Sampling strategy for train images

[Kush0301]

Thank you for this amazing work! I wanted to know if there was a reason for not sampling the train images randomly but in a sequential manner (t-1, t, t+1). I got significantly better results on a custom dataset by following nerfstudio's random sampling strategy while keeping the rest of the code same.

[maturk]

@Kush0301, thanks for pointing this out. There is really no reason to use the sequential manner, I mainly used this for debugging and sanity checks when developing the model. But you are right, this can be updated. I will work on a revision of the code-base soon.