All pull requests

Pull requests

• chore: update cubek/cubecl

  #6089·Charles23R opened Oct 15, 2026Contributor·

  ·
  1

• fix(fusion): hold an operation's read-only inputs until it is submitted

  bugfusion
  #6086·ThierryCantin-Demers opened Oct 15, 2026Member·3 tasks·

  ·
  2

• fix(fusion): tune index-reading fused blocks on zeroed indices under CompileAndAutotune

  #6076·SamuelBelanger opened Oct 15, 2026Contributor·

  ·
  4

• perf(flex): run 2D conv im2col directly on NCHW input

  #6075·Gilfeather opened Oct 15, 2026·2 of 3 tasks·

  ·
  3

• test: remove redundant tests and fix tests that did not exercise their names

  #6066·antimora opened Oct 13, 2026Collaborator·1 of 3 tasks·

  ·
  7

  

• fix(cubecl): preserve maxpool3d index strides and validate vector reductions

  #6057·HongchengZhao opened Oct 13, 2026·2 of 3 tasks·

  ·

• Fix: TryTraceFuser::fuse() not checking if bindings out of bounds for first operation added

  #6053·abroy77 opened Oct 12, 2026·3 tasks done·

  ·
  1

  4

• perf(cubecl): F32 unit convolution, im2col gather, and split depthwise wgrad in conv autotune

  #6048·znicelya opened Oct 12, 2026·3 tasks·

  ·

• Add regression metrics (MSE, MAE, R²)

  #6042·Infinty-ux opened Oct 11, 2026Contributor·1 of 3 tasks·

  ·
  1

  5

• Perf/pytorch transposed load

  store
  #6041·skewballfox opened Oct 11, 2026Contributor·3 tasks done·

  ·
  22

• feat(group): split tensors over a group of devices

  enhancement
  #6025·ThierryCantin-Demers opened Oct 9, 2026Member·1 of 3 tasks·

  ·
  1

• fix: validate quantized tensor data lengths

  #5978·Prajwal-k-tech opened Oct 5, 2026·

  ·
  1

  6

• Add 1D and 2D Lp pooling modules

  #5977·Prajwal-k-tech opened Oct 5, 2026·

  ·
  1

  11

• feat(tensor): support grouped-query attention and causal mask alignment

  enhancement
  #5953·Andy2887 opened Oct 2, 2026Contributor·

  ·
  1

  5

• feat(config): add configurable NaN policy for CubeCL backends

  enhancement
  #5952·laggui opened Oct 2, 2026Member·2 tasks·

  ·

• feat(data): add drop_last to the data loader

  datasetfeature
  #5911·ProgrammerProfi opened Sep 30, 2026·2 tasks done·

  ·
  1

  2

• feat(optim)!: accept a device learning rate for graph-captured steps

  bug
  #5872·nathanielsimard opened Sep 28, 2026Member·

  ·
  3

• feat(linalg): add Cholesky decomposition

  feature
  #5858·Mikyx-1 opened Sep 26, 2026Contributor·1 of 2 tasks·

  ·
  1

  5

• feat(linalg): add batched linear solve

  feature
  #5856·Mikyx-1 opened Sep 26, 2026Contributor·2 of 3 tasks·

  ·
  1

  3

• feat(tensor): add differentiable trilinear grid_sample_3d

  feature
  #5803·thewtex opened Sep 23, 2026·3 tasks done·

  ·
  12

• feat(core): pipeline trait to split a model by layers across devices

  feature
  #5702·ThierryCantin-Demers opened Sep 16, 2026Member·

  ·
  10

• chore(deps): validate CubeCL complex cast lowering

  dependencies
  #5541·shinaoka opened Aug 31, 2026·1 of 2 tasks·

  ·
  2

• feat(tensor): add GroupNorm backend primitive and CubeCL kernel

  feature
  #5504·bungogood opened Aug 27, 2026·

  ·
  6

• chore(bench): improve benchmark defaults and the comparison script

  benchmark
  #5367·laundmo opened Aug 14, 2026Contributor·2 tasks·

  ·
  5

• feat(cubecl): native argwhere/nonzero/mask_select via stream compaction

  feature
  #5309·Yashiru opened Aug 6, 2026Contributor·2 tasks done·

  ·
  5