MSCViT, a small hybrid ViT with multi-scale self-attention and wavelet-based convolutional fusion, reaches 84.68% on CIFAR-100 from scratch with 14.0M parameters.
Tokens-to- token vit: Training vision transformers from scratch on imagenet
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
MSCViT: A Small-size ViT architecture with Multi-Scale Self-Attention Mechanism for Tiny Datasets
MSCViT, a small hybrid ViT with multi-scale self-attention and wavelet-based convolutional fusion, reaches 84.68% on CIFAR-100 from scratch with 14.0M parameters.