All pull requests

Pull requests

• Index 0-d MakeVector inputs instead of calling .item() in numba

  #2503·syahra712 opened Oct 16, 2026·4 of 11 tasks·

  ·
  1

• Accept check_finite in cho_solve

  #2502·VenishPaneliya opened Oct 16, 2026Contributor·

  ·

• Fix local_slice_read_of_write on multi-axis advanced write slice reads

  #2499·ADITYASINGH1206 opened Oct 15, 2026·

  ·
  1

  6

• Keep unbatched scalar inputs static in JAX Blockwise

  bugjaxscanvectorization
  #2498·jessegrabowski opened Oct 14, 2026Member·

  ·
  1

  5

• Bump the github-actions group across 1 directory with 12 updates

  GitHub CI/CDmajorno releasenotes
  #2497·dependabot[bot] opened Oct 14, 2026Bot·

  ·

• Fix multi-axis sample variance and standard deviation ddof normalization

  bugNumPy compatibility
  #2495·ADITYASINGH1206 opened Oct 14, 2026·5 of 11 tasks·

  ·
  1

  3

• Support len on tensors with a known first axis length

  #2494·err0rgod opened Oct 13, 2026·7 of 11 tasks·

  ·
  1

  2

• Implement native Numba BFGS minimization

  #2492·ricardoV94 opened Oct 13, 2026Member·

  ·
  4

• Remove DimShuffle from reductions over every axis

  #2487·AnKiT-GaRG2 opened Oct 12, 2026Contributor·3 tasks done·

  ·
  3

• Speedup variadic Elemiwse Add/Mul in Python mode

  maintenanceOp implementationvectorization
  #2483·AnKiT-GaRG2 opened Oct 12, 2026Contributor·

  ·
  1

  17

• Reuse SciPy Cython special functions in the C backend

  #2482·liuyinchu opened Oct 12, 2026·

  ·
  1

• Implement lazy ifelse in non-vm backends

  backend compatibilityenhancement
  #2481·ricardoV94 opened Oct 12, 2026Member·

  ·
  1

• fix: stop coercing symbolic integer indices to int in MLX Subtensor/IncSubtensor

  indexingmlx
  #2475·AnKiT-GaRG2 opened Oct 11, 2026Contributor·5 tasks done·

  ·
  1

• fix: enable Numba boundscheck for advanced integer indexing

  #2474·AnKiT-GaRG2 opened Oct 11, 2026Contributor·6 tasks done·

  ·
  1

• Add some sparse ops and rename others to have more explicit names

  #2469·tomicapretto opened Oct 9, 2026Contributor·6 of 9 tasks·

  ·
  13

• Update Usmm, fix UsmmCscDense broadcasting, and implement them in Numba

  numbasparse variables
  #2466·tomicapretto opened Oct 9, 2026Contributor·6 of 9 tasks·

  ·
  3

• Remove lossless integer casts from Alloc and Reshape shapes

  #2463·raashish1601 opened Oct 7, 2026Contributor·5 of 11 tasks·

  ·
  1

  4

• Drop axis-invariant terms from the input of softmax and logsumexp

  #2460·raashish1601 opened Oct 7, 2026Contributor·5 of 11 tasks·

  ·
  1

• Support dense vectors in sparse structured_dot

  #2455·ska2704 opened Oct 6, 2026·6 of 11 tasks·

  ·
  1

  8

• Fix large-parameter incomplete gamma accuracy in the C backend

  #2450·liuyinchu opened Oct 3, 2026·5 tasks done·

  ·
  1

  4

• Fix axis number in CAReduce zero-size error message

  #2438·anishmehta24 opened Sep 24, 2026Contributor·4 of 11 tasks·

  ·

• Let scalar * tensor fall back to the tensor's reflected operator

  #2435·anishmehta24 opened Sep 24, 2026Contributor·4 of 11 tasks·

  ·
  1

  6

• Add experimental JavaScript backend

  #2434·ricardoV94 opened Sep 24, 2026Member·

  ·
  3

• Preserve scalar input dtypes in Python Elemwise fallback

  #2431·Ai-Eastern opened Sep 23, 2026·4 tasks done·

  ·
  1

  13

• Add general C kernels for AdvancedSubtensor and AdvancedIncSubtensor

  #2410·velochy opened Sep 12, 2026Contributor·4 of 11 tasks·

  ·
  3