paper-with-me

홈 › Papers

DMT: Comprehensive Distillation with Multiple Self-supervised Teachers

2023-12-19 · Yuang Liu, Jing Wang, Qiang Zhou, Fan Wang, Jun Wang, Wei zhang

Numerous self-supervised learning paradigms, such as contrastive learning and masked image modeling, have been proposed to acquire powerful and general representations from unlabeled data. However, these models are commonly pretrained within their specific framework alone, failing to consider the complementary nature of visual representations. To tackle this issue, we introduce Comprehensive Distillation with Multiple Self-supervised Teachers (DMT) for pretrained model compression, which leverages the strengths of multiple off-the-shelf self-supervised models. Our experimental results on prominent benchmark datasets exhibit that the proposed method significantly surpasses state-of-the-art competitors while retaining favorable efficiency metrics. On classification tasks, our DMT framework utilizing three different self-supervised ViT-Base teachers enhances the performance of both small/tiny models and the base model itself. For dense tasks, DMT elevates the AP/mIoU of standard SSL models on MS-COCO and ADE20K datasets by 4.0%.

📄 PDF Abstract BibTeX arXiv:2312.11938

Code (0)

등록된 구현이 없습니다.

Tasks

Contrastive LearningModel CompressionSelf-Supervised Learning

Methods 이 논문이 사용한 방법론

Contrastive Learning 설명 없음
BASE 설명 없음

Similar Papers 제목 키워드 기반

Automated Graph Self-supervised Learning via Multi-teacher Knowledge Distillation

2022-10-05 · Lirong Wu, Yufei Huang, Haitao Lin, Zicheng Liu 외

Self-supervised learning on graphs has recently achieved remarkable success in graph representation learning. With hundreds of self-supervised pretext tasks proposed over the past few years, the research community has gr…

Graph Representation LearningKnowledge DistillationRepresentation LearningSelf-Supervised Learning

Autoencoders as Cross-Modal Teachers: Can Pretrained 2D Image Transformers Help 3D Representation Learning?

2022-12-16 · Runpei Dong, Zekun Qi, Linfeng Zhang, Junbo Zhang 외

The success of deep learning heavily relies on large-scale data with comprehensive labels, which is more expensive and time-consuming to fetch in 3D compared to 2D images or natural languages. This promotes the potential…

3D Point Cloud ClassificationFew-Shot 3D Point Cloud ClassificationKnowledge DistillationRepresentation Learning

Multi-Level Feature Distillation of Joint Teachers Trained on Distinct Image Datasets

2024-10-29 · Adrian Iordache, Bogdan Alexe, Radu Tudor Ionescu

We propose a novel teacher-student framework to distill knowledge from multiple teachers trained on distinct datasets. Each teacher is first trained from scratch on its own dataset. Then, the teachers are combined into a…

Action RecognitionAllimage-classificationImage Classification

Continual Distillation of Teachers from Different Domains

2026-04-10 · Nicolas Michel, Maorong Wang, Jiangpeng He, Toshihiko Yamasaki arxiv

Deep learning models continue to scale, with some requiring more storage than many large-scale datasets. Thus, we introduce a new paradigm: Continual Distillation (CD), where a student learns sequentially from a stream o…

Domain Generalization

Weight Copy and Low-Rank Adaptation for Few-Shot Distillation of Vision Transformers

2024-04-14 · Diana-Nicoleta Grigore, Mariana-Iuliana Georgescu, Jon Alvarez Justo, Tor Johansen 외

Few-shot knowledge distillation recently emerged as a viable approach to harness the knowledge of large-scale pre-trained models, using limited data and computational resources. In this paper, we propose a novel few-shot…

Knowledge Distillation