microsoft / CSWin-Transformer

CSWin Transformer: A General Vision Transformer Backbone with Cross-Shaped, CVPR 2022
MIT License
539 stars 78 forks source link

Error about building 384 models #6

Closed TingquanGao closed 3 years ago

TingquanGao commented 3 years ago

The code: https://github.com/microsoft/CSWin-Transformer/blob/d8be74a7833898f7bd9c77eb8c051d1b8bd5d753/models/cswin.py#L407 shoud be:

model = CSWinTransformer(img_size=384, patch_size=4, embed_dim=96, depth=[2,4,32,2],

And as the same: https://github.com/microsoft/CSWin-Transformer/blob/d8be74a7833898f7bd9c77eb8c051d1b8bd5d753/models/cswin.py#L414

LightDXY commented 3 years ago

We set the image size as an input parameter. https://github.com/microsoft/CSWin-Transformer/blob/d8be74a7833898f7bd9c77eb8c051d1b8bd5d753/main.py#L312

and as the same https://github.com/microsoft/CSWin-Transformer/blob/d8be74a7833898f7bd9c77eb8c051d1b8bd5d753/finetune.py#L449

TingquanGao commented 3 years ago

OK, thx for reply.