Skip to content

Pull requests: mlcommons/endpoints

New pull request New
28 Open 300 Closed
28 Open 300 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

feat(metrics): full-run OSL over all agentic turns for accuracy validation size/normal PR Review Policy: <=500 non-test lines & <=20 files
#504 opened Sep 11, 2026 by arekay-nv Collaborator Loading…
1
3
feat(swebench): add trajectory-stable session routing size/normal PR Review Policy: <=500 non-test lines & <=20 files
#503 opened Sep 10, 2026 by tianmu-li Collaborator Loading…
8 of 10 tasks
7
Add submission cli test for create run size/large PR Review Policy: 501-1500 lines or 21-50 files
#502 opened Sep 10, 2026 by anandhu-eng Contributor Loading…
1
feat(load-generator): shuffle agentic conversation start order size/normal PR Review Policy: <=500 non-test lines & <=20 files
#501 opened Sep 10, 2026 by arekay-nv Collaborator Loading…
3 of 10 tasks
2
document output folder structure size/normal PR Review Policy: <=500 non-test lines & <=20 files
#498 opened Sep 9, 2026 by anandhu-eng Contributor Loading…
1
feat(swe-bench): add persistent Pyxis runtime size/very-large PR Review Policy: >1500 lines or >50 files
#496 opened Sep 5, 2026 by leopck Collaborator Loading…
1
feat(provenance): resolve source git SHA via a validated multi-channel chain size/normal PR Review Policy: <=500 non-test lines & <=20 files
#491 opened Sep 1, 2026 by arekay-nv Collaborator • Draft
3
feat(report): add events.jsonl SHA-256 to result_summary.json size/normal PR Review Policy: <=500 non-test lines & <=20 files
#490 opened Aug 31, 2026 by arekay-nv Collaborator Loading…
2
feat(agentic): auto-download official dataset size/normal PR Review Policy: <=500 non-test lines & <=20 files
#489 opened Aug 27, 2026 by tianmu-li Collaborator Loading…
7 of 10 tasks
18
feat(swe-bench): add the distributed fleet scorer size/large PR Review Policy: 501-1500 lines or 21-50 files
#485 opened Aug 27, 2026 by leopck Collaborator • 6/6 Loading…
6
feat(swe-bench): gate dispatch on endpoint readiness size/large PR Review Policy: 501-1500 lines or 21-50 files
#484 opened Aug 27, 2026 by leopck Collaborator • 5/6 Loading…
7
feat(swe-bench): classify evaluation infrastructure failures size/normal PR Review Policy: <=500 non-test lines & <=20 files
#483 opened Aug 27, 2026 by leopck Collaborator • 4/6 Loading…
1
feat(swe-bench): recover conservatively from infrastructure failures size/large PR Review Policy: 501-1500 lines or 21-50 files
#482 opened Aug 27, 2026 by leopck Collaborator • 3/6 Loading…
4
feat(swe-bench): enforce run-scoped all-or-nothing merging size/normal PR Review Policy: <=500 non-test lines & <=20 files
#481 opened Aug 27, 2026 by leopck Collaborator • 2/6 Loading…
1
feat(swe-bench): add a durable distributed work queue size/large PR Review Policy: 501-1500 lines or 21-50 files
#480 opened Aug 27, 2026 by leopck Collaborator • 1/6 Loading…
13
[docs] Benchmark suite and ruleset proposal size/normal PR Review Policy: <=500 non-test lines & <=20 files
#468 opened Aug 25, 2026 by arekay-nv Collaborator Loading…
10 tasks
42
fix(swebench-service): give Pyxis container creation its own deadline size/normal PR Review Policy: <=500 non-test lines & <=20 files
#456 opened Aug 19, 2026 by leopck Collaborator Loading…
10 tasks
2
fix(benchmark): an accuracy run that produces no number must fail size/normal PR Review Policy: <=500 non-test lines & <=20 files
#454 opened Aug 19, 2026 by leopck Collaborator Loading…
1 of 4 tasks
6
fix(benchmark): don't reject an accuracy-only run for listing many endpoints size/normal PR Review Policy: <=500 non-test lines & <=20 files
#453 opened Aug 19, 2026 by leopck Collaborator Loading…
1 of 4 tasks
2
docs: steady-state metrics reporting design proposal size/very-large PR Review Policy: >1500 lines or >50 files
#447 opened Aug 13, 2026 by nv-alicheng Collaborator Loading…
2 of 10 tasks
42
Added Conversational Aware Routing for Dynamo support
#434 opened Jul 30, 2026 by leopck Collaborator • Draft
3
Add License Scanning and Reporting for Endpoints Changes on Main.
#412 opened Jul 14, 2026 by arav-agarwal2 Collaborator Loading…
4 of 10 tasks
18
feat :Testing workflow to push images size/large PR Review Policy: 501-1500 lines or 21-50 files
#405 opened Jul 9, 2026 by arekay-nv Collaborator Loading…
10 tasks
11
fix(endpoint-client): retry pre-response connection resets (opt-in)
#404 opened Jul 8, 2026 by Palanivelg Contributor Loading…
3 tasks done
19
DRAFT: feat(power): vendor-agnostic power/energy monitoring
#383 opened Jun 30, 2026 by viraatc Collaborator Loading…
15
ProTip! no:milestone will show everything without a milestone.