute cost of the graph on the accelerator you picked — latency and throughput, nothing else. It is not an accuracy test. It will not tell you whether your model gives the correct answer, and it does not show output tensor values. Ask it "how fast", not "how correct". What it does not do, stated up front: no side-by-side accelerator comparison in one run, no per-layer or per-operator profiling, no result export, no image or real-data input. It opens .tflite files only. PRIVACY Your model never leaves the device. Loading, inspection, inference and timing all happen locally on your iPhone or iPad. No account, no sign-in — your models and your results are never uploaded. TensorFlow Lite is now called LiteRT; the format and the .tflite extension are unchanged. TensorFlow, the TensorFlow logo and any related marks are trademarks of Google Inc. Not affiliated with or endorsed by Google.