Commits

Branch selector

 main

User selector

All users

All time

Commit History

Commits on Jun 11, 2026

• Walk transparent ops when extracting input quant params (#20139)

  Show description for 27cc7e8
  ethansfng

  authored
  27cc7e8

  Copy full SHA for 27cc7e8

• Fix spotlessKotlinCheck formatting for makeExecutorchException call (#20211)

  Show description for bd7426d
  psiddh

  authored
  bd7426d

  Copy full SHA for bd7426d

• Update fused quant broadcast logic (#20171)

  Show description for 86e861c
  DrJessop

  authored
  86e861c

  Copy full SHA for 86e861c

• Preserve GraphModule.meta in ExportPass

  Show description for 39dade2
  abeakkas

  authored
  39dade2

  Copy full SHA for 39dade2

Commits on Jun 10, 2026

• Qualcomm AI Engine Direct - Adding QNN backend support for _cdist_forward core ATen op (#20195)

  Show description for 129c687
  qti-horodnic

  authored
  129c687

  Copy full SHA for 129c687

• [ExecuTorch][WebGPU] SymInt live-scalar mechanism + et_vk.select_as_symint

  Show description for 5526971
  JulianCloudNTH

  committed
  5526971

  Copy full SHA for 5526971

• [ExecuTorch][WebGPU] Add update_cache tests (native numeric + export)

  Show description for 14168a3
  JulianCloudNTH

  committed
  14168a3

  Copy full SHA for 14168a3

• [ExecuTorch][WebGPU] Add update_cache op (llama.update_cache)

  Show description for 1208a2f
  JulianCloudNTH

  committed
  1208a2f

  Copy full SHA for 1208a2f

• [ET-VK][conv2d] Auto-route SlidingWindow conv2d to im2col/GEMM via device-aware heuristic (#20190)

  Show description for caeb005
  

  2 peopleauthored and

  SS-JIA

  committed
  caeb005

  Copy full SHA for caeb005

• [ETVK] Add benchmark binary + im2col/GEMM conv2d prototype

  Show description for fd2cf88
  

  ssjia

  authored and

  SS-JIA

  committed
  fd2cf88

  Copy full SHA for fd2cf88

• [ExecuTorch][WebGPU] Add per-pass dispatch ordering + scratch buffer tests

  Show description for f0dff03
  JulianCloudNTH

  committed
  f0dff03

  Copy full SHA for f0dff03

• [ExecuTorch][WebGPU] Switch native backend from wgpu-native to Dawn (Tint) + SwiftShader

  Show description for f8bf776
  JulianCloudNTH

  committed
  f8bf776

  Copy full SHA for f8bf776

• Arm backend: add SmolLM2 VGF generation and evaluation workflow (#19971)

  Show description for c26c649
  

  xingguo01

  and

  Copilot

  authored
  c26c649

  Copy full SHA for c26c649

• Add benchmarking script (#20188) (#20188)

  Show description for 4ed31d3
  metascroy

  authored
  4ed31d3

  Copy full SHA for 4ed31d3

• Arm backend: Replace VkQueueWaitIdle with fences (#20186)

  Show description for e0be283
  wwwind

  authored
  e0be283

  Copy full SHA for e0be283

• [cuda] int4: stabilize two-layer decode test via CUDA-seeded init (#20196)

  Show description for ae4b0a4
  Gasoonjia

  authored
  ae4b0a4

  Copy full SHA for ae4b0a4

• Route Module.loadMethod through makeExecutorchException for native log enrichment (#20191)

  Show description for 4519036
  psiddh

  authored
  4519036

  Copy full SHA for 4519036

• Setuptools symlinks (#20092)

  Show description for 97c153f
  tomeuv

  authored
  97c153f

  Copy full SHA for 97c153f

• Switch to neon for interleave (#20137)

  Show description for 92e6a4c
  metascroy

  authored
  92e6a4c

  Copy full SHA for 92e6a4c

• Fix Windows CUDA build: guard extension_cuda compile options to CXX (#20184)

  Show description for eb851a5
  shoumikhin

  authored
  eb851a5

  Copy full SHA for eb851a5

• Cortex-M backend: add cortex-m0plus to the trunk e2e CI matrix (#19774)

  Show description for a1649b9
  rascani

  authored
  a1649b9

  Copy full SHA for a1649b9

• Extend CPPCHECK scope to prim ops

  Show description for cec0814
  perheld

  committed
  cec0814

  Copy full SHA for cec0814

• Cap PyTorch build parallelism for all GCC docker images (#20123)

  Show description for ceda793
  

  rascani

  and

  claude

  authored
  ceda793

  Copy full SHA for ceda793

• Arm backend: Bump tosa-tools version for Ethos-U as well

  Show description for 4229704
  per

  committed
  4229704

  Copy full SHA for 4229704

• Arm backend: Bump to 2026.05 release of tosa-tools

  Show description for 8e10687
  per

  committed
  8e10687

  Copy full SHA for 8e10687

• Cortex-M backend: add cortex-m7 to the trunk e2e CI matrix (#19730)

  Show description for 6f2331b
  rascani

  authored
  6f2331b

  Copy full SHA for 6f2331b

• Qualcomm AI Engine Direct - Verify Direct Build in External CI (#19763)

  Show description for 10bc51e
  winskuo-quic

  authored
  10bc51e

  Copy full SHA for 10bc51e

• Qualcomm AI Engine Direct - Decouple calibration and evaluation task flags (#20113)

  Show description for 286acf5
  DannyYuyang-quic

  authored
  286acf5

  Copy full SHA for 286acf5

• Arm backend: Add modified guards before retracing (#20180)

  Show description for 1434397
  

  YufengShi-dudu

  and

  martinlsm

  authored
  1434397

  Copy full SHA for 1434397

• Enable Arm VGF delegate in pybind builds (#19290)

  Show description for c7a0b68
  robell

  authored
  c7a0b68

  Copy full SHA for c7a0b68

• NXP backend: New Neutron-C flow support for ReLU (#19275)

  Show description for f07ddec
  StrycekSimon

  authored
  f07ddec

  Copy full SHA for f07ddec

• Arm backend: Add smoke test for VGF (#20175)

  Show description for 03d1818
  wwwind

  authored
  03d1818

  Copy full SHA for 03d1818

• Arm backend: Add Ethos-U65 path to testing (#20141)

  Show description for 2b9e9bf
  Erik-Lundell

  authored
  2b9e9bf

  Copy full SHA for 2b9e9bf

• Make the CUDA caller-stream guard a shared extension/cuda library (#20158) (#20158)

  Show description for 0b13b6a
  shoumikhin

  authored
  0b13b6a

  Copy full SHA for 0b13b6a

• NXP backend: Enable `cat` with new Neutron flow. (#20106)

  Show description for ce2f3f9
  MartinPavella

  authored
  ce2f3f9

  Copy full SHA for ce2f3f9