Apple, the tech giant known for its innovative products, has recently unveiled OpenELM, a groundbreaking set of small AI language models that can operate directly on smartphones. These models, described as proof-of-concept research models, could pave the way for future on-device AI offerings.
The OpenELM models, ranging from 270 million to 3 billion parameters, are now available on the Hugging Face platform under an Apple Sample Code License. What sets Apple’s models apart is their smaller size compared to other AI models, allowing for efficient operation on mobile devices.
Apple’s innovative approach with OpenELM includes a layer-wise scaling strategy that not only enhances performance but also enables training on fewer tokens. In addition, the company has made the code for CoreNet, a library used to train OpenELM, accessible to the open research community to foster collaboration and advancement in the field of AI.
While Apple has not yet integrated these AI models into its consumer devices, rumors suggest that they may be incorporated into the upcoming iOS 18 update. This update is anticipated to introduce new AI features that leverage on-device processing to prioritize user privacy. Apple is also rumored to be considering partnerships with Google or OpenAI for more complex processing capabilities.
Overall, Apple’s introduction of OpenELM marks a significant step towards bringing advanced AI capabilities directly to smartphones. As the tech industry continues to prioritize user privacy and on-device processing, Apple’s approach with OpenELM may serve as a blueprint for future developments in the field of artificial intelligence. Stay tuned for more updates as Apple continues to revolutionize the way we interact with AI technology.
“Zombie enthusiast. Subtly charming travel practitioner. Webaholic. Internet expert.”