search

sczhou / DAVANet

[CVPR 2019, Oral] DAVANet: Stereo Deblurring with View Aggregation
MIT License
137 stars 25 forks source link

Pretrained model of DeblurNet #13

Closed oaishi closed 4 years ago

oaishi commented 4 years ago

Dear Shangchen Zhou, for my research scope, I needed to infer the results produced by the single DeblurNet. I followed the procedure described (here) and was able to switch to do the Single Deblurnet configuration. I shall be grateful if you kindly help me to get the checkpoint of the single deblurnet. Specifically, I am looking for the pre-trained model which is referred as "Ours-Single" in table 1 of the paper.

codebert

Thanking you, Oaishi.

sczhou commented 4 years ago

Hi Oaishi,

I am so sorry that the Single Deblurnet model can not be found since I have left the previous company last year and I didn't keep a local copy for myself. However, it should be easy to reproduce the training process of Single Deblurnet by slightly modifying the training code. I would be glad to answer the questions if you would like to reproduce the training by yourself.