Search results

Training performance is suboptimal for images with multiple sequences and significant perspective differences.

Status: Open.

#17 In Linketic/VGGT-X;

· sly061 opened on Jan 8, 2026

Found error when facing masked images.

Status: Open.

#15 In Linketic/VGGT-X;

· Zhaoyibinn opened on Dec 9, 2025

3DGS training ---Colmap Artifacts from VGGT-X - Not good results

Status: Open.

#14 In Linketic/VGGT-X;

· Jaykumaran opened on Dec 4, 2025

index 0 is out of bounds for axis 0 with size 0

Status: Open.

#12 In Linketic/VGGT-X;

· LMeteorYu0330 opened on Dec 2, 2025

点云顶部会有很多黑色噪声点

Status: Open.

#9 In Linketic/VGGT-X;

· kuaiqushangzixiba opened on Nov 11, 2025

修改阈值

Status: Open.

#6 In Linketic/VGGT-X;

· Lfg111 opened on Nov 6, 2025

好奇论文中的“Although VGGT contains 24 AA layers, only the output features from layers 4, 11, 17, and 23 are utilized for dense prediction.”是怎么得出的？

Status: Open.

#5 In Linketic/VGGT-X;

· zxwang2018 opened on Nov 5, 2025

Sparse models have no 3D point clouds

Status: Open.

#4 In Linketic/VGGT-X;

· zwl995 opened on Nov 4, 2025

论文中1000+帧输入，显存消耗问题

Status: Open.

#3 In Linketic/VGGT-X;

· DavidQiuChao opened on Oct 31, 2025

'pycolmap._core.Image' object has no attribute 'id'

Status: Open.

#2 In Linketic/VGGT-X;

· Kuni05 opened on Oct 29, 2025

Release VGGT-X artifacts (models, dataset) on Hugging Face

Status: Open.

#1 In Linketic/VGGT-X;

· NielsRogge opened on Oct 1, 2025