GPU Unit Tests

Workflow runs for GPU Unit Tests

• Make TorchTitan local compile configuration extensible

  skipped:
  GPU Unit Tests#27:Pull request#13synchronizebyryan-li-ryanllli-aws
  ryanllli/model-local-compile-backend

  Oct 19, 20261s
  More actions for workflow run 27

• Make TorchTitan local compile configuration extensible

  skipped:
  GPU Unit Tests#26:Pull request#13synchronizebyryan-li-ryanllli-aws
  ryanllli/model-local-compile-backend

  Oct 19, 20261s
  More actions for workflow run 26

• Make TorchTitan local compile configuration extensible

  skipped:
  GPU Unit Tests#25:Pull request#13synchronizebyryan-li-ryanllli-aws
  ryanllli/model-local-compile-backend

  Oct 19, 202610s
  More actions for workflow run 25

• Make TorchTitan local compile configuration extensible

  skipped:
  GPU Unit Tests#24:Pull request#13synchronizebyryan-li-ryanllli-aws
  ryanllli/model-local-compile-backend

  Oct 19, 202610s
  More actions for workflow run 24

• Make TorchTitan local compile configuration extensible

  skipped:
  GPU Unit Tests#23:Pull request#13synchronizebyryan-li-ryanllli-aws
  ryanllli/model-local-compile-backend

  Oct 19, 20268s
  More actions for workflow run 23

• Make TorchTitan local compile configuration extensible

  skipped:
  GPU Unit Tests#22:Pull request#13synchronizebyryan-li-ryanllli-aws
  ryanllli/model-local-compile-backend

  Oct 16, 20266s
  More actions for workflow run 22

• Make TorchTitan local compile configuration extensible

  skipped:
  GPU Unit Tests#21:Pull request#13synchronizebyryan-li-ryanllli-aws
  ryanllli/model-local-compile-backend

  Oct 15, 202610s
  More actions for workflow run 21

• Make TorchTitan local compile configuration extensible

  skipped:
  GPU Unit Tests#20:Pull request#13synchronizebyryan-li-ryanllli-aws
  ryanllli/model-local-compile-backend

  Oct 15, 20261s
  More actions for workflow run 20

• Make TorchTitan local compile configuration extensible

  skipped:
  GPU Unit Tests#19:Pull request#13synchronizebyryan-li-ryanllli-aws
  ryanllli/model-local-compile-backend

  Oct 15, 20262s
  More actions for workflow run 19

• Make TorchTitan local compile configuration extensible

  skipped:
  GPU Unit Tests#18:Pull request#13openedbyryan-li-ryanllli-aws
  ryanllli/model-local-compile-backend

  Oct 15, 20261s
  More actions for workflow run 18

• Allow configuring the default torch.compile backend

  skipped:
  GPU Unit Tests#17:Pull request#12synchronizebyryan-li-ryanllli-aws
  ryanllli/custom-device-plugin-example

  Oct 15, 20262s
  More actions for workflow run 17

• Allow configuring the default torch.compile backend

  skipped:
  GPU Unit Tests#16:Pull request#12openedbyryan-li-ryanllli-aws
  ryanllli/custom-device-plugin-example

  Oct 15, 20262s
  More actions for workflow run 16

• Make local compile backend configurable

  skipped:
  GPU Unit Tests#15:Pull request#11openedbyryan-li-ryanllli-aws
  ryanllli/configurable-local-compile-backend

  Oct 15, 20268s
  More actions for workflow run 15

• Add optimizer-owned gradient clipping extension point

  skipped:
  GPU Unit Tests#14:Pull request#10synchronizebyryan-li-ryanllli-aws
  ryanllli/compiled-grad-clip-extension

  Oct 14, 20262s
  More actions for workflow run 14

• Add optimizer-owned gradient clipping extension point

  skipped:
  GPU Unit Tests#13:Pull request#10synchronizebyryan-li-ryanllli-aws
  ryanllli/compiled-grad-clip-extension

  Oct 14, 20261s
  More actions for workflow run 13

• Add optimizer-owned gradient clipping extension point

  skipped:
  GPU Unit Tests#12:Pull request#10synchronizebyryan-li-ryanllli-aws
  ryanllli/compiled-grad-clip-extension

  Oct 14, 20262s
  More actions for workflow run 12

• Add optimizer-owned gradient clipping extension point

  skipped:
  GPU Unit Tests#11:Pull request#10openedbyryan-li-ryanllli-aws
  ryanllli/compiled-grad-clip-extension

  Oct 14, 20262s
  More actions for workflow run 11

• [TorchTitan] Tell to_hf and from_hf when an HF load is quantized (#5063)

  cancelled:
  GPU Unit Tests#10:Commit3f021a1pushedbyryan-li-ryanllli-aws
  main

  Oct 14, 20261d 0h 15m 34s
  More actions for workflow run 10

• [transformers_modeling_backend] Keep the Titan MoE output for HF's ba…

  cancelled:
  GPU Unit Tests#9:Commitee3b997pushedbyryan-li-ryanllli-aws
  main

  Oct 13, 20261d 0h 21m 0s
  More actions for workflow run 9

• feat: make SAC matmul recompute frequency configurable

  skipped:
  GPU Unit Tests#8:Pull request#8synchronizebysamansg-aws
  samansg/sac-matmul-recompute-freq

  Oct 7, 20262s
  More actions for workflow run 8

• [moe] Add GlobalBatchLoadBalanceLoss: DP-global MoE load-balance loss (Qwen3)

  skipped:
  GPU Unit Tests#7:Pull request#9openedbyaws-agi-boreival
  moe/global-batch-lbl

  Oct 7, 20263s
  More actions for workflow run 7

• feat: make SAC matmul recompute frequency configurable

  skipped:
  GPU Unit Tests#6:Pull request#8openedbysamansg-aws
  samansg/sac-matmul-recompute-freq

  Oct 5, 20261s
  More actions for workflow run 6

• [observability] Profile Neuron devices with the PrivateUse1 activity

  skipped:
  GPU Unit Tests#5:Pull request#6openedbykvshbg-aws
  kvshbg/privateuse1-profiler

  Oct 5, 20261s
  More actions for workflow run 5

• [attention] Make the FlexAttention mask type configurable

  cancelled:
  GPU Unit Tests#4:Pull request#5synchronizebyaws-arashj
  attention/flex-mask-type

  Oct 5, 20261d 0h 0m 11s
  More actions for workflow run 4

• [attention] Make the FlexAttention mask type configurable

  cancelled:
  GPU Unit Tests#3:Pull request#5openedbyaws-arashj
  attention/flex-mask-type

  Oct 5, 202610m 37s
  More actions for workflow run 3