GPU tests

Workflow runs for GPU tests

• Record the plan hf_ptq's FSDP2 load distributes the model by

  failed:
  GPU tests#10781:Commit7126fa6pushedbycopy-pr-bot[bot]Bot
  pull-request/2740

  Oct 23, 202634s
  More actions for workflow run 10781

• Refuse experts with a quant wrapper of their own instead of unregiste…

  failed:
  GPU tests#10780:Commitb0a357bpushedbycopy-pr-bot[bot]Bot
  pull-request/2741

  Oct 23, 202637s
  More actions for workflow run 10780

• Test that the formats the expert-parallel load dequantizes pass its c…

  failed:
  GPU tests#10779:Commit783b133pushedbycopy-pr-bot[bot]Bot
  pull-request/2743

  Oct 23, 202637s
  More actions for workflow run 10779

• Test the keep-fused fold's per-stage amax fallback with a W4A8 expert

  failed:
  GPU tests#10778:Commit0dec764pushedbycopy-pr-bot[bot]Bot
  pull-request/2750

  Oct 23, 202635s
  More actions for workflow run 10778

• Refuse to import an untrusted checkpoint's code in the Kimi adapters

  failed:
  GPU tests#10777:Commit8a42c07pushedbycopy-pr-bot[bot]Bot
  pull-request/2748

  Oct 23, 202635s
  More actions for workflow run 10777

• Import the dense-sharding test's transformers modules at the top

  failed:
  GPU tests#10776:Commit0d976e2pushedbycopy-pr-bot[bot]Bot
  pull-request/2742

  Oct 23, 202636s
  More actions for workflow run 10776

• Date distributed_writer 2026, as LICENSE_HEADER does

  failed:
  GPU tests#10775:Commit67caebepushedbycopy-pr-bot[bot]Bot
  pull-request/2751

  Oct 23, 202637s
  More actions for workflow run 10775

• Decompress Kimi-K3 only with a quantization_config, and let its error…

  failed:
  GPU tests#10774:Commite1e929fpushedbycopy-pr-bot[bot]Bot
  pull-request/2749

  Oct 23, 202633s
  More actions for workflow run 10774

• Run hf_ptq's sharded previews for real, under expert parallelism and …

  failed:
  GPU tests#10773:Commit68d6bb3pushedbycopy-pr-bot[bot]Bot
  pull-request/2754

  Oct 23, 202636s
  More actions for workflow run 10773

• Name the way around expert parallelism when a model is below its tran…

  failed:
  GPU tests#10772:Commit55cbc94pushedbycopy-pr-bot[bot]Bot
  pull-request/2746

  Oct 23, 202631s
  More actions for workflow run 10772

• Route an expert-parallel plan to the loader in the same module

  failed:
  GPU tests#10771:Commit636d578pushedbycopy-pr-bot[bot]Bot
  pull-request/2780

  Oct 23, 202636s
  More actions for workflow run 10771

• Tell a compressed checkpoint's config that fails to parse from none

  failed:
  GPU tests#10770:Commitb2cc481pushedbycopy-pr-bot[bot]Bot
  pull-request/2747

  Oct 23, 202638s
  More actions for workflow run 10770

• Say what picks the no-gather writer in the distributed export test

  failed:
  GPU tests#10769:Commita04fa68pushedbycopy-pr-bot[bot]Bot
  pull-request/2752

  Oct 23, 202636s
  More actions for workflow run 10769

• Import the expert-parallel export tests' ModelOpt modules at module s…

  failed:
  GPU tests#10768:Commit0bb80b2pushedbycopy-pr-bot[bot]Bot
  pull-request/2753

  Oct 23, 202637s
  More actions for workflow run 10768

• Keep assistant-less rows for loss_mode full; raise chat-tokenization …

  completed successfully:
  GPU tests#10767:Commit17964edpushedbycopy-pr-bot[bot]Bot
  pull-request/2912

  Oct 23, 202628m 20s
  More actions for workflow run 10767

• Add regression coverage for subfolder state restoration and cache lookup

  completed successfully:
  GPU tests#10766:Commit86ec87fpushedbycopy-pr-bot[bot]Bot
  pull-request/2971

  Oct 23, 202627m 41s
  More actions for workflow run 10766

• Date distributed_writer 2026, as LICENSE_HEADER does

  failed:
  GPU tests#10765:Commit7095ff2pushedbycopy-pr-bot[bot]Bot
  pull-request/2751

  Oct 23, 20263m 17s
  More actions for workflow run 10765

• Import the expert-parallel export tests' ModelOpt modules at module s…

  failed:
  GPU tests#10764:Commit951faa3pushedbycopy-pr-bot[bot]Bot
  pull-request/2753

  Oct 23, 20263m 18s
  More actions for workflow run 10764

• Run hf_ptq's sharded previews for real, under expert parallelism and …

  failed:
  GPU tests#10763:Commit33998a6pushedbycopy-pr-bot[bot]Bot
  pull-request/2754

  Oct 23, 20262m 22s
  More actions for workflow run 10763

• Refuse experts with a quant wrapper of their own instead of unregiste…

  failed:
  GPU tests#10762:Commitdb14b8fpushedbycopy-pr-bot[bot]Bot
  pull-request/2741

  Oct 23, 20261m 31s
  More actions for workflow run 10762

• Decompress Kimi-K3 only with a quantization_config, and let its error…

  failed:
  GPU tests#10761:Commit94f6f78pushedbycopy-pr-bot[bot]Bot
  pull-request/2749

  Oct 23, 20262m 46s
  More actions for workflow run 10761

• Tell a compressed checkpoint's config that fails to parse from none

  failed:
  GPU tests#10760:Commitcd230a7pushedbycopy-pr-bot[bot]Bot
  pull-request/2747

  Oct 23, 20263m 12s
  More actions for workflow run 10760

• Refuse rank-local calibration that another algorithm runs for its scales

  failed:
  GPU tests#10759:Commit60b267dpushedbycopy-pr-bot[bot]Bot
  pull-request/2740

  Oct 23, 20263m 3s
  More actions for workflow run 10759

• Say what picks the no-gather writer in the distributed export test

  failed:
  GPU tests#10757:Commit8d8a821pushedbycopy-pr-bot[bot]Bot
  pull-request/2752

  Oct 23, 20262m 48s
  More actions for workflow run 10757

• Test the keep-fused fold's per-stage amax fallback with a W4A8 expert

  failed:
  GPU tests#10758:Commit07be9dcpushedbycopy-pr-bot[bot]Bot
  pull-request/2750

  Oct 23, 20263m 3s
  More actions for workflow run 10758