Build Linux CUDA Wheels

Workflow runs for Build Linux CUDA Wheels

• Add Arm Buck2 coverage and declaration checks to GitHub CI

  completed successfully:
  Build Linux CUDA Wheels#4266:Pull request#23888synchronizebyusamahz
  usamahz:codex/arm-buck2-ci

  Oct 19, 202635m 0s
  More actions for workflow run 4266

• Add Arm Buck2 coverage and declaration checks to GitHub CI

  cancelled:
  Build Linux CUDA Wheels#4265:Pull request#23888synchronizebyusamahz
  usamahz:codex/arm-buck2-ci

  Oct 19, 202610m 15s
  More actions for workflow run 4265

• [PyBindings] Fix wrong results for inputs in another layout, refuse stale or mixed state, return CUDA outputs without copies

  completed successfully:
  Build Linux CUDA Wheels#4264:Pull request#23957synchronizebyshoumikhin
  shoumikhin:pybindings-torchfree-gaps

  Oct 19, 202633m 51s
  More actions for workflow run 4264

• Arm backend: Support configurable RIFE model classes

  completed successfully:
  Build Linux CUDA Wheels#4263:Pull request#23994openedbyoscarandersson8218
  oscarandersson8218:two-block-rife-example

  Oct 19, 202632m 58s
  More actions for workflow run 4263

• 2026-10-19 nightly release (adc713a2e66a1f65eb8ea931d4fd1b9e6ef161c6)

  completed successfully:
  Build Linux CUDA Wheels#4262:Commitb39aa75pushedbypytorchbot
  nightly

  Oct 19, 202639m 23s
  More actions for workflow run 4262

• Add Arm Buck2 coverage and declaration checks to GitHub CI

  completed successfully:
  Build Linux CUDA Wheels#4261:Pull request#23888synchronizebyusamahz
  usamahz:codex/arm-buck2-ci

  Oct 19, 202635m 12s
  More actions for workflow run 4261

• Add Arm Buck2 coverage and declaration checks to GitHub CI

  cancelled:
  Build Linux CUDA Wheels#4260:Pull request#23888synchronizebyusamahz
  usamahz:codex/arm-buck2-ci

  Oct 19, 202615m 15s
  More actions for workflow run 4260

• Bump PyTorch and TorchAO pins for dev20261018

  completed successfully:
  Build Linux CUDA Wheels#4259:Pull request#23988openedbypytorchupdatebot
  automated/pytorch-pin-bump-dev20261018

  Oct 19, 202632m 17s
  More actions for workflow run 4259

• [PyBindings] Fix wrong results for inputs in another layout, refuse stale or mixed state, return CUDA outputs without copies

  completed successfully:
  Build Linux CUDA Wheels#4258:Pull request#23957synchronizebyshoumikhin
  shoumikhin:pybindings-torchfree-gaps

  Oct 19, 202647m 16s
  More actions for workflow run 4258

• [executorch][cuda] Expose raw packed steps on CudaExecutor

  completed successfully:
  Build Linux CUDA Wheels#4257:Pull request#23979synchronizebyGasoonjia
  gh/Gasoonjia/78/head

  Oct 19, 202635m 33s
  More actions for workflow run 4257

• [executorch][muse] Hash prepared image identity without CommonCrypto

  completed successfully:
  Build Linux CUDA Wheels#4256:Pull request#23730synchronizebyGasoonjia
  gh/Gasoonjia/46/head

  Oct 19, 202635m 23s
  More actions for workflow run 4256

• [executorch][muse] Run the CUDA DFlash drafter in CUDA graphs

  completed successfully:
  Build Linux CUDA Wheels#4255:Pull request#23985synchronizebyGasoonjia
  gh/Gasoonjia/84/head

