NVIDIA/TensorRT

View on GitHub

NVIDIA® TensorRT™ is an SDK for high-performance deep learning inference on NVIDIA GPUs. This repository contains the open source components of TensorRT.

Stars: 13.4k
Forks: 2.4k
Open beginner issues: 0
Indexed issues: 581
Avg merge: 5d 3h
Merged PRs (30d): 2

Dominant language: C++
License: Apache-2.0
Last GitHub push: Aug 25, 2026
Latest indexed: Sep 19, 2026
Contributing guide: Contributing guide
Code of conduct: No code of conduct
Beginner labels: No beginner labels indexed

Beginner friendly Everything

Open All

581 open issues indexed

trtllm-serve accepts `response_format` and never enforces it Open

Difficulty 3/5 1-2 days Newbie friendliness 68/100

NVIDIA/TensorRT#4855 · Opened Sep 17, 2026
Wrong FP16 results when a fused convolution, activation and residual add reads a strided concat view Open

Module:Accuracy

Difficulty 5/5 Over a week Newbie friendliness 45/100

NVIDIA/TensorRT#4854 · Opened Sep 17, 2026
TensorRT silently produces wrong results for batch > 1 on SegFormer (b0/b1/b2 verified, fp32, static and dynamic engines, TRT 10.11 & 11.1, sm89 + sm120) Open

Module:Accuracy

Difficulty 5/5 Over a week Newbie friendliness 28/100

NVIDIA/TensorRT#4853 · Opened Sep 15, 2026
[CMake] Windows package config references missing bin/tensorrt_player.exe (tensorrt_capture_player) Open

Module:OSS Build

Difficulty 3/5 1-2 days Newbie friendliness 68/100

NVIDIA/TensorRT#4851 · Opened Sep 14, 2026
AOT plugins leak CUDA modules on dynamic-shape changes in TensorRT 11.1/11.2.1 Open

Module:Plugins

Difficulty 5/5 Over a week Newbie friendliness 35/100

NVIDIA/TensorRT#4849 · Opened Sep 12, 2026
Question: How much large-scale scaling behavior can be predicted before benchmarking? Open

Difficulty 5/5 Over a week Newbie friendliness 20/100

NVIDIA/TensorRT#4848 · Opened Sep 11, 2026
Incorrect output failure of TensorRT 11.1 and 11.2 when running dynamic If → Transpose → Resize on GPU GTX 1660 SUPER and A10G Open

Module:Accuracy

Difficulty 4/5 3-5 days Newbie friendliness 35/100

NVIDIA/TensorRT#4846 · Opened Sep 6, 2026
TensorRT 10.14+ multi-GPU build regression with Cask-backed ops Open

Module:Engine Build

Difficulty 5/5 Over a week Newbie friendliness 25/100

NVIDIA/TensorRT#4845 · Opened Sep 5, 2026
BF16 strongly-typed engine build SIGSEGVs in Myelin on RTX 5090 (sm_120) for DETR-family transformer — identical graph builds fine as FP16 (reproduced on 11.0.0.114 / 11.1.0.106 / 11.2.1.2) Open

Module:Engine Build

Difficulty 5/5 Over a week Newbie friendliness 25/100

NVIDIA/TensorRT#4842 · Opened Aug 27, 2026
Serialized engine size failure of TensorRT 11.2 when building an unfused pointwise op on a large activation on GPU H100 Open

Module:Engine Build

Difficulty 4/5 3-5 days Newbie friendliness 55/100

NVIDIA/TensorRT#4841 · 1 comment · 1 reaction · Opened Aug 26, 2026
Segmentation fault in TensorRT 10.3 during Myelin compilation of PointCNN-style model on Jetson Orin NX Open

Module:Engine Build

Difficulty 5/5 Over a week Newbie friendliness 30/100

NVIDIA/TensorRT#4840 · Opened Aug 26, 2026
FP16 model compilation failure with TensorRT 11.2.1.2 on 4090 RTX Open

Module:ONNX

Difficulty 4/5 3-5 days Newbie friendliness 48/100

NVIDIA/TensorRT#4838 · 1 comment · Opened Aug 25, 2026
Silently incorrect fp16/bf16 outputs from fused MHA on consumer Blackwell (sm_120) for ViT/DETR model (SAM 3 detector) — fp32 exact, persists in 10.16.1, breaking the MHA fusion restores accuracy Open

Module:Accuracy

Difficulty 5/5 Over a week Newbie friendliness 35/100

NVIDIA/TensorRT#4837 · Opened Aug 24, 2026
W4A16 FP4 build succeeds while silently realizing FP32 (no warning, no refusal) Open

Module:Quantization

Difficulty 4/5 3-5 days Newbie friendliness 48/100

NVIDIA/TensorRT#4833 · 1 comment · Opened Aug 14, 2026
Myelin "Could not infer output types for operation: dequantize" on SM 8.7 (Orin); identical engine builds on SM 11.0 (Thor) with the same TensorRT 10.16.2.10 Open

