Summary of Weight Copy and Low-rank Adaptation For Few-shot Distillation Of Vision Transformers, by Diana-nicoleta Grigore et al.
Weight Copy and Low-Rank Adaptation for Few-Shot Distillation of Vision Transformers
by Diana-Nicoleta Grigore, Mariana-Iuliana Georgescu, Jon Alvarez Justo, Tor Johansen, Andreea Iuliana Ionescu, Radu Tudor Ionescu
First submitted to arxiv on: 14 Apr 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
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 authors propose a novel few-shot feature distillation approach for vision transformers, leveraging the consistent depth-wise structure of these models. The method involves copying weights from pre-trained teacher transformers to shallower student transformers and then using LoRA to distill knowledge in a few-shot scenario. The approach is tested on six datasets from various domains and tasks, achieving superior results compared to state-of-the-art competitors. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper introduces a new way for computers to learn by copying the “knowledge” of large pre-trained models using only a small amount of data and computer power. It’s like teaching a student how to do something by showing them how to do it, rather than explaining it in detail. The authors test their approach on different types of images and tasks and show that it works better than other methods. |
Keywords
» Artificial intelligence » Distillation » Few shot » Lora
