We explore the impact of training with more diverse datasets, characterized by the number of unique samples, on the performance of self-supervised learning (SSL) under a fixed computational budget. Our findings consistently demonstrate that increasing pretraining data diversity enhances SSL performance, albeit only when the distribution distance to the downstream data is minimal. Notably, even with an exceptionally large pretraining data diversity achieved through methods like web crawling or diffusion-generated data, among other ways, the distribution shift remains a challenge. Our experiments are comprehensive with seven SSL methods using large-scale datasets such as ImageNet and YFCC100M amounting to over 200 GPU days. Code and trained models will be available at https://github.com/hammoudhasan/DiversitySSL .

翻译：我们探讨了在固定计算预算下，通过增加数据集多样性（以独特样本数量为特征）对自监督学习（SSL）性能的影响。研究结果一致表明，增加预训练数据多样性能够提升SSL性能，但前提是数据分布与下游数据的分布距离需尽可能小。值得注意的是，即使通过网页爬取或扩散生成数据等方法实现了极高的预训练数据多样性，分布偏移问题依然存在。我们通过七种SSL方法进行了全面实验，使用了ImageNet和YFCC100M等大规模数据集，累计计算量超过200 GPU天。相关代码和预训练模型将在 https://github.com/hammoudhasan/DiversitySSL 开源。