How to put Image not noise as input of Gaussian Diffusiuon Network

[SeunghyunKim1995]

I succeeded in learning guided diffusion using my custom data set. To check the performance of my neural network, I succeeded in generating a moderately noisy image through about 100 steps from noise. Is there any way to use this noisy image as an input for a neural network instead of noise?

After fetching the image, tensorizing it, and inputting it, the range of tensor values was limited from 0 to 1, resulting in only white images

Is there any genius who knows the image preprocessing process to solve this problem?

[2144533854]

I think this method is not feasible, The noise has to be Gaussian noise.