BriefGPT.xyz
Feb, 2024
小规模数据上轻量级视觉变形器的预训练与图像最小缩放
Pre-training of Lightweight Vision Transformers on Small Datasets with Minimally Scaled Images
HTML
PDF
Jen Hong Tan
TL;DR
轻量级视觉Transformer（ViT）可以通过预训练和最小的图像缩放，实现优于ResNet等卷积神经网络在小数据集和小图像分辨率上的性能，而不需要显著地增大图像。
Abstract
Can a lightweight
vision transformer
(ViT) match or exceed the
performance
of
convolutional neural networks
(CNNs) like ResNet on
→