All pull requests

Pull requests

• [ET-VK][test] End-to-end tests of external graph I/O through the delegate

  CLA Signed
  #24912·pytorchbot opened Nov 11, 2026Collaborator·4/4·

  ·
  1

• [ET-VK][runtime] Zero-copy graph I/O through registered external resources

  CLA Signed
  #24911·pytorchbot opened Nov 11, 2026Collaborator·3/4·

  ·
  1

• [ET-VK][runtime] Make output staging visible to the host before readback

  CLA Signed
  #24910·pytorchbot opened Nov 11, 2026Collaborator·2/4·

  ·
  1

• [ET-VK][runtime] Remove the warmup_execute_after_compile option

  CLA Signed
  #24909·pytorchbot opened Nov 11, 2026Collaborator·1/4·

  ·
  2

• Add an aten.erf kernel

  CLA Signedmeta-exported
  #24908·psiddh opened Nov 11, 2026Contributor·

  ·
  4

• Run 1-D transposed convolution on the transposed 2-D kernels

  CLA Signedmeta-exported
  #24907·psiddh opened Nov 11, 2026Contributor·

  ·
  4

• Notify a real GitHub account from the PyTorch pin bump workflows

  CLA Signedrelease notes: nonetopic: not user facing
  #24905·shoumikhin opened Nov 11, 2026Contributor·

  ·
  2

• NXP backend: Add non quantized FP32 model reference

  CLA Signedmodule: nxprelease notes: nxp
  #24898·roman-janik-nxp opened Nov 11, 2026Collaborator·

  ·
  3

• Lower FP32 pow with exponents 2 and 0.5

  CLA Signedmeta-exported
  #24897·JakeStevens opened Nov 11, 2026Contributor·

  ·
  5

• Fix TorchAO source installs and automatic pin updates

  ciflow/dockerciflow/trunkCLA Signedhelp wantedpartner: armrelease notes: none
  #24896·zingo opened Nov 11, 2026Collaborator·

  ·
  1

• Run Windows Python unit tests in parallel

  ciflow/trunkCLA Signedhelp wantedmodule: windowspartner: armrelease notes: none
  #24895·zingo opened Nov 11, 2026Collaborator·

  ·
  2

• Arm backend: Fix avg_pool2d divisor_override with ceil_mode and no padding

  ciflow/trunkCLA Signedcommunity: contributionmodule: armpartner: arm
  #24894·EylonKrause opened Nov 11, 2026·

  ·
  5

• Arm backend: Fix ceil_mode pooling when the last window is dropped

  ciflow/trunkCLA Signedcommunity: contributionmodule: armpartner: arm
  #24893·EylonKrause opened Nov 11, 2026·

  ·
  5

• Add Arm Clang bare-metal smoke test to CI

  ciflow/trunkCLA Signedmodule: armpartner: armrelease notes: arm
  #24892·perheld opened Nov 11, 2026Collaborator·

  ·
  1

• [CPU] Respect output strides during broadcasting in CPU elementwise helpers

  CLA Signedcommunity: contributionrelease notes: ops & kernels
  #24891·AhmadHassan-BTed opened Nov 11, 2026·

  ·
  1

  3

• CUDA backend: run each delegate on its target device (#24887)

  CLA Signed
  #24890·Gasoonjia opened Nov 11, 2026Contributor·

  ·
  1

  22

• Respect CPU quotas when sizing quantized GEMM compiler workers

  CLA Signed
  #24889·Gasoonjia opened Nov 11, 2026Contributor·

  ·
  2

• Adapt sym quantizer to ET

  CLA Signedmeta-exported
  #24875·mgiordy opened Nov 11, 2026Contributor·

  ·
  5

• Bump node-forge from 1.3.1 to 1.4.3 in /examples/demo-apps/react-native/rnllama

  CLA Signeddependenciesjavascript
  #24873·dependabot[bot] opened Nov 10, 2026Bot·

  ·
  2

• Qualcomm AI Engine Direct - Allow ontarget compilation

  CLA Signed
  #24871·qti-mmadhava opened Nov 10, 2026Contributor·

  ·
  2

• Validate non-empty tensor data pointers during Method execution

  CLA Signedcommunity: contribution
  #24870·professor-moody opened Nov 10, 2026·

  ·
  2

• Stream safe text without fixed stop-marker holdback

  CLA Signed
  #24862·kiymetakdemir opened Nov 10, 2026Contributor·

  ·
  9

• Explain ExecuTorch in the ESP-IDF component's README

  CLA Signedmodule: microcontrollers
  #24861·rascani opened Nov 10, 2026Contributor·14/14·

  ·
  8

• [ET-VK][conv2d] Use row-aligned work groups for non-reducing pointwise convs on Mali

  CLA Signedmeta-exported
  #24859·SS-JIA opened Nov 10, 2026Contributor·

  ·
  2

• [ET-VK][conv2d] Fold the accumulation of Half pointwise convs on Adreno

  CLA Signedmeta-exported
  #24858·SS-JIA opened Nov 10, 2026Contributor·

  ·
  2