Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

The configuration of the architecture variants is inconsistent with the papers and weights files. #2

Closed
lartpang opened this issue Feb 19, 2022 · 1 comment

Comments

@lartpang
Copy link

lartpang commented Feb 19, 2022

@tangchuanxin

model = ShiftViT(embed_dim=96, depths=(3, 3, 24, 4), mlp_ratio=2, drop_path_rate=0.2, n_div=12)

The code is inconsistent with the content of the paper:

image

and the weight file. The content of this pth file is the same as the architecture variant -S in the figure above, ie, depths=(6, 8, 18, 6).

https://github.com/microsoft/SPACH/releases/download/v1.0/shiftvit_tiny_r2.pth

@wgting96
Copy link
Contributor

Thank you for pointing this error out. The configuration [depths=(6, 8, 18, 6)] is correct. I will fix this bug soon.

The configuration of (3, 3, 24, 4) is another version of ShiftViT-T we have tried. It has the similar number of parameters and FLOPs.

Config MLP ratio Num of params. (M) FLOPs (G)
[6, 8, 18, 6] 2.0 29 4.6
[3, 3, 24, 4] 2.0 27 4.1

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants