search

dvlab-research / ECCV22-P3AFormer-Tracking-Objects-as-Pixel-wise-Distributions

The official code for our ECCV22 oral paper: tracking objects as pixel-wise distributions.
Other
159 stars 4 forks source link

One question about Swin-B Transformer #17

Open monstre0731 opened 1 year ago

monstre0731 commented 1 year ago

Hi,

Thanks a lot for your amazing work.

I have a question about the implementation of P3AFormer on KITTI. As shown in your paper, Swin-B is used as the backbone. How did you handle the resolution problem that the shape of the images in KITTI are 1280 * 384, but the pre-trained models of Swin-Transformer are all with a square resolution.

Thanks!

Hi again,

Could you please upload the code on KITTI dataset?

Thanks!

ddenglina commented 4 months ago

@monstre0731 Have you solved this problem?I have the same problem.