Skip to content

Pull requests: NVIDIA/TensorRT-LLM

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
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort

Pull requests list

[None][test] Remove UCX cases on qa side
#17389 opened Aug 7, 2026 by fredricz-20070104 Collaborator Loading…
[None][test] Cover KVCacheManagerV2 C++ pool rebalance path
#17387 opened Aug 7, 2026 by thorjohnsen Collaborator Loading…
1 task done
[TRTLLM-15157][infra] automate stale PR cleanup
#17383 opened Aug 6, 2026 by karljang Collaborator Loading…
[None][fix] Select FlashInfer hybrid primary KV layer
#17382 opened Aug 6, 2026 by mihai-chiorean Contributor Loading…
[None][fix] Restore Eagle derived query counts
#17381 opened Aug 6, 2026 by mihai-chiorean Contributor Loading…
[TRTLLM-12680][feat] add exit code telemetry
#17377 opened Aug 6, 2026 by Mgluhovskoi Collaborator Draft
[https://nvbugs/6482566][fix] Make MoE all-to-all completion-flag timeout phase-aware
#17376 opened Aug 6, 2026 by xwang233 Collaborator Loading…
4 tasks done
[https://nvbugs/6428096][fix] Unwaive DeepSeekV3Lite compiled tp2pp2
#17375 opened Aug 6, 2026 by cascade812 Collaborator Loading…
1 task done
[None][fix] Use HND mapping for MiniMax-M3 MSA KV cache
#17374 opened Aug 6, 2026 by peihu-nv Collaborator Loading…
1 task done
[https://nvbugs/6551020][fix] Avoid SIGBUS when loading safetensors over network FS
#17368 opened Aug 6, 2026 by chenfeiz0326 Collaborator Loading…
1 task done
[TRTLLM-14388][refactor] BREAKING: Force 2 model spec dec to fall back to 1 model api-breaking Accepted LLM API contract change that is backwards-incompatible
#17366 opened Aug 6, 2026 by mikeiovine Collaborator Draft
1 task done
[None][perf] optimize qwen3.5 perf by removing gather and scatter op
#17365 opened Aug 6, 2026 by nv-guomingz Collaborator Loading…
1 task done
[https://nvbugs/6517844][test] Unwaive DeepSeek V3 Lite RTX Pro 6000D test
#17364 opened Aug 6, 2026 by nv-lschneider Collaborator Loading…
1 task done
[None][fix] Pass num_fused_shared_experts in FP8 MoE fallback-tactic path
#17362 opened Aug 6, 2026 by brnguyen2 Collaborator Loading…
1 task done
[None][feat] Default MiniMax M2 to KV cache manager V2
#17360 opened Aug 6, 2026 by liji-nv Collaborator Loading…
1 task done
ProTip! Filter pull requests by the default branch with base:main.