Skip to content

Pull requests: radixark/miles

New pull request New
844 Open 1,769 Closed
844 Open 1,769 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

[AMD] Fall back to mcore async checkpointing when NVRx is unavailable
#2816 opened Aug 30, 2026 by michaelzhang-ai Contributor Loading…
dashboard: land Rollouts on the newest step that has data
#2815 opened Aug 30, 2026 by Shi-Dong Contributor Loading…
fix(rollout): drop groups with missing rewards
#2814 opened Aug 30, 2026 by guapisolo Collaborator Loading…
feat(ci): add user allowlists for command tiers
#2813 opened Aug 30, 2026 by guapisolo Collaborator Loading…
openenv: stop the episode when the session crosses --max-seq-len
#2812 opened Aug 29, 2026 by chethanuk Loading…
1
openenv: raise the tbench2_env contract floor to OpenEnv#1025
#2811 opened Aug 29, 2026 by nblintao Contributor Loading…
openenv e2b backend: declare the SDK as an extra, floor it at 2.12
#2809 opened Aug 29, 2026 by nblintao Contributor Loading…
examples: move the Harbor-on-Daytona recipe onto the in-process path
#2807 opened Aug 28, 2026 by nblintao Contributor • 3/3 • Draft
examples: run Harbor in-process on cloud sandbox backends
#2806 opened Aug 28, 2026 by nblintao Contributor • 2/3 • Draft
rollout: shared miles-side layer for agent-function legs (session URL, sandbox credentials)
#2805 opened Aug 28, 2026 by nblintao Contributor • 1/3 • Draft
agent functions: one failure contract for the OpenEnv, NeMo Gym and Harbor legs
#2802 opened Aug 28, 2026 by nblintao Contributor • 2/2 • Draft
rollout: let agent functions discard infra-aborted samples
#2801 opened Aug 28, 2026 by nblintao Contributor • 1/2 • Draft
Add Qwen3.8-Flash-Next vision RL support
#2799 opened Aug 28, 2026 by nanjiangwill Contributor Loading…
2
fix(replay): align BSHD data with Megatron token order
#2798 opened Aug 28, 2026 by Shikhargupta Loading…
fix(lora): exclude adapters when loading base checkpoint
#2797 opened Aug 28, 2026 by Shikhargupta Loading…
fix(loss): exclude masked samples from token normalizer
#2796 opened Aug 28, 2026 by Shikhargupta Loading…
docker: hold the wheels release in its own stage, so 1.7 GB of it stops shipping in every published image
#2795 opened Aug 28, 2026 by sujeito-operator Loading…
fix(deepseek_v4): chunk indexer queries on demand
#2793 opened Aug 28, 2026 by EazyReal Loading…
Add GLM-5.3 vision-language RL support
#2792 opened Aug 28, 2026 by nanjiangwill Contributor Loading…
test(tito): use FP8 Qwen3.6 checkpoint
#2791 opened Aug 28, 2026 by guapisolo Collaborator Loading…
2
Add low-precision profiler summaries
#2789 opened Aug 27, 2026 by anson627 Loading…
GLM 5.3 Flash RL Support run-ci-megatron run-ci-model-scripts Run model script smoke tests
#2786 opened Aug 27, 2026 by Zhichenzzz Contributor Loading…
2
[AMD][LoRA] Add a Qwen3-30B-A3B MoE LoRA launcher
#2784 opened Aug 27, 2026 by Arist12 Contributor Loading…
Partition the rollout radix cache by the session's initial weight version
#2783 opened Aug 27, 2026 by fzyzcjy Collaborator Loading…
Default fully-async recipes to retract weight-update pauses
#2782 opened Aug 27, 2026 by fzyzcjy Collaborator Loading…
ProTip! Type g i on any issue or pull request to go back to the issue listing page.