Issues

• Possible train–inference discrepancy in CameraHead KV-cache organization

  wzzheng/StreamVGGT#40 · Filter by author richardchen225 opened on Jul 24, 2026

• Question about reproducing Camera Pose Estimation results on CO3Dv2

  wzzheng/StreamVGGT#36 · Filter by author ShangNH opened on May 31, 2026

• No image_collection in ARKitScenes_Multi dataset

  wzzheng/StreamVGGT#34 · Filter by author CheungBH opened on Feb 26, 2026

• Training Dataset Scripts

  wzzheng/StreamVGGT#33 · Filter by author HITchina opened on Jan 20, 2026

• Apply for version of flash attention

  wzzheng/StreamVGGT#32 · Filter by author coldsakrua opened on Dec 17, 2025

• An earnest invitation about further explore. From an intern at Tsinghua AIR

  wzzheng/StreamVGGT#31 · Filter by author HuanLi0311 opened on Nov 16, 2025

• Possible to deploy it as TensorRT

  wzzheng/StreamVGGT#30 · Filter by author lipilian opened on Nov 11, 2025

• Duplicate Point cloud per camera

  wzzheng/StreamVGGT#29 · Filter by author Xchanger0321 opened on Nov 10, 2025

• compute_pose_loss not work

  wzzheng/StreamVGGT#28 · Filter by author vvuonghn opened on Oct 28, 2025

• Question about kv cache in camera head

  wzzheng/StreamVGGT#27 · Filter by author LinZhuoChen opened on Oct 13, 2025

• How much training time is required?

  wzzheng/StreamVGGT#26 · Filter by author kaikai23 opened on Oct 12, 2025

• Which tokens to be cached?

  wzzheng/StreamVGGT#25 · Filter by author kaikai23 opened on Oct 12, 2025

• How to process long videos

  wzzheng/StreamVGGT#24 · Filter by author CanCanZeng opened on Sep 26, 2025

• Issue with los script

  wzzheng/StreamVGGT#23 · Filter by author clementinegrethen opened on Sep 18, 2025

• Test Set and Validation

  wzzheng/StreamVGGT#22 · Filter by author StephenYangjz opened on Sep 1, 2025

• Unable to Reproduce the Constant-time incoming frame inference with released codebase

  wzzheng/StreamVGGT#21 · Filter by author MarkChenYutian opened on Aug 27, 2025
  3

• Visualization

  wzzheng/StreamVGGT#20 · Filter by author missTL opened on Aug 22, 2025

• different video lengths

  wzzheng/StreamVGGT#19 · Filter by author Kamp-Sun opened on Aug 22, 2025

• Abnormal multi-layer reconstruction results

  wzzheng/StreamVGGT#18 · Filter by author togekiss0603 opened on Aug 21, 2025
  2

• The target size of image in camera pose estimation

  wzzheng/StreamVGGT#17 · Filter by author silencexw opened on Aug 20, 2025

• Why "XFORMERS_AVAILABLE = False"?

  wzzheng/StreamVGGT#16 · Filter by author tyhuang98 opened on Aug 18, 2025

• Question about total training epoch numbers

  wzzheng/StreamVGGT#15 · Filter by author zhangshuoneu opened on Aug 18, 2025

• Dynamic Pointcloud Visualization

  wzzheng/StreamVGGT#13 · Filter by author ShunkaiZhou opened on Aug 14, 2025
  1

• Misalignment between points generated by each frame

  wzzheng/StreamVGGT#12 · Filter by author xiaoyuanzi22333 opened on Aug 5, 2025
  5

• OOM after ~70 images

  wzzheng/StreamVGGT#8 · Filter by author urbste opened on Jul 23, 2025
  1