What should the structure of the dataset be like?

[UriyMihailov]

Hi! I have a question regarding the datasets. As far as I understand, the minimum amount of data that is required to work with H-Loc is just a number of images located together in some folder. But what should be the preferable structure of the dataset, aside from the images themselves? As I understand, it would be nice to also have information about the camera's intrinsic parameters. Also, as I figured, it would be great to have prior information about the camera poses. If what I said is right, then I have another important question: what should be the preferable structure of the dataset? What I mean is - how should the files be located inside the folders, and how should the files be structured themselves (the info contained inside, its order, file's extensions, etc.). It would be really great to know that to be able to correctly prepare the necessary data.