-
Notifications
You must be signed in to change notification settings - Fork 595
Pull requests: Gen-Verse/OpenClaw-RL
Pull requests list
fix: skip DDP and optimizer for prm_teacher (inference-only mode)
#115
opened May 14, 2026 by
HeJW0303
Loading...
3 tasks done
feat: implement single-GPU (24GB) INT4 QLoRA + offload-rollout + external PRM for OpenClaw-Combine and test
#97
opened Apr 20, 2026 by
wuwanchun
Loading...
fix: repad rollout samples after sample filtering
#79
opened Apr 15, 2026 by
xuyangc03
Contributor
Loading...
Fix "strip_thinking function" and "default model name for OpenClaw requests"
#53
opened Mar 30, 2026 by
zlh-source
Loading...
Add initial Qwen3.5-4B support and SGLang text-runtime patch
#39
opened Mar 19, 2026 by
ShenAC-SAC
Loading...
docs: add Chinese documentation for core RL methods
#33
opened Mar 17, 2026 by
guozhihao-224
Loading...
feat(openclaw-rl): add reward model feedback training
#28
opened Mar 14, 2026 by
chengzidl
Loading...
fix: add cd to SLIME_ROOT and derive MEGATRON_ROOT dynamically
#21
opened Mar 12, 2026 by
JasonOA888
Loading...
2 tasks done
feat: add openclaw-training-free β zero-GPU experience loop for LLMs
#6
opened Mar 5, 2026 by
zcyynl
Loading...
fix: add cd to SLIME_ROOT before train_async.py in all training scripts
#5
opened Mar 4, 2026 by
kaysonyu
Loading...
fix: derive Megatron-LM PYTHONPATH from script location
#3
opened Mar 1, 2026 by
koriyoshi2041
Loading...
4 tasks done
docs: fix kl-loss-coef default value in instructions
#2
opened Mar 1, 2026 by
koriyoshi2041
Loading...
2 tasks done
fix: correct PYTHONBUFFERED typo to PYTHONUNBUFFERED
#1
opened Mar 1, 2026 by
koriyoshi2041
Loading...
3 tasks done
ProTip!
Type g i on any issue or pull request to go back to the issue listing page.
You canβt perform that action at this time.