Skip to content

Pull requests: PrimeIntellect-ai/prime-rl

New pull request New
113 Open 2,377 Closed
113 Open 2,377 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

feat(configs): Kimi-K2.7-Code PD-disaggregated wide-EP inference config
#2827 opened Jun 15, 2026 by hallerite Member • Draft
[ENG-4094] docker: include cuda toolkit with runtime image
#2826 opened Jun 15, 2026 by eexwhyzee Collaborator Loading…
chore: bump vLLM to 0.23.0 and prune/adapt monkeypatches
#2825 opened Jun 15, 2026 by hallerite Member Loading…
chore: point verifiers env paths at the flattened environments/ dir
#2824 opened Jun 15, 2026 by mikasenghaas Member • Draft
[codex] add sparse filesystem weight sync
#2823 opened Jun 15, 2026 by samsja Member Loading…
3
[codex] Fix Qwen3.5 MRoPE VLM model
#2819 opened Jun 14, 2026 by eligotts Contributor • Draft
[codex] Truncate inference rank logs with tee
#2817 opened Jun 14, 2026 by samsja Member • Draft
explicit del and malloc
#2816 opened Jun 14, 2026 by cdreetz Contributor Loading…
@samsja @mikasenghaas
2
feat: pause inference during trainer checkpoints
#2815 opened Jun 14, 2026 by samsja Member • Draft
4
[codex] compact training sample transport payloads
#2809 opened Jun 14, 2026 by samsja Member Loading…
[codex] reduce retained train trace memory
#2807 opened Jun 14, 2026 by samsja Member Loading…
7
fix: surface deep_gemm load failures instead of silently disabling FP8
#2803 opened Jun 13, 2026 by S1ro1 Collaborator • Draft
fix(laguna): match vLLM YaRN attention scaling to fix RL KL mismatch
#2802 opened Jun 13, 2026 by S1ro1 Collaborator • Draft
[codex] add fake R3 orchestrator debug mode
#2801 opened Jun 13, 2026 by samsja Member Loading…
2
feat(models): declarative invertible conversion-op framework for all models
#2797 opened Jun 13, 2026 by S1ro1 Collaborator • Draft
fix(orchestrator): bound rollout buffer memory (OOM at 1 TiB on long …
#2796 opened Jun 13, 2026 by samsja Member Loading…
2
feat(orchestrator): add static dataset SFT
#2795 opened Jun 12, 2026 by tim0120 Contributor • Draft
Validate requested attention implementation per model at load time
#2793 opened Jun 12, 2026 by hallerite Member Loading…
3
Add Xiaomi MiMo-V2.5 MoE model support
#2791 opened Jun 12, 2026 by hallerite Member • Draft
Add Tencent Hy3 (HYV3) MoE model support
#2789 opened Jun 12, 2026 by hallerite Member • Draft
feat(algo): RLCSD — contrastive on-policy self-distillation
#2788 opened Jun 12, 2026 by hallerite Member • Draft
Fix/readme trainer torchrun
#2787 opened Jun 12, 2026 by anravich13-cloud Loading…
6
fix: is_tt_moe_model checks text_config for VLM-style MoE models
#2780 opened Jun 11, 2026 by Levichev Loading…
feat: pull-based NIXL weight broadcast with lazy-tensor plan discovery
#2779 opened Jun 11, 2026 by S1ro1 Collaborator • Draft
Add inference backend A/B benchmark
#2777 opened Jun 11, 2026 by RitwijParmar • Draft
1
ProTip! Add no:assignee to see everything that’s not assigned.