Koan-Sin Tan

MobilenetV4 was made public last week, see https://arxiv.org/abs/2404.10518 or https://arxiv.org/html/2404.10518v1 According to numbers in the paper, it should be able to get > 300 qps for iPhone 13.

@freedomtan to try it on iPhone 13 again.

> @freedomtan to try it on iPhone 13 again. As I got before, on iPhone 13, it's about 220 qps

Let's try to have PyTorch model (with weights from the TensorFlow model).

@RSMNYS With Xcode 16.0 beta and iOS 18 + MLPackage targeting iOS 15 or later, it's possible to get per-op time. Please check https://developer.apple.com/videos/play/wwdc2024/10161/?time=927

Per-op profiling actually is possible on iOS 17.4+ / MacOS 14.4+. I wrote a little command line program and tested it on my Macbook Pro M1, see https://github.com/freedomtan/coreml_modelc_profling

> FWIW There's still no official weights from the paper authors, but I've trained a number of PyTorch native MobileNetV4 models and made them available in `timm`. The conv-medium runs...

@RSMNYS `pip install git+https://github.com/huggingface/pytorch-image-models.git` then ```python import timm import torch import coremltools as ct torch_model = timm.create_model("hf-hub:timm/mobilenetv4_conv_large.e600_r384_in1k", pretrained=True) torch_model.eval() # Trace the model with random data. example_input = torch.rand(1, 3,...

@RSMNYS and @anhappdev According to coremltools [8.0b1 doc on quantization](https://apple.github.io/coremltools/docs-guides/source/opt-workflow.html#with-calibration-dataset), it's possible to create a calibrated quantized A8W8 PTQ model from an existing Core ML model. I used random data...

@RSMNYS Please update screenshots with new UI for both Google Play and Apple App Store,