-
Notifications
You must be signed in to change notification settings - Fork 524
Pull requests: areal-project/AReaL
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
feat(cli): add experimental cli scaffold for service-style subcommands
#1440
opened Jun 24, 2026 by
sitabulaixizawaluduo
Collaborator
Loading…
5 of 16 tasks
fix(CI): fix vlm_grpo CI OOM bug
safe-to-test
Ready to run unit-tests in a PR.
#1438
opened Jun 23, 2026 by
sitabulaixizawaluduo
Collaborator
Loading…
1 of 15 tasks
Feat/training service cli
#1435
opened Jun 22, 2026 by
sitabulaixizawaluduo
Collaborator
•
Draft
2 of 15 tasks
Feat/inference service cli
#1434
opened Jun 22, 2026 by
sitabulaixizawaluduo
Collaborator
•
Draft
2 of 15 tasks
fix(io_struct): support multi-EOS models in stop-token handling
#1433
opened Jun 22, 2026 by
PheelaV
Loading…
8 of 15 tasks
fix(stats): export(key=..., reset=True) pops the wrong dict and raises KeyError
#1432
opened Jun 22, 2026 by
EazyReal
Contributor
Loading…
4 tasks done
docs: mirgate and clean the documents
#1431
opened Jun 22, 2026 by
mingcheng
Contributor
Loading…
8 of 15 tasks
feat: add W&B worker GPU system metrics
#1428
opened Jun 21, 2026 by
EazyReal
Contributor
Loading…
9 of 16 tasks
fix(dataset): correct GSM8K SFT loss-mask boundary for merged tokens
#1427
opened Jun 21, 2026 by
EazyReal
Contributor
Loading…
8 of 9 tasks
fix(reward): bound MathVerifyWorker.verify wall-clock on a hung verification
#1426
opened Jun 20, 2026 by
EazyReal
Contributor
Loading…
7 of 9 tasks
fix: don't inject None into stop_token_ids when pad/eos token is None
#1425
opened Jun 20, 2026 by
EazyReal
Contributor
Loading…
7 of 9 tasks
refactor(workflow): extract grouped rollout wrapper
#1418
opened Jun 16, 2026 by
RanranranQAQ
Loading…
5 of 15 tasks
feat(ppo): configurable loss aggregation — token / seq / prompt mean (ScaleRL §3.2)
#1417
opened Jun 16, 2026 by
EazyReal
Contributor
Loading…
8 of 9 tasks
feat(rollout): add min_valid_group_size to drop under-filled rollout groups
#1416
opened Jun 16, 2026 by
EazyReal
Contributor
Loading…
fix(ppo): group-normalize by actual group sizes for partial groups
#1415
opened Jun 16, 2026 by
EazyReal
Contributor
Loading…
refactor(ppo): derive group-norm group_size from gconfig.n_samples
#1413
opened Jun 16, 2026 by
EazyReal
Contributor
Loading…
8 of 9 tasks
fix(openai): render tool-call arguments as a mapping for HF chat templates
#1411
opened Jun 16, 2026 by
EazyReal
Contributor
Loading…
7 of 9 tasks
feat(experimental): Diffusion RL post-training — Phase 1 PoC (SD1.5 + LoRA + REINFORCE)
#1410
opened Jun 15, 2026 by
Fyrgo8
Loading…
5 of 6 tasks
feat: trajectory dump/replay for offline training-loop debugging
#1407
opened Jun 12, 2026 by
Fyrgo8
Loading…
5 of 9 tasks
Support Megatron FP8 weight transfer in AWEX colocate mode
#1406
opened Jun 11, 2026 by
equation314
Loading…
8 of 14 tasks
ci: add PyPI publish workflow and fix Megatron deps 🚀
stale
#1404
opened Jun 10, 2026 by
mingcheng
Contributor
Loading…
7 of 15 tasks
feat(distillation): Multi-Teacher On-Policy Distillation Support
#1400
opened Jun 8, 2026 by
zahrayousefijamarani
Contributor
Loading…
6 of 15 tasks
fix: Prevent workers from applying dp-scaled staleness to fix rollout hanging issues caused by zero local capacity
#1396
opened Jun 8, 2026 by
zcsh
Loading…
3 of 14 tasks
fix: add group_id to StartSessionRequest for online GRPO session grouping
#1392
opened Jun 5, 2026 by
Oxygen56
Loading…
feat(experimental): enable DTA training for Archon DP
stale
#1391
opened Jun 5, 2026 by
ezoicoder
Collaborator
Loading…
8 of 15 tasks
Previous Next
ProTip!
Find all pull requests that aren't related to any open issues with -linked:issue.