Loading Now

Summary of Accessing Vision Foundation Models Via Imagenet-1k, by Yitian Zhang et al.


Accessing Vision Foundation Models via ImageNet-1K

by Yitian Zhang, Xu Ma, Yue Bai, Huan Wang, Yun Fu

First submitted to arxiv on: 15 Jul 2024

Categories

  • Main: Computer Vision and Pattern Recognition (cs.CV)
  • Secondary: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)

     Abstract of paper      PDF of paper


GrooveSquid.com Paper Summaries

GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!

Summary difficulty Written by Summary
High Paper authors High Difficulty Summary
Read the original abstract here
Medium GrooveSquid.com (original content) Medium Difficulty Summary
The paper presents Proteus, a simple and general solution to distill foundation models into smaller equivalents on ImageNet-1K without access to the original training data. This is achieved by removing dataset bias from conventional knowledge distillation settings and introducing three levels of training objectives: token, patch, and feature. The resulting Proteus model is trained at ImageNet-level costs with surprising ability, making it accessible for broader research community use. When using DINOv2-g/14 as the teacher, Proteus-L/14 matches the performance of the Oracle method DINOv2-L/14 (142M training data) across 19 benchmarks and outperforms other vision foundation models including CLIP-L/14 (400M), OpenCLIP-L/14 (400M/2B), and SynCLR-L/14 (600M) with a significantly smaller training set of 1.2M images.
Low GrooveSquid.com (original content) Low Difficulty Summary
Proteus is a new way to make big AI models smaller, so they can be used by more people without needing huge amounts of data or powerful computers. This is important because it makes it easier for researchers to use these powerful models in their own work. The team developed Proteus by removing things that can make the training process biased and adding new ways to train the model. They tested Proteus with a small dataset and found that it works really well, even better than some of the biggest AI models out there.

Keywords

* Artificial intelligence  * Knowledge distillation  * Token