Pull requests: NVIDIA/TensorRT-LLM
Pull requests list
[None][test] Enable gen_only + ctx_only DeepSeek-V4-Pro perf-sanity on GB300
#16795
opened Jul 23, 2026 by
chenfeiz0326
Collaborator
Loading…
2 tasks
[TRTLLM-14571][infra] Enable container-local AutoTuner cache in CI
#16794
opened Jul 23, 2026 by
YihuiLu512
Collaborator
Loading…
1 task done
[https://nvbugs/6490033][fix] Relaxed the assertion to accept both
is_eagle3() and…
#16793
opened Jul 23, 2026 by
trtllm-agent
Collaborator
Loading…
2 tasks done
[None][perf] Skip DeepGEMM clean_logits in DSA indexer prefill on custom top-k path
#16789
opened Jul 23, 2026 by
dc3671
Collaborator
Loading…
1 task done
[https://nvbugs/6490028][fix] Bump only the cross-library
cublas_tolerance from 1.05 to 1.10 in the test…
#16788
opened Jul 23, 2026 by
trtllm-agent
Collaborator
Loading…
2 tasks done
[None][feat] Prefer V2 transceiver backend for Gemma and Llama
#16787
opened Jul 23, 2026 by
moraxu
Collaborator
Loading…
1 task done
[https://nvbugs/6485885][fix] Stop thinking-budget processor re-forcing the reasoning end tag
#16785
opened Jul 23, 2026 by
Wanli-Jiang
Collaborator
Loading…
1 task done
[None][infra] Fix release check failure for .test_durations
#16784
opened Jul 23, 2026 by
EmmaQiaoCh
Collaborator
Loading…
1 task done
[None][perf] Preserve default V2 KV cache pool sizing
#16783
opened Jul 23, 2026 by
2ez4bz
Collaborator
Loading…
1 task done
[TRTLLM-14541][fix] VisualGen: byte-identical tune-run vs cache-load on multi-GPU
VisualGen
#16782
opened Jul 23, 2026 by
luyiyun1021
Collaborator
Loading…
1 task done
[TRTLLM-14417][fix] MoE: fp32 accumulation in deferred MoEAllReduce finalize
#16778
opened Jul 23, 2026 by
xwang233
Collaborator
Loading…
2 tasks done
[TRTLLM-12838][infra] CBTS: coverage-based test selection
#16776
opened Jul 23, 2026 by
crazydemo
Collaborator
Loading…
1 task done
[https://nvbugs/6463822][fix] Fix LTX2 CUDA graph test leak issue
#16775
opened Jul 23, 2026 by
yibinl-nvidia
Collaborator
•
Draft
1 task
[None][feat] Support DeepSeek-V4 in layer_wise_benchmarks
#16774
opened Jul 23, 2026 by
ruodil
Collaborator
Loading…
1 task done
[None][feat] Add Cosmos3-Edge (Nemotron-dense) support
VisualGen
#16773
opened Jul 23, 2026 by
ishovkun
Loading…
1 task done
[#16767][fix] Fix DSpark rolling-window slot collision in disaggregated serving
#16772
opened Jul 23, 2026 by
longlee0622
Collaborator
Loading…
1 task done
[None][perf] Remove spurious sync in sparse fmha forward
#16771
opened Jul 23, 2026 by
brb-nv
Collaborator
Loading…
1 task done
[None][test] Enable session prefetch for all test stages
#16770
opened Jul 23, 2026 by
sunnyqgg
Collaborator
Loading…
3 tasks done
[TRTLLM-14345][feat] Improve the GDN Replay Kernel Under Low Latency
#16768
opened Jul 23, 2026 by
JadoTu
Collaborator
Loading…
1 task done
[NVBUG-6448152][test] TEST ONLY; DO NOT REVIEW Python peer-ready activation validation
#16766
opened Jul 23, 2026 by
chienchunhung
Collaborator
•
Draft
Draft: [None][refactor] Mixed Modality Support for Nemotron Nano Omni V3
#16764
opened Jul 22, 2026 by
aswinvisva
Collaborator
•
Draft
1 task
[https://nvbugs/6198785][fix] Unify phase-1 CUDA graph cleanup
#16763
opened Jul 22, 2026 by
Mgluhovskoi
Contributor
Loading…
[None][fix] SpecDecOneEngineForCausalLM: accept optional hidden_size/vocab_size for composite configs
#16762
opened Jul 22, 2026 by
brnguyen2
Collaborator
Loading…
[None][fix] Fix GPT-OSS router token identity
#16760
opened Jul 22, 2026 by
SimengLiu-nv
Collaborator
Loading…
1 task done
ProTip!
Updated in the last three days: updated:>2026-07-20.