Apple has unveiled a new open-source language model, OpenELM, that works locally on devices instead of clouds.
Image Credit: ReutersApple, often perceived as lagging behind in the AI race, has surprised many with its recent unveiling of OpenELM, a cutting-edge generative AI model optimised for on-device use.
OpenELM is a family of open-source language models specifically designed to run efficiently on Apple devices such as iPhones and Macs.
With tech giants like Microsoft and now Apple investing in small-model technology, the release of OpenELM underscores the growing importance of on-device AI.
Apple’s adoption of on-device AI not only enhances user privacy but also signifies a new era of innovation in the AI landscape.
Apple has unveiled a new open-source language model, OpenELM, that works locally on devices instead of clouds. This is likely to power the AI functions of the upcoming iPhone read more
Apple has unveiled a bunch of new language models, one of which, OpenELM, is likely to power its upcoming iPhones. Image Credit: Reuters
Apple, often perceived as lagging behind in the AI race, has surprised many with its recent unveiling of OpenELM, a cutting-edge generative AI model optimised for on-device use.
While Apple’s AI endeavours have been shrouded in secrecy, the introduction of OpenELM showcases the company’s long-standing commitment to AI innovation.
OpenELM is a family of open-source language models specifically designed to run efficiently on Apple devices such as iPhones and Macs. Using a layer-wise scaling strategy, OpenELM allocates parameters within each layer of the transformer model to enhance accuracy.
Advertisement
With eight models available in four different parameter sizes, ranging from 270M to 3B, OpenELM has been trained on public datasets, setting it apart from its predecessors.
What sets OpenELM apart is its ability to handle AI-powered tasks without relying on cloud servers, making it ideal for on-device use.
Despite requiring half the amount of training data compared to similar models like OLMo, OpenELM has demonstrated superior performance. It has been trained on CoreNet, an open-source library, alongside other models to enable efficient inference and fine-tuning on Apple devices.
In a departure from traditional practices, Apple has made a comprehensive release of OpenELM, providing not only model weights and inference code but also the complete framework for training and evaluation on publicly available datasets.
This move aims to empower the open research community and foster future endeavours in AI research.
The timing of OpenELM’s release, just weeks ahead of Apple’s WWDC event, hints at the potential AI features to be unveiled in iOS 18. This open-source initiative reflects a significant shift in Apple’s approach, moving away from its previously secretive nature to embrace transparency and collaboration within the research community.
With tech giants like Microsoft and now Apple investing in small-model technology, the release of OpenELM underscores the growing importance of on-device AI.
Apple’s adoption of on-device AI not only enhances user privacy but also signifies a new era of innovation in the AI landscape.
Advertisement
(With inputs from agencies)
)
