Improved Baselines with Momentum Contrastive Learning
This incremental work makes state-of-the-art unsupervised learning more accessible for researchers by enhancing existing methods.
The authors tackled the problem of improving unsupervised learning baselines by integrating SimCLR's design improvements into the Momentum Contrast (MoCo) framework, resulting in stronger baselines that outperform SimCLR without needing large training batches.
Contrastive unsupervised learning has recently shown encouraging progress, e.g., in Momentum Contrast (MoCo) and SimCLR. In this note, we verify the effectiveness of two of SimCLR's design improvements by implementing them in the MoCo framework. With simple modifications to MoCo---namely, using an MLP projection head and more data augmentation---we establish stronger baselines that outperform SimCLR and do not require large training batches. We hope this will make state-of-the-art unsupervised learning research more accessible. Code will be made public.