苹果公司近期在人工智能领域取得新进展，推出了一款名为自回归视觉模型（AIM）的新技术。该技术由苹果的研究人员在最新论文《Scalable Pre-training of Large Autoregressive Image Models》中提出，旨在探讨如何利用自回归目标训练视觉转换器（ViT）模型，以获得与大语言模型（LLMs）相似的表征学习扩展能力。研究结果显示，AIM模型能够轻松扩展至数十亿个参数，并且能够有效利用大量未整理的图像数据。这一技术的出现，标志着苹果在图像处理和人工智能领域迈出了重要的一步，未来可能为图像理解和生成技术带来新的突破。
Title: Apple Unveils Autoregressive Image Model AIM
Keywords: Apple AI, Autoregressive Model, Image Processing
News content:
Apple has made new progress in the field of artificial intelligence, introducing a new technology called the Autoregressive Image Model (AIM). The technology, proposed by Apple’s researchers in their latest paper titled “Scalable Pre-training of Large Autoregressive Image Models,” aims to explore how to train visual transformers (ViT) models using an autoregressive objective to achieve similar scaling capabilities in representation learning as large language models (LLMs). The research results show that the AIM model can easily scale to billions of parameters and can effectively utilize large amounts of uncurated image data. The advent of this technology marks a significant step forward for Apple in the field of image processing and artificial intelligence, and it may bring new breakthroughs to image understanding and generation technologies in the future.
【来源】https://www.jiqizhixin.com/articles/2024-01-18-7
Views: 1