Skip to content

Pull requests: THUDM/slime

New pull request New
100 Open 1,312 Closed
100 Open 1,312 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 reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

Add GLM5 SFT support
#1844 opened Apr 20, 2026 by samaritan1998 Loading…
2
Fix double prepare_grads / loss-scaler-double-update in train_one_step
#1842 opened Apr 17, 2026 by jthomy Loading…
1
fix(sft): enable max-length filtering for messages datasets
#1841 opened Apr 17, 2026 by none0663 Contributor Loading…
3 tasks done
Add support for NVIDIA DGX Spark (GB10 / sm_121a, arm64)
#1835 opened Apr 15, 2026 by boots-coder Loading…
2
Fix missing activation checkpointing (recompute) parameters in bridge mode
#1833 opened Apr 14, 2026 by XJL010622 Loading…
[build] Add A100 support: patch set, offline-friendly conda build, and examples
#1832 opened Apr 14, 2026 by jason9693 Loading…
fix(gemma3): use GeGLU activation instead of SwiGLU
#1825 opened Apr 10, 2026 by leofan-lab Loading…
feat: add GLM-5 SFT loss mask support
#1824 opened Apr 10, 2026 by stevewx Contributor Loading…
4 tasks done
fix address already in use
#1819 opened Apr 8, 2026 by xutianming Contributor Loading…
Add Qwen3.5 VLM CI run-ci-changed
#1814 opened Apr 7, 2026 by zhuzilin Contributor Loading…
3
fix: auto-fallback to flash_attn for Qwen3.5 on pre-Hopper GPUs (head_dim=256)
#1808 opened Apr 6, 2026 by dadiaomengmeimei Loading…
2
feat: delta compression for weight sync
#1806 opened Apr 5, 2026 by nanjiangwill Collaborator Loading…
3 of 5 tasks
Supporting FIPO (Future-KL Influenced Policy Optimization)
#1801 opened Apr 3, 2026 by SeungyounShin Loading…
feat: add checkpoint retention limit to automatically clean up old checkpoints
#1798 opened Apr 2, 2026 by stevewx Contributor Loading…
4 tasks done
Add rollout sampling-mask support run-ci-short
#1795 opened Apr 2, 2026 by yitianlian Collaborator Loading…
10
Enhanced Off-Policy Async Rollout with Staleness Control and Partial Rollout Support
#1781 opened Mar 30, 2026 by huang3eng Contributor Loading…
3
Hook proposal
#1774 opened Mar 27, 2026 by andrija-s • Draft
[kimi25 rl part4] Support K25 HF weight conversion between BF16\FP8\INT4
#1757 opened Mar 23, 2026 by Gao016 Contributor Loading…
[kimi25 rl part3] support K25 VL rollout processor and train-time token expansion
#1755 opened Mar 23, 2026 by GeLee-Q Contributor • Draft
[kimi25 rl part2] pass megatron bridge provider args from slime config
#1754 opened Mar 23, 2026 by GeLee-Q Contributor Loading…
[kimi25 rl part1.2] support kimi25 q-lora pairing in bridge update path (weight update for train-infer colocate)
#1753 opened Mar 23, 2026 by GeLee-Q Contributor Loading…
[docker] fix qwen3_vl visual module loading
#1727 opened Mar 15, 2026 by ZHZisZZ Loading…
3
Add Mooncake Backend for Rollout Data Transfer run-ci-megatron
#1709 opened Mar 11, 2026 by zxpdemonio Loading…
6 tasks done
PipelineRL -- keep cache on weight update
#1694 opened Mar 9, 2026 by hari-hm Contributor Loading…
1
fix: normalize rewards per-group when sample counts are unequal
#1655 opened Mar 2, 2026 by dubin555 Loading…
2 of 3 tasks
1
ProTip! Find all pull requests that aren't related to any open issues with -linked:issue.