Open muximuxi opened 6 months ago
Thanks for your interest, it looks like the pose is not correctly aligned, have you tried to see what is the initial visualization? i.e. at step 0, only init the GS on the SMPL mesh and not optimized yet, in that case, we can check whether the pose is roughly aligned with the video.
Thanks for your reply, I have fixed this pose issue,it seems the poses/anim_nerf_train.npz, anim_nerf_test.npz does not correct, how can I get this data or how to calculate it ?
I use the data processed by the instantAvatar, changed the code to read the data, after using bash ./script/fit_ubc.sh. I can't get result like this: can you give me some suggestions?
how can I get the cameras.npz and train_pose.npz? when I use the data which processed by the instantAvatar, after ./scripts/fit.sh , the resutl is bad, like this: can you give me some suggestions?
Where is fit.sh? Which yaml profile should I choose to run on my custom data?
how can I get the cameras.npz and train_pose.npz? when I use the data which processed by the instantAvatar, after ./scripts/fit.sh , the resutl is bad, like this: can you give me some suggestions?![test](https://github.com/JiahuiLei/GART/assets/19643523/5e45edf7-3846-4051-95eb-80f7ff6e56ef)