Thank you for the amazing work on this project! I’ve been training the vanilla GS model using the COLMAP dataset you provided, and I’m seeing a significant performance boost (+2 PSNR) compared to training on my own dataset.
I suspect the difference in performance might be due to variations in the pre-processing steps. Would it be possible for you to share the COLMAP pre-processing script you used for preparing the data?
Thanks again for your help!
Thank you for the amazing work on this project! I’ve been training the vanilla GS model using the COLMAP dataset you provided, and I’m seeing a significant performance boost (+2 PSNR) compared to training on my own dataset.
I suspect the difference in performance might be due to variations in the pre-processing steps. Would it be possible for you to share the COLMAP pre-processing script you used for preparing the data?
Thanks again for your help!