From Nasdaq: 2024-04-25 15:09:14

Apple has released OpenELM, open source large language models on Hugging Face Hub. These models can run on-device instead of the cloud, improving efficiency. The models come in four sizes – 270 million, 450 million, 1.1 billion, and 3 billion parameters, using a layer-wise scaling strategy for better accuracy.

Apple’s development of large language models will be entirely on-device, powered by iPhone processors. The OpenELM models, both pre-trained and instruction-tuned, aim to empower the open research community. Apple provides the complete framework for training and evaluation of the language model on public datasets, diverging from prior practices.

In an effort to enhance efficiency and accuracy, Apple’s OpenELM models use a layer-wise scaling strategy to allocate parameters within each layer of the transformer model. The tech giant is set to introduce iOS 18 with AI capabilities at the Worldwide Developers Conference, expanding its AI offerings to a wider audience.



Read more at Nasdaq: Apple Releases Open-Source LLMs To Run On-Device