Three Guidelines You Should Know for Universally Slimmable Self-Supervised Learning
This work addresses the challenge of efficient model deployment for self-supervised learning across diverse devices, representing an incremental improvement with specific guidelines for loss design.
The paper tackles the problem of deploying self-supervised models across devices by proposing universally slimmable self-supervised learning (US3L), which achieves better accuracy-efficiency trade-offs, outperforming state-of-the-art methods on benchmarks like recognition, object detection, and instance segmentation with only one training and one set of weights.
We propose universally slimmable self-supervised learning (dubbed as US3L) to achieve better accuracy-efficiency trade-offs for deploying self-supervised models across different devices. We observe that direct adaptation of self-supervised learning (SSL) to universally slimmable networks misbehaves as the training process frequently collapses. We then discover that temporal consistent guidance is the key to the success of SSL for universally slimmable networks, and we propose three guidelines for the loss design to ensure this temporal consistency from a unified gradient perspective. Moreover, we propose dynamic sampling and group regularization strategies to simultaneously improve training efficiency and accuracy. Our US3L method has been empirically validated on both convolutional neural networks and vision transformers. With only once training and one copy of weights, our method outperforms various state-of-the-art methods (individually trained or not) on benchmarks including recognition, object detection and instance segmentation. Our code is available at https://github.com/megvii-research/US3L-CVPR2023.