Skip to content

Pull requests: oumi-ai/oumi

New pull request New
30 Open 2,483 Closed
30 Open 2,483 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

Add MedQA example
#2648 opened Sep 5, 2026 by shanghongsim Contributor • Queued
2 of 4 tasks
1
feat(verl): route agent-loop rows via agent_name and validate the config
#2613 opened Aug 14, 2026 by aniruddh-alt Contributor • 4/4 • Draft
1
feat(verl): UserSimToolAgentLoop for simulated-user rollouts
#2612 opened Aug 14, 2026 by aniruddh-alt Contributor • 3/4 • Draft
1
feat(verl): process-wide user-simulator inference engine
#2611 opened Aug 14, 2026 by aniruddh-alt Contributor • 2/4 • Draft
1
aniruddh alt/usersim 01 rollout core
#2610 opened Aug 14, 2026 by aniruddh-alt Contributor • 1/4 Loading…
1
feat(tools): reject unsupported JSON Schema keywords in tool definitions
#2597 opened Aug 12, 2026 by wizeng23 Collaborator Loading…
3 of 4 tasks
8
fix(verl): make the Countdown GRPO example runnable and checkpoint-producing
#2592 opened Aug 10, 2026 by aniruddh-alt Contributor Loading…
8
Tool response in assistant span detection and tests
#2590 opened Aug 7, 2026 by shanghongsim Contributor Loading…
2 of 4 tasks
30
feat(collators): opt-in end-of-turn supervision + nested tool-result masking
#2589 opened Aug 6, 2026 by ryan-arman Contributor • Draft
4 tasks
1
fix(quantize): repair CLI output_format and output_path defaults
#2585 opened Aug 4, 2026 by ryan-arman Contributor • Draft
1
fix(launcher): reserve CPU/memory on Modal sandboxes from JobConfig resources
#2572 opened Jul 28, 2026 by lefft Contributor • Draft
3 of 4 tasks
1
Remove ceiling on transformers for uv override only
#2567 opened Jul 24, 2026 by wizeng23 Collaborator Loading…
6
fix(train): cap Ray CPUs via OUMI_RAY_NUM_CPUS
#2566 opened Jul 24, 2026 by aniruddh-alt Contributor • Draft
3
feat(examples): conversational GRPO customer-support demo
#2565 opened Jul 24, 2026 by aniruddh-alt Contributor • Draft
3
feat(grpo): conversation LLM-judge reward
#2564 opened Jul 24, 2026 by aniruddh-alt Contributor • Draft
5
feat(verl): UserSimToolAgentLoop for tool + simulated-user GRPO rollouts
#2563 opened Jul 24, 2026 by aniruddh-alt Contributor Loading…
24
feat(examples): NL2SQL Spider tool-agent GRPO demo (configs + row builder)
#2562 opened Jul 24, 2026 by aniruddh-alt Contributor Loading…
15
feat(environments): NL2SQL run_sql executor + sql_execution_match reward
#2561 opened Jul 24, 2026 by aniruddh-alt Contributor Loading…
5
feat(models): add text_only opt-in for dual-mode checkpoints (e.g. Qwen3.5)
#2557 opened Jul 23, 2026 by wizeng23 Collaborator • Draft
2 of 4 tasks
1
build(deps): raise vLLM floor to >=0.19,<0.20 (both extras)
#2546 opened Jul 17, 2026 by jgreer013 Contributor • Draft
1
Update community links
#2534 opened Jul 8, 2026 by taenin Collaborator Loading…
1 of 4 tasks
4
docs: add lm_eval MAX_PATH and cp1252 errors to Windows troubleshooting
#2532 opened Jul 7, 2026 by ppotash Loading…
2 of 4 tasks
3
feat(environments): BrowserExecutableEnvironment (Kernel cloud browser for web-agent training)
#2529 opened Jun 30, 2026 by aniruddh-alt Contributor • Draft
9
fix(launcher): downgrade Slurm connection-check log from error to debug
#2525 opened Jun 23, 2026 by rlehman221 Contributor • Draft
2 of 4 tasks
build(deps): update torchvision requirement from <0.26,>=0.21 to >=0.21,<0.28
#2510 opened Jun 15, 2026 by dependabot Bot Loading…
@oelachqar @wizeng23
1
ProTip! Type g i on any issue or pull request to go back to the issue listing page.