Apple Uploads 20 New Core Machine Learning Models on HuggingFace

According to TapTechNews on June 20, Apple has further supplemented its existing public models and research papers, uploading 20 new core machine learning models on the open-source artificial intelligence knowledge base HuggingFace.

Apple publicly disclosed 4 open-source efficient LLMs (OpenELMs) in April this year and shared them on the HuggingFace platform. Now, Apple has added 20 CoreML models and a series of data sets on this platform to expand its AI lineup.

TapTechNews translated the content of HuggingFace's co-founder and chief executive Clement Delangue as follows:

This is a major update. Apple has uploaded a large number of models in its HuggingFace repo. These models use its CoreML framework and are mainly for text and images (image classification or deep segmentation, etc.).

Imagine an application that can effortlessly remove the unwanted background in photos or immediately identify the objects in front of you and provide their names in other languages and so on.

Apple has also updated Ferret, the large language model (LLM) released in October 2023, which is mainly used for image queries. Apple has also published research papers on generative artificial intelligence animation tools and creating artificial intelligence avatars.