Hello,
Thank you for providing the code. This code is really helpful.
I find the training code is too long (for a single file) and it is hard to follow :
https://github.com/ericsujw/KpSFR/blob/3d4b7968e2ab34239bf4c175f17a29efa46a691f/train_nn.py
Hence, I tried to modularize it so that it can be easily followed and understandable. Please note that I have hardly modified some code with respect to the original code. I only have modularised it for better understanding
These codes are running like the original code. I thought some other people may feel the same issues in understanding and following the code. That's why I share these codes. Please feel free to explore them and let me know if you find any issues. This will be helpful
The Python files are attached herein.
https://drive.google.com/file/d/14vYzEI9PAfN6kL4TvWcsdM9GPDlpaUes/view?usp=sharing
https://drive.google.com/file/d/1QwQF75HA3E0W5EVU_mTzVgDb2c30qYdl/view?usp=sharing