Apple has been pretty mum about its plans for generative AI, but with the release of new AI models today, it seems like the company’s ambitions lie in making AI run locally on Apple devices.
The researchers at Apple recently released OpenELM, a series of four very small language models on the Hugging Face model library, on Wednesday. Apple has stated on its Hugging Face model page that OpenELM (Open-Source Efficient Language Model) can perform very efficiently on any text-related tasks like writing up emails, chatting, and much more.
These models from Apple are open-source and developers are free to use them to develop their own applications, integrating this tiny AI into their future apps.
The OpenELM model comes in four different sizes, i.e., 270 million parameters, 450 million parameters, 1.1 billion parameters, and 3 billion parameters. Usually, the more parameters a language model has, the more complex conversations and tasks it can understand.
Apple’s CEO, Tim Cook, has teased that new generative AI features might be coming to the company’s new lineup of devices very soon. Cook stated back in February that they have put in a ton of time and effort into these projects.
But it has not yet been clear how the company is going to use the AI models. The company has released other AI models before, though it has not released any AI foundational models for commercial use like its competitors have done.