SimMIM: A Simple Framework for Masked Image Modeling
This work addresses the challenge of efficient and effective pre-training for computer vision models, offering a simpler alternative to existing masked image modeling approaches, though it is incremental in simplifying prior methods.
The paper tackles the problem of masked image modeling for representation learning by proposing SimMIM, a simplified framework that eliminates complex designs like block-wise masking and tokenization, and finds that simple components such as random masking, direct regression of raw pixels, and a linear prediction head achieve strong performance. It results in 83.8% top-1 accuracy on ImageNet-1K with ViT-B, surpassing previous methods by +0.6%, and enables state-of-the-art results on larger models like SwinV2-H and SwinV2-G with less data.
This paper presents SimMIM, a simple framework for masked image modeling. We simplify recently proposed related approaches without special designs such as block-wise masking and tokenization via discrete VAE or clustering. To study what let the masked image modeling task learn good representations, we systematically study the major components in our framework, and find that simple designs of each component have revealed very strong representation learning performance: 1) random masking of the input image with a moderately large masked patch size (e.g., 32) makes a strong pre-text task; 2) predicting raw pixels of RGB values by direct regression performs no worse than the patch classification approaches with complex designs; 3) the prediction head can be as light as a linear layer, with no worse performance than heavier ones. Using ViT-B, our approach achieves 83.8% top-1 fine-tuning accuracy on ImageNet-1K by pre-training also on this dataset, surpassing previous best approach by +0.6%. When applied on a larger model of about 650 million parameters, SwinV2-H, it achieves 87.1% top-1 accuracy on ImageNet-1K using only ImageNet-1K data. We also leverage this approach to facilitate the training of a 3B model (SwinV2-G), that by $40\times$ less data than that in previous practice, we achieve the state-of-the-art on four representative vision benchmarks. The code and models will be publicly available at https://github.com/microsoft/SimMIM.