All pull requests

Pull requests

• fix(sglang): validate deployment config before launch

  #2423·CharlesXu-HQ opened Sep 25, 2026·

  ·

• fix(agent): terminate timed-out sandbox process groups

  #2422·CharlesXu-HQ opened Sep 25, 2026·

  ·

• fix(ray): roll back partial train group creation

  #2421·CharlesXu-HQ opened Sep 25, 2026·

  ·

• fix(sglang): bound local server startup waits

  #2420·CharlesXu-HQ opened Sep 25, 2026·

  ·

• fix(rollout): scope remote RM sessions to event loops

  #2419·CharlesXu-HQ opened Sep 25, 2026·

  ·

• fix(agent): linearize session shutdown with request admission

  #2418·CharlesXu-HQ opened Sep 25, 2026·

  ·

• fix(weight-sync): publish restored baseline on delta resume

  #2417·CharlesXu-HQ opened Sep 25, 2026·

  ·

• perf(reward): offload rule-based scoring from event loop

  #2416·CharlesXu-HQ opened Sep 25, 2026·

  ·

• fix(checkpoint): reject inconsistent PPO resume state

  #2415·CharlesXu-HQ opened Sep 25, 2026·

  ·

• fix(rollout): support fan-out results in group reward models

  #2414·CharlesXu-HQ opened Sep 25, 2026·

  ·

• fix(rollout): handle nested aborted fully-async results

  #2413·CharlesXu-HQ opened Sep 25, 2026·

  ·
  1

• fix(rollout): score fan-out samples as a flat group

  #2412·yuchenwang3 opened Sep 24, 2026·

  ·

• fix(rm_hub): stop substring-matching gpqa letter labels and short answers

  #2411·shaurya416 opened Sep 24, 2026·

  ·
  1

• Support distributed rollout for further scaling

  #2410·zhuzilin opened Sep 24, 2026Contributor·

  ·

• fix: preserve per-sample training metadata across DP transport

  #2403·heyxingxing opened Sep 22, 2026·

  ·
  1

• fix: propagate fully async worker failures to training

  #2401·heyxingxing opened Sep 22, 2026·

  ·
  1

• refactor: migrate Slime MUSA bootstrap to musa_titan

  #2398·ForAxel opened Sep 21, 2026Contributor·

  ·

• feat(profile): safer torch.profiler defaults (rank-0 only + env-gated metadata)

  #2397·leofan-i opened Sep 21, 2026Contributor·

  ·

• fix(args): reject a non-positive --eval-temperature

  #2396·PerryLink opened Sep 20, 2026·

  ·

• fix: isolate dataset shuffle random state

  #2395·LZC-BELIEVER opened Sep 19, 2026·

  ·

• fix(vlm): preserve token IDs in multi-turn Qwen rollouts

  #2389·zupengwang opened Sep 16, 2026·

  ·
  1

• fix(R3): reduce host memory usage for rollout routing replay

  #2385·JayPritchet opened Sep 14, 2026·

  ·
  1

• feat: add fused Triton log-probability backend for Megatron

  #2383·YolandaLyj opened Sep 14, 2026·

  ·

• fix: add an opt-out for final checkpoint saves

  #2382·NOAH1ARK opened Sep 13, 2026·

  ·
  1

• fix(rollout): handle fan-out dynamic filter groups

  #2381·QIU-Guanzong opened Sep 13, 2026·

  ·
  1