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

CNN vs. ViT, 牛久先生 #489

Open
AkihikoWatanabe opened this issue Oct 27, 2022 · 2 comments
Open

CNN vs. ViT, 牛久先生 #489

AkihikoWatanabe opened this issue Oct 27, 2022 · 2 comments

Comments

@AkihikoWatanabe
Copy link
Owner

https://speakerdeck.com/yushiku/cnn-vs-vit

@AkihikoWatanabe
Copy link
Owner Author

・Swin Transformer, Depth-wise conv, ConvNeXt, ViTとCNNのロバスト性の違いの話があり勉強になる
・最終的な結論が、CNNもTransformerも変わらない(明確な勝者はいない; 今のところ引き分け)というのはおもしろかった

@AkihikoWatanabe
Copy link
Owner Author

depth-wise conv, point-wise convの解説記事:https://agirobots.com/depthwise-pointwise-convolution/

通常のCNNのフィルタによるfeature map計算を、空間方向(depth-wise conv)とチャネル方向(point-wise conv; 1x1 conv)に分解することで大幅にパラメータ数削減

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

No branches or pull requests

1 participant