You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I use the 'finetune_realesrgan_x4plus_pairdata.yml' configuration file, and I only train the real-ESRGANx4 model instead of training the real-ESRNetx4 model first. However, I cant get a better result of finetune the model on validation dataset. The total number of my training dataset is about 6000 images while the number of validation dataset is about 800. The average psnr value is always about 24 no matter how many epochs I trained. Should I train the real-ESRNetx4 first, or should I use the multi scale script to make the data more abundant? Or should I lower the learning rate? Here's my configuration file looks like:
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
-
I use the 'finetune_realesrgan_x4plus_pairdata.yml' configuration file, and I only train the real-ESRGANx4 model instead of training the real-ESRNetx4 model first. However, I cant get a better result of finetune the model on validation dataset. The total number of my training dataset is about 6000 images while the number of validation dataset is about 800. The average psnr value is always about 24 no matter how many epochs I trained. Should I train the real-ESRNetx4 first, or should I use the multi scale script to make the data more abundant? Or should I lower the learning rate? Here's my configuration file looks like:
finetune_realesrgan_x4plus_pairdata.txt
Beta Was this translation helpful? Give feedback.
All reactions