Question about result of generalization model

[blue-531]

Thank you for sharing your interesting work!

Using generalization model parameter uploaded in the github, I had difference between the results in the paper and my results.
PSNR of Nerf synthetic in paper: 27.29
PSNR of Nerf synthetic reproduced : 25.327
PSNR of llff in paper: 25.86
PSNR of llff reproduced : 25.642

Is it correct that the checkpoints of the performance published in the paper were uploaded properly? If not, can I get a checkpoint for that performance?

[MukundVarmaT]

#23 (comment)