Search results

[Feature]: AutoDeploy: Nemotron nano-v3 fp8 functional support

AutoDeploy<NV> AutoDeploy Backend<NV> AutoDeploy Backend

Status: Open.

Task

#8512 In NVIDIA/TensorRT-LLM;

· lucaslie opened on Oct 20, 2025

[AutoDeploy]: Dynamically select a tile size for <code>fused_mlp_moe_kernel</code>

AutoDeploy<NV> AutoDeploy Backend<NV> AutoDeploy Backend

Status: Open.

#8511 In NVIDIA/TensorRT-LLM;

· nzmora-nvidia opened on Oct 20, 2025

[Feature]: Structured outputs: JDecoding With a Compressed Finite State Machine.

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

Speculative Decoding<NV>MTP/Eagle/Medusa/Lookahead/Prompt-Lookup-Decoding/Draft-Target-Model/ReDrafter<NV>MTP/Eagle/Medusa/Lookahead/Prompt-Lookup-Decoding/Draft-Target-Model/ReDrafter

Status: Open.

#8504 In NVIDIA/TensorRT-LLM;

· july8023 opened on Oct 20, 2025

[Usage]: XQA Kernel License Clarification

Doc<NV>TRTLLM's textual/illustrative materials: API refs, guides, tutorials. Improvement & clarity.<NV>TRTLLM's textual/illustrative materials: API refs, guides, tutorials. Improvement & clarity.

questionFurther information is requestedFurther information is requested

Status: Open.

#8476 In NVIDIA/TensorRT-LLM;

· aikitoria opened on Oct 19, 2025

[Bug]: Can't run GPT-OSS models on DGX Spark

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.

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

Status: Open.

#8474 In NVIDIA/TensorRT-LLM;

· kalelegs opened on Oct 18, 2025

[Feature]: AutoDeploy: graph visualization tooling

AutoDeploy<NV> AutoDeploy Backend<NV> AutoDeploy Backend

Community Engagementhelp/insights needed from communityhelp/insights needed from community

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

#8460 In NVIDIA/TensorRT-LLM;

· lucaslie opened on Oct 17, 2025

IGNORE PLEASE. testing my pipeline

nvbug_testJust for testingJust for testing

Status: Open.

#8441 In NVIDIA/TensorRT-LLM;

· karljang opened on Oct 16, 2025

[AutoDeploy]: Nemotron-6 8K/16K (ISL/OSL) on B200; BF16

AutoDeploy<NV> AutoDeploy Backend<NV> AutoDeploy Backend

Status: Open.

#8436 In NVIDIA/TensorRT-LLM;

· nzmora-nvidia opened on Oct 16, 2025

[AutoDeploy]: Nemotron-6 8K/16K (ISL/OSL) on H100; BF16

AutoDeploy<NV> AutoDeploy Backend<NV> AutoDeploy Backend

Status: Open.

#8435 In NVIDIA/TensorRT-LLM;

· nzmora-nvidia opened on Oct 16, 2025

[Feature]: Enable MLA prefill-only operator

AutoDeploy<NV> AutoDeploy Backend<NV> AutoDeploy Backend

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.

#8424 In NVIDIA/TensorRT-LLM;

· nzmora-nvidia opened on Oct 16, 2025

[Feature]: support for nvidia/Qwen2.5-VL-7B-Instruct-FP4

questionFurther information is requestedFurther information is requested

waiting for feedback

Status: Open.

#8404 In NVIDIA/TensorRT-LLM;

· pythonjavaerlang opened on Oct 15, 2025

[AutoDeploy]: Build the connection between BTK and TRTLLM-Bench (with AD as the backend)

AutoDeploy<NV> AutoDeploy Backend<NV> AutoDeploy Backend

Status: Open.

#8393 In NVIDIA/TensorRT-LLM;

· nzmora-nvidia opened on Oct 15, 2025