Issues

• Mixed-precision training workflow for DrivAerML?

  thuml/Transolver-3#6 · Filter by author vpuri3 opened on Aug 14, 2026

• Eval mode 2 does part-wise Rel-L2 averaging, not joint full-mesh Rel-L2

  thuml/Transolver-3#5 · Filter by author vpuri3 opened on Aug 4, 2026

• Clarify train loss vs reported Rel-L2 (NASA-CRM / AhmedML / DrivAerML)

  thuml/Transolver-3#4 · Filter by author vpuri3 opened on Aug 1, 2026

• Request for exact AhmedML and DrivAerML train/validation/test splits

  thuml/Transolver-3#3 · Filter by author vpuri3 opened on Jul 28, 2026
  2

• Release Transolver-3 checkpoints on Hugging Face

  thuml/Transolver-3#2 · Filter by author NielsRogge opened on Jun 30, 2026

• About DriveML Dataset

  thuml/Transolver-3#1 · Filter by author AbramsDkr opened on Jun 29, 2026