Skip to content

Pull requests: PrimeIntellect-ai/prime-rl

New pull request New
74 Open 1,958 Closed
74 Open 1,958 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

[DO NOT MERGE] Dummy PR to get tests to run with secrets on external PRs
#2313 opened Apr 17, 2026 by Jackmin801 Member Loading…
hrl dev90 cherry-pick #2308
#2309 opened Apr 17, 2026 by JannikSt Member Loading…
Add FA4 support to ring attention in context parallel mode
#2307 opened Apr 17, 2026 by jupiterepoch Loading…
5
feat(orchestrator): MultiAgentEnv extraction for debate protocols
#2306 opened Apr 17, 2026 by joanvelja • Draft
Pin Triton JIT cache to node-local /tmp per job
#2305 opened Apr 17, 2026 by rasdani Contributor Loading…
feat: proper context parallelism for GLM-5 sparse MLA
#2293 opened Apr 17, 2026 by samsja Member • Draft
refactor: propagate shared config via before validator
#2291 opened Apr 16, 2026 by mikasenghaas Member Loading…
3
refactor: regex-based auto-resolution for tool call and reasoning parsers
#2290 opened Apr 16, 2026 by mikasenghaas Member • Draft
fix: log inference metrics on time axis instead of step
#2288 opened Apr 15, 2026 by samsja Member • Draft
1
feat: renderers v2 — client-side tokenization for training
#2278 opened Apr 14, 2026 by hallerite Member Loading…
8
feat: drop filtered rollouts instead of masking
#2277 opened Apr 14, 2026 by mikasenghaas Member Loading…
1
feat: per-batch loss type override for hosted SFT hard distill
#2271 opened Apr 14, 2026 by eligotts Contributor Loading…
11
feat: shared rollout limiter for train and eval concurrency
#2263 opened Apr 13, 2026 by mikasenghaas Member Loading…
8
Codex/gb200 build pipeline
#2262 opened Apr 13, 2026 by S1ro1 Collaborator • Draft
fix: handle malformed preprocessing payloads
#2253 opened Apr 10, 2026 by teilomillet • Draft
chore: add qwen3_vl_moe to VLM registry docs table
#2250 opened Apr 10, 2026 by hallerite Member Loading…
fix: propagate mm_token_type_ids through VLM training pipeline
#2249 opened Apr 10, 2026 by hallerite Member Loading…
1
feat: add num_epochs config to SFT trainer
#2246 opened Apr 10, 2026 by hallerite Member • Draft
8
feat: VLM On-Policy Distillation
#2245 opened Apr 10, 2026 by JacobHelwig Loading…
5
feat: ExtraExpert PEFT for MoE
#2224 opened Apr 7, 2026 by samsja Member • Draft
3 tasks
feat: add trainer/orchestrator step summary panes to tmux
#2222 opened Apr 7, 2026 by samsja Member • Draft
fix: force weight broadcast at step 0 for NCCL mode
#2209 opened Apr 6, 2026 by samsja Member Loading…
9
feat: add env args scheduler hot reload
#2207 opened Apr 6, 2026 by snimu Collaborator Loading…
7
Fix Prime monitor auth headers for default RFT API
#2205 opened Apr 6, 2026 by d42me Collaborator Loading…
Align DeepGEMM vLLM environment variable with InferenceConfig
#2202 opened Apr 5, 2026 by JacobHelwig Loading…
2
ProTip! Find all pull requests that aren't related to any open issues with -linked:issue.