Shortcuts

MoCo v3

Abstract

This paper does not describe a novel method. Instead, it studies a straightforward, incremental, yet must-know baseline given the recent progress in computer vision: self-supervised learning for Vision Transformers (ViT). While the training recipes for standard convolutional networks have been highly mature and robust, the recipes for ViT are yet to be built, especially in the self-supervised scenarios where training becomes more challenging. In this work, we go back to basics and investigate the effects of several fundamental components for training self-supervised ViT. We observe that instability is a major issue that degrades accuracy, and it can be hidden by apparently good results. We reveal that these results are indeed partial failure, and they can be improved when training is made more stable. We benchmark ViT results in MoCo v3 and several other self-supervised frameworks, with ablations in various aspects. We discuss the currently positive evidence as well as challenges and open questions. We hope that this work will provide useful data points and experience for future research.

Models and Benchmarks

In this page, we provide benchmarks as much as possible to evaluate our pre-trained models. If not mentioned, all models are pre-trained on ImageNet-1k dataset.

Algorithm Backbone Epoch Batch Size Results (Top-1 %) Links
Linear Eval Fine-tuning Pretrain Linear Eval Fine-tuning
MoCo v3 ResNet50 100 4096 69.6 / config | model | log config | model | log /
ResNet50 300 4096 72.8 / config | model | log config | model | log /
ResNet50 800 4096 74.4 / config | model | log config | model | log /
ViT-small 300 4096 73.6 / config | model | log config | model | log /
ViT-base 300 4096 76.9 83.0 config | model | log config | model | log config | model | log
ViT-large 300 4096 / 83.7 config | model | log / config | model | log

Citation

@InProceedings{Chen_2021_ICCV,
    title     = {An Empirical Study of Training Self-Supervised Vision Transformers},
    author    = {Chen, Xinlei and Xie, Saining and He, Kaiming},
    booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)},
    year      = {2021}
}
Read the Docs v: dev-1.x
Versions
latest
stable
1.x
dev-1.x
dev
Downloads
pdf
html
epub
On Read the Docs
Project Home
Builds

Free document hosting provided by Read the Docs.