Skip to content

Pull requests: THUDM/slime

New pull requestNew
214 Open 1,559 Closed
214 Open 1,559 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviewsReview requiredApproved reviewChanges requested
Assignee
Filter by who’s assigned
Assigned to nobodyLoading
Sort
Sort by
NewestOldestMost commentedLeast commentedRecently updatedLeast recently updatedBest match
Most reactions
👍👎😄🎉😕❤️🚀👀

Pull requests list

feat(agent): optionally preserve native parallel tool calls
#2283 opened Aug 18, 2026 by Zeng-WeijunLoading…
fix(agent): restore a re-rendered assistant echo before rendering the prompt
#2282 opened Aug 18, 2026 by pollyloopLoading…
feat: track policy versions in fully async rollout
#2279 opened Aug 17, 2026 by 0z5aLoading…
feat: add deterministic fully-async staleness control
#2278 opened Aug 17, 2026 by 0z5a • Draft
fix: preserve text-only rows in multimodal datasets
#2277 opened Aug 16, 2026 by VanyMoLoading…
feat: support streaming external rollouts
#2272 opened Aug 14, 2026 by AphohLoading…
[Fix] Support origin HF weight backfilling in parallel converter
#2270 opened Aug 13, 2026 by albaNnaksqrContributorLoading…
Fix O(#segments) per-comm memory probe that slows weight sync & training under expandable_segments
#2269 opened Aug 13, 2026 by yszhliLoading…
1
[docker] Upgrade to sglang v0.5.17
#2265 opened Aug 12, 2026 by zhuzilinContributorLoading…
fix: load critic from policy checkpoints without value head
#2259 opened Aug 9, 2026 by DodojordiLoading…
Fix: skip KL backward graph when coefficient is zero
#2258 opened Aug 9, 2026 by DodojordiLoading…
fix(rollout): guard zero-length response against x[-0:] returning the whole sequence
#2255 opened Aug 6, 2026 by hobostayContributorLoading…
fix: pool speculative-decoding metrics instead of averaging per-sample ratios
#2240 opened Jul 26, 2026 by keepkeenContributorLoading…
[Multimodal][Model] Make Qwen3.5-VL work with packed sequences
#2233 opened Jul 25, 2026 by TobyYang7Loading…
fix: avoid in-place mutation of param in MTP weight conversion (related to #2131)
#2227 opened Jul 21, 2026 by botbikamordehai2-sketchLoading…
1
Fix ref weights being broadcast on the first async update after resume
#2224 opened Jul 20, 2026 by ishzguLoading…
[docker] backport output-gate slicing when query groups < TP
#2221 opened Jul 20, 2026 by LLMSharkLoading…
fix(rollout): make dynamic refills granular
#2218 opened Jul 18, 2026 by EazyRealContributorLoading…
feat: add backend-aware MUSA support
#2216 opened Jul 17, 2026 by ForAxelLoading…
1
Support NemotronH hybrid Mamba+Attention+MoE training (e.g. Nemotron Nano-30B-A3B)
#2211 opened Jul 16, 2026 by HelloWorldLTYContributorLoading…
[Rollout] Add opt-in group-scoped session affinity
#2206 opened Jul 14, 2026 by chengcuipingLoading…
1
fix: normalize rewards by explicit sample groups
#2204 opened Jul 14, 2026 by morlutoContributorLoading…
1
1
fix: validate DAPO integer labels without float coercion
#2203 opened Jul 14, 2026 by morlutoContributorLoading…
1
1
Add critic load fallback for resuming training
#2200 opened Jul 13, 2026 by coding-famerContributorLoading…
fix: guard zero rollout temperature logprob scaling
#2197 opened Jul 12, 2026 by EazyRealContributorLoading…
ProTip! Add no:assignee to see everything that’s not assigned.