[1] MLP-Mixer: An all-MLP Architecture for Vision - Google Research
[2] Beyond Self-attention: External Attention using Two Linear Layers for Visual Tasks - 清华大学
[3] Do You Even Need Attention? A Stack of Feed-Forward Layers Does Surprisingly Well on ImageNet - 牛津大学
[4] Are Pre-trained Convolutions Better than Pre-trained Transformers? - Google Research
[5] ResMLP: Feedforward networks for image classification with data-efficient training - Facebook AI
[6] FNet: Mixing Tokens with Fourier Transforms - Google Research
[7] Pay Attention to MLPs - Google Research
对于以上论文的浅谈:也来盘点一些最近的非Transformer工作
MLP替代Transformer浅谈
最新推荐文章于 2025-09-15 15:44:07 发布
本文总结了Google、清华大学、牛津大学等机构近期的研究成果,包括MLP-Mixer、外部注意力模型、全连接层替代注意力、预训练卷积与Transformer的比较、ResMLP等。这些工作展示了非Transformer架构在图像识别任务中的潜力,探讨了其在数据效率和性能上的新突破。
3012

被折叠的 条评论
为什么被折叠?



