The new betas have Apple Silicon optimizations for the Stable Diffusion AI image generator.

AI on Mac


AppleInsider may earn affiliate commissions for purchases made through links on our site.

The latest beta updates for macOS Ventura, iOS 16, and iPadOS will improve the performance of Stable Diffusion, according to an Apple engineer in the machine learning team. AI art creation routines.

On Thursday, Apple released the fourth developer beta of macOS Ventura 13.1, as well as beta versions for the company's other operating systems.

Tweeter Atila, whose biography lists him as a machine learning specialist at Apple, tweeted about the new macOS Ventura beta.

He says the latest beta releases include optimizations that allow stable diffusion to run at increased efficiency on the Apple Neural Engine and Apple Silicon GPU.

Apple has shared sample code for converting a neural network model from PyTorch to the company's Core ML system. They tested the base MacBook Air M2 with 8GB of RAM and completed 50 iterations of the model in 18 seconds.

For the “distilled” version of Stable Diffusion 2, the M2 can generate an image in less than one second.

Stable Diffusion is a text-to-speech machine learning model that can be used to create digital illustrations. Users can type in something like “cat piloting a submarine” and the program will create an image based on the text prompt.

Generation speed depends on the prompt and the hardware the code is running on.

Apple has a GitHub code repository for running Stable Diffusion on Apple Silicon with Core ML.

Leave a Reply

Your email address will not be published. Required fields are marked *