search

BR-IDL / PaddleViT

:robot: PaddleViT: State-of-the-art Visual Transformer and MLP Models for PaddlePaddle 2.0+
https://github.com/BR-IDL/PaddleViT
Apache License 2.0
1.22k stars 318 forks source link

trans2seg and focal #88

Closed GuoQuanhao closed 2 years ago

GuoQuanhao commented 2 years ago

增加新的数据读取方式，通过get_dataloader获取数据迭代

增加新的数据增强，通过get_transforms获取数据增强，仅新增Trans10kV2并保留伍老师之前的transforms_train

增加新的损失函数，用于对齐Trans2Seg，get_loss_function获取损失函数

增加新的学习率优化方法，通过get_scheduler获取

增加新的优化器获取方法，通过get_optimizer获取

占位Focal Transformer for segmentation，能train

Trans2seg PaddleViT训练精度0.7597高于原始作者0.72