Apple launches OpenELM, artificial intelligence that can run on your Mac or iPhone

Manzana entered the growing field of large language models (LLM) on devices with the release OpenELM, a family of open source models. designed to work efficiently on personal devices such as laptops or mobile phones.

OpenELM (Efficient Open Source Language Models) offer a significant advantage: unlike many other LLMs, these models do not require constant connection to cloud servers for processing. This can lead to faster response and better privacy for users.

Family consists of eight models ranging in size from 270 million to 3 billion parameters. While smaller models typically have less capacity than their larger counterparts, Apple’s approach uses a special technique called layered scaling to ensure good accuracy despite the reduced size.

OpenELM is designed for laptops or mobile devices. According to the official document, Apple engineers They managed to run them on a MacBook Pro with an M2 Max chip and 64GB of RAM.as well as on a computer with an Intel Core i9-13900KF processor with 64 GB of RAM and an NVIDIA RTX 4090 GPU. After running various performance tests, one of the options achieved the result. performance exceeds expectations.

The models were trained on a set containing data from Reddit, Github, Wikipedia and other sources.. OpenELM is available at HugFace under a sample code license that permits both commercial use and modification. However, it is important to note that Apple makes no guarantees regarding the safety or accuracy of the results. which these models can generate.

Apple gives a big boost to the artificial intelligence market in devices with OpenELM

This launch marks an unexpected step for Apple in the field of artificial intelligence. A few weeks ago, the Cupertino company introduced MGIE, an open-source model that allows image editing using natural language. Developed with researchers at the University of California, Santa Barbara, MGIE can interpret an image and edit it without much context.

Apple also launched a tool for creating machine learning models running on Apple Silicon chips. Like OpenELM and MGIE, the tech company’s engineers chose an open model and posted their work to a GitHub repository. Years of secrecy are behind us, at least when it comes to artificial intelligence, as Apple takes a position similar to Meta and Microsoft. opening up your progress for others to contribute.

OpenELM is the latest step in Apple’s strategy to position itself in the artificial intelligence market. While the new models don’t match the performance of GPT-4 or Gemini, they do run directly on the device. This means we won’t need to connect to the cloud, so our data will be more secure.

Apple is working on implementing artificial intelligence that matches your needs and values. A few days ago there was a rumor that the company’s engineers were developing a special artificial intelligence processor for servers. This will reduce dependence on third parties and give the company full control to ensure productivity and security.