Training Similar to ARKit data

[nmerrillq]

Hello,

Thank you for open-sourcing this great work. I am trying to train on some custom data with noisy poses similar to your ARKit experiments. I found the config for CamP with ARKit poses, but I am wondering which ZIP-NeRF config you used for that training? For now I am using configs/zipnerf/360.gin, and training with configs/zipnerf/use_multiscale.gin just threw errors. Any guidance here as to what config was used for the paper results would be helpful. Thanks.