Hello authors, thank you for your great work! I would like to ask whether the Multi-res Feature Grids and Hash Grids in Representations in sdfstudio-methods.md are the same? Or do you just use the hash encoding in instant-ngp to replace part of the MLP? Is it normal that the training speed does not improve when I use this ‘--pipeline.model.sdf-field.use-grid-feature True’? I would also like to ask if you have implemented the method in NeuS2 about applying multi-resolution hash encodings to SDF-based volume rendering?Looking forward to your reply!
Hello authors, thank you for your great work! I would like to ask whether the Multi-res Feature Grids and Hash Grids in Representations in sdfstudio-methods.md are the same? Or do you just use the hash encoding in instant-ngp to replace part of the MLP? Is it normal that the training speed does not improve when I use this ‘--pipeline.model.sdf-field.use-grid-feature True’? I would also like to ask if you have implemented the method in NeuS2 about applying multi-resolution hash encodings to SDF-based volume rendering?Looking forward to your reply!