Hello! i've attempted to follow your instructions, but i can't seem to find where the calibration folder in the dataset is created from! it's not present in the RegNeRF guide you've given link to.

Great paper, can't wait to try it out!

Yiftach

Thanks for your excellent work! I would like to ask if the training and testing dataset were conducted separately on the DTU, LLFF, and RealEstate datasets. For example, the model trained on LLFF and then tested on LLFF.

Thanks for your contribution! Now I want to test this method on sparse urban scene. I used the pretrained mipnerf360 checkpoint because it is suited for the unbounded scene. Now I want to finetune on my own urban datasets. Could you provide the finetune scripts which is not in this repro. Any response will be greatly appreciated.

Hello, I have some questions about model training that I would like to ask you.
Why the loss I trained with the dtu dataset cannot converge？
How should I use other datasets for training?
Is it easy to train?
Will the effect be the same as the effect you showed?

Thanks @haofeixu, for making this amazing work publicly available!
I'm just trying to run training on Realestate10k datasets (I got the full set already), but I noticed that the config yaml files for training on realestate or mixed datasets are not there. Will it be shared? If it's not, it's fine, I just want to ask if the learning rates, optimizer , loss weights the same to the setting used for DTU.

Thanks!