All workflows

Workflow runs

• feat(dsv4): unified BI joint sparse attention with sink and Graph records (item 24)

  completed successfully:
  Open Code Review#295:Pull request#628openedbyinaniloquentee
  Oct 28, 20261m 18s
  More actions for workflow run 295

• feat(dsv4): BI grouped output projection with native aiter VJPs (item 23)

  completed successfully:
  Open Code Review#294:Pull request#627openedbyinaniloquentee
  Oct 28, 20261m 19s
  More actions for workflow run 294

• [DSV4][ROCm] Reuse native operators for Indexer state and MXFP4 cache (task 22)

  completed successfully:
  Open Code Review#293:Pull request#626openedbyinaniloquentee
  Oct 28, 20261m 24s
  More actions for workflow run 293

• [DSV4][ROCm] Add versioned compressor state and single cache commit (task 21)

  completed successfully:
  Open Code Review#292:Pull request#625openedbyinaniloquentee
  Oct 28, 20262m 10s
  More actions for workflow run 292

• [DSV4][ROCm] Reuse C4 and native quantization for separate Indexer compression (task 20)

  completed successfully:
  Open Code Review#291:Pull request#624openedbyinaniloquentee
  Oct 28, 20261m 17s
  More actions for workflow run 291

• [DSV4][ROCm] Add contract-bound C4 KV compressor (task 19)

  completed successfully:
  Open Code Review#290:Pull request#618synchronizebyinaniloquentee
  Oct 28, 20261m 40s
  More actions for workflow run 290

• [MiniMax-H3][CUDA] Add deterministic FFN down projection candidate

  completed successfully:
  Open Code Review#289:Pull request#535synchronizebymaxiaosong1124
  Oct 28, 20261m 32s
  More actions for workflow run 289

• H3 sigma schedule

  completed successfully:
  Open Code Review#288:Pull request#566synchronizebyfxyan
  Oct 28, 20261m 6s
  More actions for workflow run 288

• [WS1][CUDA][MiniMax-H3] Add FP32 video input projection

  completed successfully:
  Open Code Review#287:Pull request#562synchronizebymaxiaosong1124
  Oct 28, 20261m 24s
  More actions for workflow run 287

• [WS1][CUDA][MiniMax-H3] h3_rmsnorm: RMSNorm with fused AdaLN modulation (RFC #420 stack 5/10)

  completed successfully:
  Open Code Review#286:Pull request#487synchronizebymaxiaosong1124
  Oct 28, 20261m 22s
  More actions for workflow run 286

• [WS1][CUDA][MiniMax-H3] adaln_row_gather: fused six-way AdaLN row gather (RFC #420 stack 4/10)

  completed successfully:
  Open Code Review#285:Pull request#486synchronizebymaxiaosong1124
  Oct 28, 20261m 10s
  More actions for workflow run 285

• [WS1][CUDA][MiniMax-H3] adaln_projection_3mod: three-modality AdaLN projection (RFC #420 stack 3/10)

  completed successfully:
  Open Code Review#284:Pull request#485synchronizebymaxiaosong1124
  Oct 28, 20261m 3s
  More actions for workflow run 284

• [WS1][CUDA][MiniMax-H3] timestep_mlp_fp32: deterministic FP32 timestep MLP (RFC #420 stack 2/10)

  completed successfully:
  Open Code Review#283:Pull request#484synchronizebymaxiaosong1124
  Oct 28, 20261m 9s
  More actions for workflow run 283

• H3 sigma schedule

  completed successfully:
  Open Code Review#282:Pull request#566synchronizebyfxyan
  Oct 28, 20261m 19s
  More actions for workflow run 282

• feat(adaln): add indexed CUDA/Triton modulation and explicit strict qualification

  completed successfully:
  Open Code Review#281:Pull request#621synchronizebyhongleng
  Oct 28, 20261m 28s
  More actions for workflow run 281

• [WS1][Qwen-Image][CUDA] Final AdaLN output (final_adaln_out)

  completed successfully:
  Open Code Review#280:Pull request#538synchronizebyfusheng-ji
  Oct 28, 20261m 19s
  More actions for workflow run 280

• [WS1][Qwen-Image][CUDA] AdaLN modulation projection (adaln_mod_projection)

  completed successfully:
  Open Code Review#279:Pull request#537synchronizebyfusheng-ji
  Oct 28, 20261m 8s
  More actions for workflow run 279

• H3 sigma schedule

  completed successfully:
  Open Code Review#278:Pull request#566synchronizebyfxyan
  Oct 28, 20261m 27s
  More actions for workflow run 278

• feat(adaln): add indexed CUDA/Triton modulation and explicit strict qualification

  completed successfully:
  Open Code Review#277:Pull request#621synchronizebyhongleng
  Oct 28, 20261m 48s
  More actions for workflow run 277

• [WS1][Qwen-Image][CUDA] Add stride-aware multi-axis RoPE with compile support

  requires action with the application:
  Qwen-Image RoPE#5:Pull request#623synchronizebydandan20031031
  dandan20031031:feat/qwenimage-multi-axis-rope

  Oct 28, 2026Action required
  More actions for workflow run 5

• [WS1][Qwen-Image][CUDA] Add stride-aware multi-axis RoPE with compile support

  completed successfully:
  Open Code Review#276:Pull request#623synchronizebydandan20031031
  Oct 28, 20261m 24s
  More actions for workflow run 276

• [WS1][Qwen-Image][CUDA] Add stride-aware multi-axis RoPE with compile support

  completed successfully:
  Open Code Review#275:Pull request#623ready_for_reviewbydandan20031031
  Oct 28, 20261m 13s
  More actions for workflow run 275

• [WS1][Qwen-Image][CUDA] Add stride-aware multi-axis RoPE with compile support

  requires action with the application:
  Qwen-Image RoPE#4:Pull request#623ready_for_reviewbydandan20031031
  dandan20031031:feat/qwenimage-multi-axis-rope

  Oct 28, 2026Action required
  More actions for workflow run 4

• [WS1][Qwen-Image][CUDA] Add stride-aware multi-axis RoPE with compile support

  requires action with the application:
  Qwen-Image RoPE#3:Pull request#623synchronizebydandan20031031
  dandan20031031:feat/qwenimage-multi-axis-rope

  Oct 28, 2026Action required
  More actions for workflow run 3

• [WS1][Qwen-Image][CUDA] Add stride-aware multi-axis RoPE with compile support

  skipped:
  Open Code Review#274:Pull request#623synchronizebydandan20031031
  Oct 28, 20262s
  More actions for workflow run 274