search

justimyhxu / GRM

Large Gaussian Reconstruction Model for Efficient 3D Reconstruction and Generation
https://justimyhxu.github.io/projects/grm/
503 stars 27 forks source link

How many additional views used for each step of training ? #10

Open lizhiqi49 opened 3 months ago

justimyhxu commented 3 months ago

Typically, we use four additional views for supervision.