alvin870203/improved-ViT

Improving Vision Transformer (ViT) Accuracy and Converged Time | 圖形分析辨認 | Pattern Analysis and Classification | NTU | 洪一平教授 | 2021 Spring

Improving ViT accuracy and converged time

FInal Project for: 圖形分析辨認 | Pattern Analysis and Classification | NTU | 洪一平教授 | 2021 Spring

Slides

Cover

Goal

Experiment

Dataset

Optimizer

Metric Learning

Multi-task Method

Conclusion

Reference