Search results

[Bug]: weights preloading makes Llama-4-Scout-17B-16E-Instruct to freeze

AutoDeploy/Dashboard<NV> AutoDeploy Backend: Dashboard related<NV> AutoDeploy Backend: Dashboard related

bugSomething isn't workingSomething isn't working

Customized kernels<NV>Specialized/modified CUDA kernels in TRTLLM for LLM ops, beyond standard TRT. Dev & perf.<NV>Specialized/modified CUDA kernels in TRTLLM for LLM ops, beyond standard TRT. Dev & perf.

Status: Open.

#11819 In NVIDIA/TensorRT-LLM;

· tcherckez-nvidia opened on Mar 2, 2026

[Feature]: Compile trtllm-gen FMHA cubins for SM120/SM121 (consumer Blackwell)

Customized kernels<NV>Specialized/modified CUDA kernels in TRTLLM for LLM ops, beyond standard TRT. Dev & perf.<NV>Specialized/modified CUDA kernels in TRTLLM for LLM ops, beyond standard TRT. Dev & perf.

Status: Open.

#11799 In NVIDIA/TensorRT-LLM;

· blake-snc opened on Feb 28, 2026

[Bug]: AutoDeploy: trtllm attention backend produces incorrect output with <code>torch-simple</code>

AutoDeploy<NV> AutoDeploy Backend<NV> AutoDeploy Backend

bugSomething isn't workingSomething isn't working

Customized kernels<NV>Specialized/modified CUDA kernels in TRTLLM for LLM ops, beyond standard TRT. Dev & perf.<NV>Specialized/modified CUDA kernels in TRTLLM for LLM ops, beyond standard TRT. Dev & perf.

Status: Open.

Bug

#11791 In NVIDIA/TensorRT-LLM;

· lucaslie opened on Feb 27, 2026

[Feature]: Add a Model Onboarding Agent

feature requestNew feature or request. This includes new model, dtype, functionality supportNew feature or request. This includes new model, dtype, functionality support

Status: Open.

#11755 In NVIDIA/TensorRT-LLM;

· bmarimuthu-nv opened on Feb 26, 2026

[AutoDeploy]: Add GLM4.7-flash (nvfp4) and Qwen3.5 (nvfp4) to BTK dashboard

AutoDeploy/Dashboard<NV> AutoDeploy Backend: Dashboard related<NV> AutoDeploy Backend: Dashboard related

feature requestNew feature or request. This includes new model, dtype, functionality supportNew feature or request. This includes new model, dtype, functionality support

Status: Open.

#11734 In NVIDIA/TensorRT-LLM;

· galagam opened on Feb 26, 2026

[Feature]: AutoDeploy Fuse gemms in GDN

AutoDeploy<NV> AutoDeploy Backend<NV> AutoDeploy Backend

feature requestNew feature or request. This includes new model, dtype, functionality supportNew feature or request. This includes new model, dtype, functionality support

Status: Open.

Task

#11726 In NVIDIA/TensorRT-LLM;

· taylor-yb-lee opened on Feb 25, 2026

[New Model]: Support Kimi K2.5

new modelRequest to add a new modelRequest to add a new model

Status: Open.

#11709 In NVIDIA/TensorRT-LLM;

· RobertLou opened on Feb 25, 2026

[Feature]: AutoDeploy - Piecewise CG peak memory improvement

AutoDeploy<NV> AutoDeploy Backend<NV> AutoDeploy Backend

feature requestNew feature or request. This includes new model, dtype, functionality supportNew feature or request. This includes new model, dtype, functionality support

Status: Open.

#11694 In NVIDIA/TensorRT-LLM;

· nvchenghaoz opened on Feb 24, 2026

AutoDeploy Dashboard: issue daily report email

AutoDeploy/Dashboard<NV> AutoDeploy Backend: Dashboard related<NV> AutoDeploy Backend: Dashboard related

feature requestNew feature or request. This includes new model, dtype, functionality supportNew feature or request. This includes new model, dtype, functionality support

Status: Open.

#11680 In NVIDIA/TensorRT-LLM;

· galagam opened on Feb 24, 2026

[New Model]: Support Qwen3.5 model in torch backend

new modelRequest to add a new modelRequest to add a new model

Pytorch<NV>Pytorch backend related issues<NV>Pytorch backend related issues

Status: Open.

#11674 In NVIDIA/TensorRT-LLM;

· RobertLou opened on Feb 24, 2026

[Bug]: Incorrect detection order for model_dir type causes IN_MEMORY loading failure

bugSomething isn't workingSomething isn't working

Inference runtime<NV>General operational aspects of TRTLLM execution not in other categories.<NV>General operational aspects of TRTLLM execution not in other categories.

waiting for feedback

Status: Open.

#11666 In NVIDIA/TensorRT-LLM;

· capyun007 opened on Feb 24, 2026

[RFC]: User Telemetry in TensorRT LLM

RFC

Status: Open.

#11659 In NVIDIA/TensorRT-LLM;

· arysef opened on Feb 23, 2026