This paper introduces AIM, a collection of vision models pre-trained with an autoregressive objective. The models exhibit similar scaling properties to Large Language Models (LLMs) and achieve high performance on downstream tasks. Pre-training AIM does not require image-specific strategies.

https://arxiv.org/abs//2401.08541

YouTube: https://www.youtube.com/@ArxivPapers

TikTok: https://www.tiktok.com/@arxiv_papers

Apple Podcasts: https://podcasts.apple.com/us/podcast/arxiv-papers/id1692476016

Spotify: https://podcasters.spotify.com/pod/show/arxiv-papers