Juno

I appreciate the detailed check. I agree with your observations. In my experiments, I used a sparse point cloud, which might have left some parts of the point cloud uninitialized....

I think this issue is also related to random initialization because some ["dead Gaussians"](https://ubc-vision.github.io/3dgs-mcmc/) are not entirely pruned. As far as I know, with around 30MB, there is a difference...

> I checked the dynamic_point_cloud.ply of the pre-trained model and my trained model and found some unusual differences, in my dynamic point cloud my `motion_xyz` is only 36 time points,...

I understand this issue arises because CPU image loading is faster than GPU processing. Additionally, the dataloading process relies on `joblib`. Unfortunately, I have not found a complete solution within...