ljzycmd / vdtr Goto Github PK

Dear authors,

Thank you for sharing your code and paper. Where can I find the checkpoints of the VDTR trained on a real blur dataset (BSD)? I only noticed GoPro and DVD checkpoints.

Sincerely,
Magauiya

Hello! Can you share how you calculate the GFLOPs and Params. I used thop to calculate and got your model GFLOPs = 1362.78 and Params = 16.01M.

Hi, when I use the testing script python test.py ./configs/vdtr/vdtr_dvd.yaml $Checkpoint_path you provided for testing on the DVD dataset, only 300 images are de-blurred, but the test set for the DVD dataset has 1000 images. I wonder if your code can deblur each image when testing?

This looks very interesting.
Will there be code available for testing soon?