All pull requests

Pull requests

• [API] Require a declared stop for --mode mp4

  #667·spaingnv opened Sep 28, 2026Collaborator·

  ·
  1

  3

• [Naming] Change teahv to taehv in code-base

  #666·ArielG-NV opened Sep 28, 2026Collaborator·

  ·
  1

  6

• [Demo - Lingbot] Add a new 1.3b config with model knobs to balance performance/accuracy

  #660·ArielG-NV opened Sep 25, 2026Collaborator·

  ·
  8

• 'preload' command-line option & static analysis to verify preload effectiveness for application

  #658·ArielG-NV opened Sep 24, 2026Collaborator·

  ·
  1

  4

• Support view-scoped multi-view conditioning

  #656·aidanfnv opened Sep 23, 2026Collaborator·

  ·
  7

• [Crazy Robotaxi] Pre-encode bounded dynamic prompts

  #655·aidanfnv opened Sep 23, 2026Collaborator·

  ·
  1

  12

• [Cam2V] Finish the overlay after --total-blocks

  #652·spaingnv opened Sep 22, 2026Collaborator·

  ·
  1

  15

• Add Crazy Robotaxi prompt fragment toggles

  #643·aidanfnv opened Sep 21, 2026Collaborator·

  ·
  1

  9

• [Crazy Robotaxi] Pace model generation at the declared 30 FPS

  #642·aidanfnv opened Sep 21, 2026Collaborator·

  ·
  1

  6

• [Crazy RoboTaxi] Decouple display and model resolutions

  #641·aidanfnv opened Sep 21, 2026Collaborator·

  ·
  1

  15

• [Feat] sageattention2 backend

  #639·Shn0930 opened Sep 21, 2026·

  ·
  10

• feat(omnidreams): add ClipGT-to-USDZ and AlpaSim closed-loop workflow for OmniDreams

  #638·Shn0930 opened Sep 18, 2026·

  ·
  4

• feat(omnidreams): add layer-wise DiT parameter offload

  #637·Shn0930 opened Sep 18, 2026·

  ·
  13

• [API] Add NVTX ranges and a flashdreams-profile command

  #634·psamin opened Sep 18, 2026·

  ·
  14

• Add memory-efficient dense attention and lazy encoder imports

  #632·jmccaffrey-nv opened Sep 17, 2026Collaborator·

  ·
  5

• [API] Return list[StepResult] from both loop steps (#562)

  #630·spaingnv opened Sep 15, 2026Collaborator·

  ·
  1

  9

• Add tensor and context parallel inference to the core runtime

  #619·jmccaffrey-nv opened Sep 11, 2026Collaborator·

  ·
  20

• [Models] Add LongSANA Runtime V2 text-to-video pipeline

  #591·jmccaffrey-nv opened Sep 5, 2026Collaborator·

  ·
  9

• Backfill total_s from Cam2V's model_step_wall_s for LingBot benchmark scenarios

  #589·AmirF194 opened Sep 4, 2026·

  ·
  13

• Gamelayer

  #587·3a1b2c3 opened Sep 4, 2026Contributor·

  ·
  4

• Add correlated V2 runtime latency profiles

  #574·ZenAlexa opened Sep 3, 2026·

  ·
  1

  23

• Collect model input after cadence pacing

  #573·ZenAlexa opened Sep 3, 2026·

  ·
  16

• [Demos] Complete Waypoint V2 migration with Cam2V HUD

  #572·jmccaffrey-nv opened Sep 3, 2026Collaborator·

  ·
  5

• Persist torch.compile cache across GPU CI jobs

  #570·spaingnv opened Sep 2, 2026Collaborator·

  ·
  1

  7

• Reduce Robotaxi input sampling latency

  #560·gtong-nv opened Sep 1, 2026Collaborator·

  ·
  2