no-torch

Workflow runs for no-torch

• Add RVV vectorized FP32 kernel to CPU Adam

  requires action with the application:
  no-torch#1543:Pull request#8817synchronizebyzhangjian29
  zte-riscv:add-rvv-cpuamda-kernel

  Oct 21, 2026Action required
  More actions for workflow run 1543

• Make the Triton device guards accelerator-agnostic

  requires action with the application:
  no-torch#1542:Pull request#8868synchronizebyVenusTZZ
  VenusTZZ:npu-triton-support

  Oct 21, 2026Action required
  More actions for workflow run 1542

• no-torch

  completed successfully:
  no-torch#1541:Scheduled
  master

  Oct 21, 202635s
  More actions for workflow run 1541

• Make the Triton device guards accelerator-agnostic

  requires action with the application:
  no-torch#1540:Pull request#8868synchronizebyVenusTZZ
  VenusTZZ:npu-triton-support

  Oct 20, 2026Action required
  More actions for workflow run 1540

• Make the Triton device guards accelerator-agnostic

  requires action with the application:
  no-torch#1539:Pull request#8868openedbyVenusTZZ
  VenusTZZ:npu-triton-support

  Oct 20, 2026Action required
  More actions for workflow run 1539

• no-torch

  completed successfully:
  no-torch#1538:Scheduled
  master

  Oct 20, 202628s
  More actions for workflow run 1538

• Add KLXPU (KUNLUNXIN XPU) accelerator

  completed successfully:
  no-torch#1537:Pull request#8799synchronizebyGuCarpenter
  GuCarpenter:klxpu_dev

  Oct 19, 202634s
  More actions for workflow run 1537

• no-torch

  completed successfully:
  no-torch#1536:Scheduled
  master

  Oct 19, 202629s
  More actions for workflow run 1536

• Measure first-call and post-update costs of continuous rollout CUDA Graphs

  requires action with the application:
  no-torch#1535:Pull request#8853openedbynathon-lee
  nathon-lee:bench/rollout-cuda-graph-first-call

  Oct 18, 2026Action required
  More actions for workflow run 1535

• Support Transformers 5.20 cache query offsets in DeepSpeedStaticCache

  requires action with the application:
  no-torch#1534:Pull request#8852openedbynathon-lee
  nathon-lee:fix/static-cache-query-offset

  Oct 18, 2026Action required
  More actions for workflow run 1534

• no-torch

  completed successfully:
  no-torch#1533:Scheduled
  master

  Oct 18, 202630s
  More actions for workflow run 1533

• Enable CUDA Graph for continuous rollouts

  completed successfully:
  no-torch#1532:Pull request#8732synchronizebynathon-lee
  nathon-lee:feat/rollout-cuda-graph

  Oct 17, 202634s
  More actions for workflow run 1532

• no-torch

  completed successfully:
  no-torch#1531:Scheduled
  master

  Oct 17, 202633s
  More actions for workflow run 1531

• feat(rollout): segment-based kernel injection for hybrid-engine greedy decode

  completed successfully:
  no-torch#1530:Pull request#8602synchronizebydelock
  gma/qwen_ki_expr

  Oct 16, 202636s
  More actions for workflow run 1530

• feat(rollout): segment-based kernel injection for hybrid-engine greedy decode

  completed successfully:
  no-torch#1529:Pull request#8602synchronizebydelock
  gma/qwen_ki_expr

  Oct 16, 202633s
  More actions for workflow run 1529

• feat(rollout): segment-based kernel injection for hybrid-engine greedy decode

  completed successfully:
  no-torch#1528:Pull request#8602synchronizebydelock
  gma/qwen_ki_expr

  Oct 16, 202638s
  More actions for workflow run 1528

• feat(rollout): segment-based kernel injection for hybrid-engine greedy decode

  completed successfully:
  no-torch#1527:Pull request#8602synchronizebydelock
  gma/qwen_ki_expr

  Oct 16, 202640s
  More actions for workflow run 1527

• no-torch

  completed successfully:
  no-torch#1526:Scheduled
  master

  Oct 16, 202636s
  More actions for workflow run 1526

• Add opt-in TorchAO HiFloat8 Linear and AutoEP training

  requires action with the application:
  no-torch#1525:Pull request#8784synchronizebyhazelduan
  hazelduan:deepspeed_ao

  Oct 15, 2026Action required
  More actions for workflow run 1525

• Add opt-in TorchAO HiFloat8 Linear and AutoEP training

  requires action with the application:
  no-torch#1524:Pull request#8784synchronizebyhazelduan
  hazelduan:deepspeed_ao

  Oct 15, 2026Action required
  More actions for workflow run 1524

• feat(rollout): segment-based kernel injection for hybrid-engine greedy decode

  completed successfully:
  no-torch#1523:Pull request#8602synchronizebydelock
  gma/qwen_ki_expr

  Oct 15, 202639s
  More actions for workflow run 1523

• no-torch

  completed successfully:
  no-torch#1522:Scheduled
  master

  Oct 15, 202635s
  More actions for workflow run 1522

• feat(rollout): segment-based kernel injection for hybrid-engine greedy decode

  completed successfully:
  no-torch#1521:Pull request#8602synchronizebydelock
  gma/qwen_ki_expr

  Oct 14, 202638s
  More actions for workflow run 1521

• Add opt-in TorchAO HiFloat8 Linear and AutoEP training

  requires action with the application:
  no-torch#1520:Pull request#8784synchronizebyhazelduan
  hazelduan:deepspeed_ao

  Oct 14, 2026Action required
  More actions for workflow run 1520

• Add opt-in TorchAO HiFloat8 Linear and AutoEP training

  requires action with the application:
  no-torch#1519:Pull request#8784synchronizebyhazelduan
  hazelduan:deepspeed_ao

  Oct 14, 2026Action required
  More actions for workflow run 1519