English

Large-Scale Data-Free Knowledge Distillation for ImageNet via Multi-Resolution Data Generation

Computer Vision and Pattern Recognition 2024-11-27 v1

Abstract

Data-Free Knowledge Distillation (DFKD) is an advanced technique that enables knowledge transfer from a teacher model to a student model without relying on original training data. While DFKD methods have achieved success on smaller datasets like CIFAR10 and CIFAR100, they encounter challenges on larger, high-resolution datasets such as ImageNet. A primary issue with previous approaches is their generation of synthetic images at high resolutions (e.g., 224×224224 \times 224) without leveraging information from real images, often resulting in noisy images that lack essential class-specific features in large datasets. Additionally, the computational cost of generating the extensive data needed for effective knowledge transfer can be prohibitive. In this paper, we introduce MUlti-reSolution data-freE (MUSE) to address these limitations. MUSE generates images at lower resolutions while using Class Activation Maps (CAMs) to ensure that the generated images retain critical, class-specific features. To further enhance model diversity, we propose multi-resolution generation and embedding diversity techniques that strengthen latent space representations, leading to significant performance improvements. Experimental results demonstrate that MUSE achieves state-of-the-art performance across both small- and large-scale datasets, with notable performance gains of up to two digits in nearly all ImageNet and subset experiments. Code is available at https://github.com/tmtuan1307/muse.

Keywords

Cite

@article{arxiv.2411.17046,
  title  = {Large-Scale Data-Free Knowledge Distillation for ImageNet via Multi-Resolution Data Generation},
  author = {Minh-Tuan Tran and Trung Le and Xuan-May Le and Jianfei Cai and Mehrtash Harandi and Dinh Phung},
  journal= {arXiv preprint arXiv:2411.17046},
  year   = {2024}
}
R2 v1 2026-06-28T20:12:30.653Z