All pull requests

Pull requests

• [TMTensor] Bufferize ins operands at the op that reads them

  #4820·nmalimban opened Oct 9, 2026Contributor·

  ·
  1

• Add Decomposition for unique consecutive op

  #4819·meenakshiramanathan1068 opened Oct 9, 2026·

  ·

• [TorchToTosa] Add avg_pool3d lowering

  #4818·aakhilanvarsadat opened Oct 8, 2026·

  ·

• [Torch] Decompose aten.kthvalue op

  #4817·Kamalrajkannan opened Oct 8, 2026·

  ·

• [Torch] Add decomposition and lowering support for aten.index_put_

  #4816·prempk229 opened Oct 8, 2026·

  ·

• [RefBackend] Pass contiguous inputs with identity-layout strides

  #4815·felixtensor opened Oct 8, 2026·

  ·
  1

• [Torch] Decompose aten.select_copy.int

  #4814·Kamalrajkannan opened Oct 8, 2026·

  ·
  1

• [Torch] Add decomposition pattern for aten.masked_select

  #4813·Jayshree-Jha opened Oct 7, 2026·

  ·

• [TorchToLinalg] Correct nearest GridSample with zero padding

  #4810·Prajit-Rahul opened Oct 6, 2026·

  ·
  1

• [Torch] Decompose scalar max.dim

  #4809·gokcan-karakus opened Oct 5, 2026Collaborator·

  ·

• [Torch] Decompose scalar gather through rank-one tensors

  #4808·gokcan-karakus opened Oct 5, 2026Collaborator·

  ·

• [TorchToTosa] Lower static 1D replication padding

  #4806·gokcan-karakus opened Oct 5, 2026Collaborator·

  ·

• [Torch] Add decomposition for aten.scatter.reduce

  #4804·shashank-mcw opened Oct 4, 2026·

  ·

• [TorchToLinalg] Add first-class PT2E per-tensor.tensor/2 quantize/dequantize/choose_qparams ops

  #4803·guikorol opened Oct 2, 2026Contributor·

  ·
  11

• Integrate LLVM at 39b72080e0399a15b0313b4eb24e3dbb0aee3c08

  #4802·jackv1ck opened Oct 2, 2026Contributor·

  ·
  2

• Add support for upsample_nearest1d

  #4795·prempk229 opened Oct 1, 2026·

  ·

• Add decomposition for aten.meshgrid.indexing

  #4794·meenakshiramanathan1068 opened Oct 1, 2026·

  ·

• Add decomposition for aten.multinomial

  #4793·Kamalrajkannan opened Sep 30, 2026·

  ·
  4

• [TorchOnnxToTorch] Extract single-element runtime reduce axes via aten.item

  #4792·liuyalong opened Sep 29, 2026·

  ·
  3

• [TorchToLinalg] Handle quantized inputs in aten.bmm lowering

  #4791·Prajit-Rahul opened Sep 28, 2026·

  ·
  2

• [TorchToLinalg] Avoid over-padding pooling inputs when ceil_mode=true

  #4790·sahas3 opened Sep 28, 2026Member·

  ·
  5

• [TorchToTosa] Fix ThresholdBackward dynamic-shape crash

  #4788·alexxony opened Sep 25, 2026Contributor·

  ·

• [Torch] Add decomposition for aten.sinc

  #4786·shashank-mcw opened Sep 24, 2026·

  ·
  4

• [TorchToTosa] Preserve int64 lowering

  #4783·davsva01 opened Sep 23, 2026Contributor·

  ·
  11

• [Python] Make the torch-mlir package prefix configurable

  #4781·sgjzfzzf opened Sep 23, 2026Contributor·

  ·
  2