Search results

[Bug] Native Jetson Orin (sm_87) build: explicit -DCMAKE_CUDA_ARCHITECTURES ignored + CuTe DSL --wrap/cudart shim not propagated to static-lib consumers

Status: Open.

#117 In NVIDIA/TensorRT-Edge-LLM;

· suharvest opened on Jun 28, 2026

stbir_resize in resizeImage maybe calc wrong result

bugSomething isn't workingSomething isn't working

Status: Open.

#116 In NVIDIA/TensorRT-Edge-LLM;

· lix19937 opened on Jun 26, 2026

Perf: Qwen3-4B-Instruct-2507 NVFP4 BS=1 prefill/decode throughput ~25–33% below published benchmarks on Jetson AGX Thor

bugSomething isn't workingSomething isn't working

Status: Open.

#115 In NVIDIA/TensorRT-Edge-LLM;

· satabios opened on Jun 23, 2026

Processing of frames that are already on device, in memory

New Feature Request

Status: Open.

Enhancement

#114 In NVIDIA/TensorRT-Edge-LLM;

· joost-de-vries opened on Jun 21, 2026

enable_thinking=True produces reasoning text without literal <think>...</think> tags — _ThinkingStateMachine cannot separate it from the final answer

Status: Open.

#113 In NVIDIA/TensorRT-Edge-LLM;

· micDKpara opened on Jun 19, 2026

[qwen3_5_moe] GPTQ-Int4 MoE: export crash on shared_expert_gate + garbage inference on Jetson Orin

bugSomething isn't workingSomething isn't working

Status: Open.

#110 In NVIDIA/TensorRT-Edge-LLM;

· hanwhale11 opened on Jun 18, 2026

CustomVoice language conditioning bug confirmed in v0.8.0 — working fix applied (ref #87 / PR #98)

Status: Open.

#109 In NVIDIA/TensorRT-Edge-LLM;

· robertvandervoort opened on Jun 17, 2026

Support for nemotron ASR

New Model

Status: Open.

Enhancement

#108 In NVIDIA/TensorRT-Edge-LLM;

· WilliamTambellini opened on Jun 15, 2026

Missing model_config argument in VLM visual ONNX export

Status: Open.

#107 In NVIDIA/TensorRT-Edge-LLM;

· StefanoS90 opened on Jun 12, 2026

Garbage output from Quick Start Guide example with release 0.8.0

bugSomething isn't workingSomething isn't working

Status: Open.

#105 In NVIDIA/TensorRT-Edge-LLM;

· StefanoS90 opened on Jun 10, 2026

[Bug] Qwen3.5-4B-AWQ uses ~13GB RAM on Orin NX 16GB — GDN mamba workspace pool prevents ASR+TTS+LLM co-residency

Status: Open.

#99 In NVIDIA/TensorRT-Edge-LLM;

· suharvest opened on May 27, 2026

Does the current implementation support prefix/encoder cache?

Status: Open.

#94 In NVIDIA/TensorRT-Edge-LLM;

· Syd-Q opened on May 25, 2026