Hi every one,

can someone help me to understand how can I calculate the probability of an image in PyTorch?

If there is some function in PyTorch that calculates it immediately?

Could you explain a bit more what you mean by “probability of an image”?

Are you interested in the histogram or some sampling strategy?

I don’t know if it’s a histogram, that is a question in itself, but the mean idea is to strengthen the non-similarity between images, so I want to do it with probability and without the traditional way like euclidean distance.

Based on the description it seems you are generating images and would like to do it in a diverse way?

I guess a GAN could do the job and base on the released number of papers I assume someone might have even worked on it.