Shortcuts

MixMIM

Abstract

In this study, we propose Mixed and Masked Image Modeling (MixMIM), a simple but efficient MIM method that is applicable to various hierarchical Vision Transformers. Existing MIM methods replace a random subset of input tokens with a special [MASK] symbol and aim at reconstructing original image tokens from the corrupted image. However, we find that using the [MASK] symbol greatly slows down the training and causes training-finetuning inconsistency, due to the large masking ratio (e.g., 40% in BEiT). In contrast, we replace the masked tokens of one image with visible tokens of another image, i.e., creating a mixed image. We then conduct dual reconstruction to reconstruct the original two images from the mixed input, which significantly improves efficiency. While MixMIM can be applied to various architectures, this paper explores a simpler but stronger hierarchical Transformer, and scales with MixMIM-B, -L, and -H. Empirical results demonstrate that MixMIM can learn high-quality visual representations efficiently. Notably, MixMIM-B with 88M parameters achieves 85.1% top-1 accuracy on ImageNet-1K by pretraining for 600 epochs, setting a new record for neural networks with comparable model sizes (e.g., ViT-B) among MIM methods. Besides, its transferring performances on the other 6 datasets show MixMIM has better FLOPs / performance tradeoff than previous MIM methods

Models and Benchmarks

Here, we report the results of the model on ImageNet, the details are below:

Algorithm Backbone Epoch Batch Size Results (Top-1 %) Links
Fine-tuning Pretrain Fine-tuning
MixMIM MixMIM-base 300 2048 84.63 config | model | log config | model | log

Citation

@article{MixMIM2022,
  author  = {Jihao Liu, Xin Huang, Yu Liu, Hongsheng Li},
  journal = {arXiv:2205.13137},
  title   = {MixMIM: Mixed and Masked Image Modeling for Efficient Visual Representation Learning},
  year    = {2022},
}
Read the Docs v: stable
Versions
latest
stable
1.x
dev-1.x
0.x
Downloads
pdf
html
epub
On Read the Docs
Project Home
Builds

Free document hosting provided by Read the Docs.