Implementation of convolutional projection

[leoxiaobin]

hi, @rishikksh20

Thank you for your quick implementation!

I notice that your depth-wise separable convolution is implemented as depth-wise conv --> point-wise conv. In the paper, CvT's depth-wise seprable convolution is implemented as depth-wise conv --> bn --> point-wise conv.

[rishikksh20]

@leoxiaobin oh my bad I just reuse my older SepCon2d class in this repo, I will update that class.

[rishikksh20]

Done, do let me know if you find any other vulnerabilities in code.

[huixiancheng]

@leoxiaobin Hi！ When will the official implementation be open source?

[youngwanLEE]

@leoxiaobin
The official code URL denoted in the paper is broken. Do you have a plan to open the code and model weights?

[leoxiaobin]

hi, @youngwanLEE, you can get the pre-trained model at https://github.com/microsoft/CvT

[leoxiaobin]

@leoxiaobin Hi！ When will the official implementation be open source?

Please find it at https://github.com/microsoft/CvT

[rishikksh20]

Thanks @leoxiaobin I am adding repo on ReadMe.