All pull requests

Pull requests

• [TorchToTosa] Support avg_pool2d divisor_override

  #4824·gokcan-karakus opened Oct 12, 2026Collaborator·

  ·

• [TorchToTosa] Support max_pool2d with indices

  #4823·gokcan-karakus opened Oct 12, 2026Collaborator·

  ·

• [TorchToLinalg] Use linalg.index to lower aten.reflection_pad1d

  #4822·davidmcevadsp opened Oct 12, 2026·

  ·

• [TorchToLinalg] Fix adaptive pooling non-finite option

  #4821·Lallapallooza opened Oct 12, 2026Member·

  ·
  1

• [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·

  ·
  1

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

  #4816·prempk229 opened Oct 8, 2026·

  ·

• [Torch] Decompose aten.select_copy.int

  #4814·Kamalrajkannan opened Oct 8, 2026·

  ·
  2

• [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] 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·

  ·
  12

• [LLVM] Integrate LLVM at 2fc577681ca1c896597252941593da9b61e0960b

  #4802·jackv1ck opened Oct 2, 2026Contributor·

  ·
  4

• 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·

  ·
  5

• [Torch] RestructureNonConstantAxes: eliminate orphan constant.int from single-element select.int chain

  #4792·liuyalong opened Sep 29, 2026·

  ·
  4

• [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·

  ·
  6

• [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

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

  #4781·sgjzfzzf opened Sep 23, 2026Contributor·

  ·
  2

• [TorchToLinalg] Fix legacy aten.view narrowing a dynamic dim to static 1

  #4776·felixtensor opened Sep 21, 2026Contributor·

  ·
  1

  2