How's the pre-trained model for Conversion from day-to-night trained?

[ankur-chr]

Hi,

There is a pre-trained model linked in the description which converts day images to nighttime images. How was this model trained?
I read the paper but am confused about the way the CycleGAN model was trained for the day-to-night conversion of images in the pre-processing step.

@sauradip could you please share more information about this? Which datasets were used for the model training in this case? What was the input and output size of images for this pre-trained model?