Issues

• 🐛[BUG]: Darcy2D Jacobi kernel divides grad(k)·grad(u) by 2dx instead of 4dx², so the data doesn't solve the Darcy equation

  NVIDIA/physicsnemo#2036 · Filter by author aaguadob opened on Sep 30, 2026

• 🐛[BUG]: Darcy2D permeability: all Fourier coefficients of a sample share one random value (bare wp.tid() in rand_init)

  NVIDIA/physicsnemo#2035 · Filter by author aaguadob opened on Sep 30, 2026

• 🚀[FEA]: Update FLARE++ to the v2 reference design and add token-sharded context parallelism for FLARE/FLARE++

  NVIDIA/physicsnemo#2024 · Filter by author vpuri3 opened on Sep 29, 2026
  2

• 🐛[BUG]: MHD_PDE equation "B_grad_u" uses v * Bx.diff(y) instead of By * u.diff(y)

  ? - Needs Triagebug
  NVIDIA/physicsnemo#2021 · Filter by author aaguadob opened on Sep 28, 2026

• 🐛[BUG]: Diffusion(time=True) in darcy_physics_informed/utils.py silently drops the time-derivative term

  ? - Needs Triagebug
  NVIDIA/physicsnemo#2020 · Filter by author aaguadob opened on Sep 28, 2026

• Checkpoint loading can mix model weights and training state from different epochs

  bug
  NVIDIA/physicsnemo#2012 · Filter by author peterdsharpe opened on Sep 19, 2026

• 🐛[BUG]: power_spectrum in metrics.general puts the zero frequency half a cell off along an odd width

  NVIDIA/physicsnemo#2007 · Filter by author Nicholas022400701 opened on Sep 18, 2026
  2

• 🐛[BUG]: cdf in metrics.general.histogram drops earlier inputs from the last cumulative bin

  NVIDIA/physicsnemo#2005 · Filter by author Nicholas022400701 opened on Sep 18, 2026

• 🐛[BUG]: Variance in metrics.general.ensemble_metrics ignores the dim argument

  NVIDIA/physicsnemo#2003 · Filter by author Nicholas022400701 opened on Sep 18, 2026

• GradientsFiniteDifference/GradientsSpectral (and PhysicsInformer) silently assume a periodic domain — non-periodic BCs get a wrong residual at the boundary with no warning

  NVIDIA/physicsnemo#2001 · Filter by author barrosyan opened on Sep 17, 2026
  1

• 🐛[BUG]: AeroJEPA ignores unbatched TokenSet masks and decode_field_chunked accepts invalid arguments

  NVIDIA/physicsnemo#1998 · Filter by author harshaa765 opened on Sep 17, 2026

• 🐛[BUG]: ValueError: infer_schema(func) in @torch.library.custom_op due to stringified type annotations (from __future__ import annotations)

  ? - Needs Triagebug
  NVIDIA/physicsnemo#1990 · Filter by author DenizCanELCI opened on Sep 14, 2026
  

• 🐛[BUG]: torch.compile + ShardTensor failures with regional model recipe

  ? - Needs Triagebug
  NVIDIA/physicsnemo#1986 · Filter by author jleinonen opened on Sep 11, 2026
  3

• DistributedManager initialization fails on CPU-only hosts

  NVIDIA/physicsnemo#1936 · Filter by author pjaysaval opened on Aug 19, 2026

• DiffusionUNet3D cannot run under torch.autocast (amp_mode is not exposed)

  NVIDIA/physicsnemo#1935 · Filter by author pjaysaval opened on Aug 19, 2026

• 🐛[BUG]: DefaultTrainingLoop performs two optimizer steps per minibatch with static capture

  ? - Needs Triagebug
  NVIDIA/physicsnemo#1911 · Filter by author mnabian opened on Aug 12, 2026

• 🚀[FEA]: Add non-periodic boundary mode to grid-gradient operators

  ? - Needs Triageenhancementstale
  NVIDIA/physicsnemo#1852 · Filter by author loliverhennigh opened on Jul 17, 2026
  1

• 🚀[FEA]: PhysicsNeMo-Mesh Code Cleanups

  1 - On Deckenhancement
  NVIDIA/physicsnemo#1851 · Filter by author peterdsharpe opened on Jul 17, 2026
  1

  

• ShardTensor is not torch.compile-able: graph break in from_local resumes mid-construction before _spec is attached

  NVIDIA/physicsnemo#1796 · Filter by author negin513 opened on Jul 6, 2026
  2

• 🚀[FEA]: allow control over checkpoint filename format in save_checkoint

  enhancement
  Enhancement

  NVIDIA/physicsnemo#1175 · Filter by author CharlelieLrt opened on Oct 23, 2025
  5

  

• 🚀[FEA]: allow renaming state-dict keys in from_checkpoint when loading backward-compatible checkpoints

  enhancement
  Enhancement

  NVIDIA/physicsnemo#1173 · Filter by author CharlelieLrt opened on Oct 23, 2025
  5