  Oct 19, 202634m 58s
  More actions for workflow run 4255

• [executorch][cuda] Test prefix-cache reuse on CudaExecutor

  completed successfully:
  Build Linux CUDA Wheels#4254:Pull request#23732synchronizebyGasoonjia
  gh/Gasoonjia/48/head

  Oct 19, 202635m 49s
  More actions for workflow run 4254

• [executorch][cuda] Test CudaExecutor on every step width up to 65

  completed successfully:
  Build Linux CUDA Wheels#4253:Pull request#23819synchronizebyGasoonjia
  gh/Gasoonjia/59/head

  Oct 19, 202636m 58s
  More actions for workflow run 4253

• [executorch][muse] Copy only committed DFlash hidden rows to host

  completed successfully:
  Build Linux CUDA Wheels#4252:Pull request#23982synchronizebyGasoonjia
  gh/Gasoonjia/81/head

  Oct 19, 202635m 46s
  More actions for workflow run 4252

• [executorch][muse] Export the batching contract for CUDA with --use-offgraph-kv-cache

  completed successfully:
  Build Linux CUDA Wheels#4251:Pull request#23729synchronizebyGasoonjia
  gh/Gasoonjia/45/head

  Oct 19, 202643m 4s
  More actions for workflow run 4251

• [executorch][muse] Add cuda dflash executor

  completed successfully:
  Build Linux CUDA Wheels#4250:Pull request#23981synchronizebyGasoonjia
  gh/Gasoonjia/80/head

  Oct 19, 202636m 24s
  More actions for workflow run 4250

• [executorch][muse] Verify CUDA DFlash drafts on device under any sampling

  completed successfully:
  Build Linux CUDA Wheels#4249:Pull request#23984synchronizebyGasoonjia
  gh/Gasoonjia/83/head

  Oct 19, 202639m 3s
  More actions for workflow run 4249

• [executorch][muse] Run the MG batching runtime on CudaExecutor

  completed successfully:
  Build Linux CUDA Wheels#4248:Pull request#23731synchronizebyGasoonjia
  gh/Gasoonjia/47/head

  Oct 19, 202638m 57s
  More actions for workflow run 4248

• [executorch][muse] Export packed DFlash drafts for CUDA batching

  completed successfully:
  Build Linux CUDA Wheels#4247:Pull request#23980synchronizebyGasoonjia
  gh/Gasoonjia/79/head

  Oct 19, 202638m 37s
  More actions for workflow run 4247

• [executorch][cuda] Run packed noncausal DFlash draft steps on the cell cache

  completed successfully:
  Build Linux CUDA Wheels#4246:Pull request#23978synchronizebyGasoonjia
  gh/Gasoonjia/77/head

  Oct 19, 202633m 13s
  More actions for workflow run 4246

• [executorch][muse] Serve MG batching on CUDA from the launcher

  completed successfully:
  Build Linux CUDA Wheels#4245:Pull request#23733synchronizebyGasoonjia
  gh/Gasoonjia/49/head

  Oct 19, 202634m 34s
  More actions for workflow run 4245

• [executorch][ci] Test Muse Glimmer batching on CUDA in OSS CI

  completed successfully:
  Build Linux CUDA Wheels#4244:Pull request#23818synchronizebyGasoonjia
  gh/Gasoonjia/58/head

  Oct 19, 202639m 39s
  More actions for workflow run 4244

• [executorch][muse] Draft CUDA DFlash decode from pending tokens

  completed successfully:
  Build Linux CUDA Wheels#4243:Pull request#23983synchronizebyGasoonjia
  gh/Gasoonjia/82/head

  Oct 19, 202636m 59s
  More actions for workflow run 4243

• Arm backend: Guard ASan address ranges against integer overflow

  completed successfully:
  Build Linux CUDA Wheels#4242:Pull request#23986openedbySebastian-Larsson
  Sebastian-Larsson:Arm_backend_Guard_ASan_address_ranges_against_integer_overflow

  Oct 19, 202632m 47s
  More actions for workflow run 4242