Module:Quantization

Difficulty 5/5 Over a week Newbie friendliness 35/100

NVIDIA/TensorRT#4832 · 1 comment · Opened Aug 14, 2026
[Architecture Discussion] Overhead and safety of asynchronous soft-stop signaling for KV-cache reclamation during generation loops Open

Feature Request

Difficulty 5/5 Over a week Newbie friendliness 25/100

NVIDIA/TensorRT#4831 · 1 comment · Opened Aug 14, 2026
Myelin NVRTC compilation failure on Jetson Thor (SM110) building Fast-FoundationStereo - reproduces on TRT 10.14.1 and 11.2.1.2 Open

Module:Embedded Module:Engine Build

Difficulty 5/5 Over a week Newbie friendliness 25/100

NVIDIA/TensorRT#4830 · 1 comment · Opened Aug 13, 2026
TensorRT 11.0.0.114 builder finds no valid tactic for BF16 ConvTranspose2d (works fine in FP16) Open

Module:Engine Build

Difficulty 4/5 3-5 days Newbie friendliness 38/100

NVIDIA/TensorRT#4829 · 2 comments · Opened Aug 12, 2026
Deprecate demoDiffusion Open

Module:Demo

Difficulty 2/5 1-3 hours Newbie friendliness 72/100

NVIDIA/TensorRT#4827 · Opened Aug 11, 2026
TRT 11.2: `Transpose` of a network input folded into a fused MHA operand with the permutation applied to the shape but not to the strides Open

Module:Engine Build

Difficulty 5/5 Over a week Newbie friendliness 35/100

NVIDIA/TensorRT#4826 · Opened Aug 10, 2026
Wrong encoder output for NeMo Parakeet (FastConformer) on RTX 5090 (sm_120a) — all precisions, all opt levels, TRT 10.14 and 11.1 Open

Module:Accuracy

Difficulty 5/5 Over a week Newbie friendliness 25/100

NVIDIA/TensorRT#4822 · 1 comment · Opened Jul 23, 2026
AveragePool fails at the exact batch boundary N=65536 on RTX 5090 Open

Module:ONNX

Difficulty 4/5 3-5 days Newbie friendliness 48/100

NVIDIA/TensorRT#4821 · Opened Jul 22, 2026
YOLOv10 Error with TensorRT 10.13 Open

Module:Accuracy

Difficulty 4/5 3-5 days Newbie friendliness 42/100

NVIDIA/TensorRT#4820 · Opened Jul 20, 2026
TRT 11.1 strongly-typed FP32 engine silently produces wrong results for a DETR-style detection model (correct in ONNXRuntime and TRT 10.13) Open

Module:Accuracy

Difficulty 5/5 Over a week Newbie friendliness 30/100

NVIDIA/TensorRT#4813 · 10 comments · Opened Jul 1, 2026
ResizeWithPadPlugin — Aspect-Ratio-Preserving Resize with Padding Open

Feature Request

Difficulty 5/5 Over a week Newbie friendliness 35/100

NVIDIA/TensorRT#4811 · 1 comment · Opened Jun 28, 2026
instanceNormalizationPlugin: use public __half2float instead of private __internal_half2float Open

Module:Plugins

Difficulty 2/5 1-3 hours Newbie friendliness 88/100

NVIDIA/TensorRT#4810 · Opened Jun 25, 2026
Concerns about TensorRT 11 mixed-precision workflow Open

Module:Quantization

Difficulty 5/5 Over a week Newbie friendliness 25/100

NVIDIA/TensorRT#4807 · 1 comment · 3 reactions · Opened Jun 18, 2026
Assertion error in convertExplicitDDSPluginToImplicit.cpp: need help understanding the error. Open

Difficulty 4/5 3-5 days Newbie friendliness 25/100

NVIDIA/TensorRT#4806 · Opened Jun 17, 2026
[Feature Request] TRT_MoE Support for older GPU architectures (sm90, sm89, sm80) Open

Feature Request

Difficulty 5/5 Over a week Newbie friendliness 25/100

NVIDIA/TensorRT#4805 · Opened Jun 12, 2026
How to close part of fusions in TensorRT 10.16 Open

Module:ONNX

Difficulty 4/5 3-5 days Newbie friendliness 35/100

NVIDIA/TensorRT#4804 · 1 comment · Opened Jun 10, 2026
Broken Python Plugin Guide link in quickly_deployable_plugins README Open

Module:Documentation

Difficulty 1/5 Under an hour Newbie friendliness 68/100

NVIDIA/TensorRT#4802 · 1 comment · Opened Jun 10, 2026
Effect of Caching .timing File on Engine Rebuild Open

Module:Engine Build

Difficulty 4/5 3-5 days Newbie friendliness 35/100

NVIDIA/TensorRT#4798 · Opened Jun 2, 2026
INT8 calibration failure of TensorRT 10.3 when running trtexec --int8 on Jetson Orin (CC 8.7) Open

