Pull requests: NVIDIA/TensorRT-LLM
Pull requests list
[TRTLLM-11272][fix] Account for the existing multimodal placeholder tokens in a text prompt
#12827
opened Apr 8, 2026 by
moraxu
Loading…
1 task done
[https://nvbugs/6055847][fix] Preserve Nemotron HF mamba cache dtype in bench tuning
#12826
opened Apr 8, 2026 by
hyukn
Loading…
1 task
[None][fix] Fix DeepSeekV32 test_fp8_blockscale[baseline_mtp1] OOM on Blackwell
#12823
opened Apr 8, 2026 by
sunnyqgg
Loading…
[None][fix] Add bounded timeout to gen-side KV transfer in C++ CacheTransceiver
Community want to contribute
PRs initiated from Community
[https://nvbugs/6025177][fix] Fix KV cache issue (cherry-pick to release/1.3.0rc5.post2)
#12819
opened Apr 7, 2026 by
thorjohnsen
Loading…
3 tasks done
[https://nvbugs/5448464][fix] Partially fix LoRA overallocation for Nemotron NAS
#12817
opened Apr 7, 2026 by
brb-nv
Loading…
1 task done
[TRTLLM-10939][feat] Enable block reuse with overlap scheduler
#12816
opened Apr 7, 2026 by
chienchunhung
Loading…
1 task done
[https://nvbugs/5658258][fix] Fix OOM with large number of LoRA adapters
#12815
opened Apr 7, 2026 by
brb-nv
Loading…
1 task done
[None][feat] dual-pool KV cache with SWA block eviction for gemma4
#12813
opened Apr 7, 2026 by
suyoggupta
Loading…
6 tasks done
[None][feat] Upgrade xgrammar and lock pillow
#12812
opened Apr 7, 2026 by
yuanjingx87
Loading…
1 task done
[None][feat] AutoDeploy: Gemma4 vision support
#12810
opened Apr 7, 2026 by
bmarimuthu-nv
•
Draft
1 task
[TRTLLM-11804][feat] Mechanical refactoring VisualGen API
VisualGen
#12807
opened Apr 7, 2026 by
zhenhuaw-me
Loading…
1 task done
[https://nvbugs/6018647][test] Add unit test for Lifecycle Race Condition error in disagg sever
#12803
opened Apr 7, 2026 by
yingguo-trt
Loading…
1 task done
[None][perf] Add GreenContext SM-partitioned overlap for MoE DenseGEMM FC1+Router
#12802
opened Apr 7, 2026 by
JacobHu-NV
•
Draft
[None][feat] Add llm.encode() fast path for encoder-only models
Community want to contribute
PRs initiated from Community
[TRTLLM-11797][feat] Add cutedsl moe backend supporting for qwen3.5.
#12799
opened Apr 7, 2026 by
nv-guomingz
Loading…
1 task done
[None][feat]: Add test_moe_semantics.py to help agent understand the …
#12797
opened Apr 7, 2026 by
WeiHaocheng
Loading…
1 task
[None][test] add unit test and e2e test for gpt_oss_20b MHA kernel
#12796
opened Apr 7, 2026 by
ruodil
Loading…
1 task done
[https://nvbugs/5945047][fix] Fix Eagle3 one-model hang on SM120 via extend_ctx
#12795
opened Apr 7, 2026 by
ziyixiong-nv
Loading…
1 task done
[TRTLLM-11228][feat] Support DFlash in one-model spec dec
#12794
opened Apr 7, 2026 by
ziyixiong-nv
•
Draft
1 task
ProTip!
no:milestone will show everything without a milestone.