meyerscetbon / deep-k-svd Goto Github PK

In load_model.py, the code import Deepnet_Denoising, but no model named Deepnet_Denosing. Did you forget to upload the file Deepnet_Denoising.py?

Hello, dear author, I am a student majoring in Statistics and also a new coder. After reading your article, I have a strong interest in turning the classical algorithm K-SVD into a differentiable and learnable end-to-end depth architecture. And your research inspired my work, so I want to cite your article in my work and compare it with my results in the experiment comparison part. However, I encountered the problem of too long training time when reproducing your experimental results. I hope you can give me some guidance on how to shorten the training time. You mentioned in the article that it takes two days to train a model with Titan xp GPU. But after data processing, I have obtained a total of 29668032 pictures. Even though there are only three epochs, it will take me more than two months to train a model with 2080Ti GPU. At present, I only want to increase the print "every" in the experiment. I don't know if I made any mistakes in the operation? I hope you can reply my question as soon as possible.

There are only three .py files that can be run in this code,Why is there not a predictive .py file？

How long does it take the model to be trained over CPU?