Module:Engine Build Module:Quantization

Difficulty 5/5 Over a week Newbie friendliness 30/100

NVIDIA/TensorRT#4797 · Opened May 28, 2026
TensorRT produces different outputs for identical samples in the same batch on GPU L20 Open

Module:Accuracy

Difficulty 4/5 3-5 days Newbie friendliness 45/100

NVIDIA/TensorRT#4796 · 1 comment · Opened May 25, 2026
TensorRT Lean cannot deserialize engine built in full TensorRT (ReformatRunner error) Open

Module:Runtime

Difficulty 4/5 3-5 days Newbie friendliness 30/100

NVIDIA/TensorRT#4790 · Opened May 19, 2026
Incorrect results of TensorRT 10.16.1 when running with `--best` optimized resnet152 on GPU RTX PRO 6000 Blackwell Open

Module:Accuracy

Difficulty 4/5 3-5 days Newbie friendliness 38/100

NVIDIA/TensorRT#4780 · Opened May 17, 2026
Incorrect Cast result of TensorRT 10.16.1.11 when running ONNX Cast float-to-int32 on GPU Open

Module:ONNX

Difficulty 4/5 3-5 days Newbie friendliness 45/100

NVIDIA/TensorRT#4775 · Opened May 13, 2026
Incorrect Gather out-of-bounds handling of TensorRT 10.16.1.11 when running ONNX Gather on GPU Open

Module:ONNX

Difficulty 4/5 3-5 days Newbie friendliness 48/100

NVIDIA/TensorRT#4774 · Opened May 13, 2026
Incorrect Clip NaN handling of TensorRT 10.16.1.11 when running ONNX Clip on GPU Open

Module:ONNX

Difficulty 3/5 1-2 days Newbie friendliness 48/100

NVIDIA/TensorRT#4773 · Opened May 13, 2026
Incorrect ReduceLogSumExp overflow of TensorRT 10.16.1.11 when running ONNX ReduceLogSumExp on GPU Open

Module:ONNX

Difficulty 4/5 3-5 days Newbie friendliness 48/100

NVIDIA/TensorRT#4772 · 1 comment · Opened May 13, 2026
Incorrect Min NaN handling of TensorRT 10.16.1.11 when running ONNX Min on GPU Open

Module:ONNX

Difficulty 4/5 3-5 days Newbie friendliness 45/100

NVIDIA/TensorRT#4758 · Opened May 10, 2026
Build failure of TensorRT 10.16.1.11 when running ONNX Attention inside If on GPU Open

Module:Engine Build Module:ONNX

Difficulty 4/5 3-5 days Newbie friendliness 38/100

NVIDIA/TensorRT#4757 · Opened May 10, 2026
Incorrect GroupNormalization result of TensorRT 10.16.1.11 when running ONNX GroupNormalization(num_groups=1) on GPU Open

Module:ONNX

Difficulty 4/5 3-5 days Newbie friendliness 68/100

NVIDIA/TensorRT#4756 · Opened May 10, 2026
nan output in fp16 mode of TensorRT 10.16 when running trtexec on GPU RTX 5080 Open

Module:Samples

Difficulty 4/5 3-5 days Newbie friendliness 48/100

NVIDIA/TensorRT#4748 · Opened May 8, 2026
'notifyShape' API provides a -1 dimension on TensorRT 10.16.1 when running Qwen2.5-VL-3B on Jetson Thor Open

Module:Runtime

Difficulty 4/5 3-5 days Newbie friendliness 45/100

NVIDIA/TensorRT#4744 · Opened May 4, 2026
TensorRT 10.16 Myelin crash on Blackwell (sm_120) — Broadcast stride mismatch in SoVITS decoder Open

Module:Engine Build

Difficulty 4/5 3-5 days Newbie friendliness 38/100

NVIDIA/TensorRT#4743 · 1 reaction · Opened May 4, 2026
Wrong build of TensorRT 10.4 running DINOv3 ViT-B/16 on RTX A6000 Open

Module:ONNX

Difficulty 5/5 Over a week Newbie friendliness 25/100

NVIDIA/TensorRT#4742 · Opened May 3, 2026
Accuracy mismatch between ONNX Runtime and TensorRT for SegVit Open

Module:Accuracy Module:ONNX

Difficulty 4/5 3-5 days Newbie friendliness 42/100

NVIDIA/TensorRT#4741 · Opened Apr 28, 2026
TensorRT 10.16: opset-23 Attention op fails inside ONNX If subgraph (myelin "Unnamed Layer* N [ElementWise]_output" error) Open

Module:ONNX

Difficulty 4/5 3-5 days Newbie friendliness 45/100

NVIDIA/TensorRT#4739 · Opened Apr 26, 2026
ONNX parser failure of TensorRT 10.16 when running Attention node on any GPU Open

Module:ONNX

Difficulty 4/5 3-5 days Newbie friendliness 45/100

NVIDIA/TensorRT#4738 · Opened Apr 23, 2026

NVIDIA/TensorRT

Get new issues in your inbox