Issues

• [ALERT] Access to NVIDIA's Self-Hosted Runners is Expiring

  nvidia-runners
  NVIDIA/TensorRT-Incubator#845 · Filter by author nvidia-runner-mgmt opened on Oct 6, 2026

• cuda-python resolves to 13.0.0 (PyPI) or 12.9.0 (download.pytorch.org/whl/cpu) depending on install path (pip install .` with the index configuration committed in the repo)

  NVIDIA/TensorRT-Incubator#844 · Filter by author cuiliaomei-beep opened on Sep 12, 2026

• Conv Fails in Float16 but Passes in Float32

  Bug

  NVIDIA/TensorRT-Incubator#694 · Filter by author Mgluhovskoi opened on Aug 14, 2025

• The header trt_plugin_python.h does not seem to be included with TensorRT or with this repo

  NVIDIA/TensorRT-Incubator#658 · Filter by author matthewfl opened on Jun 16, 2025
  3

• Add infrastructure for serialization tests

  tripy
  Task

  NVIDIA/TensorRT-Incubator#653 · Filter by author pranavm-nvidia opened on Jun 9, 2025

• Decide if fp8/int4 are supported in operators

  tripy
  Enhancement

  NVIDIA/TensorRT-Incubator#642 · Filter by author yizhuoz004 opened on May 21, 2025

• Improve performance profiling experience

  tripy
  Enhancement

  NVIDIA/TensorRT-Incubator#640 · Filter by author pranavm-nvidia opened on May 20, 2025
  1

• tensorrt.host_tensor and tensorrt.shape_profile are not recognized correctly

  mlir-tensorrt
  Bug

  NVIDIA/TensorRT-Incubator#630 · Filter by author yizhuoz004 opened on May 13, 2025
  3

• Add an API to implicitly compile modules?

  tripy
  Enhancement

  NVIDIA/TensorRT-Incubator#629 · Filter by author pranavm-nvidia opened on May 13, 2025
  1

• Add an API to change the datatype of an initialized module

  tripy
  Enhancement

  NVIDIA/TensorRT-Incubator#627 · Filter by author pranavm-nvidia opened on May 13, 2025
  2

• TensorRT doesn't work with Convolution with dynamic batch in some cases

  tensorrt
  Bug

  NVIDIA/TensorRT-Incubator#617 · Filter by author pranavm-nvidia opened on Apr 23, 2025
  1

• Preserve argument names in MLIR function

  tripy
  Enhancement

  NVIDIA/TensorRT-Incubator#609 · Filter by author pranavm-nvidia opened on Apr 10, 2025

• Optimize eager mode evaluation for branchy graphs

  tripy
  Enhancement

  NVIDIA/TensorRT-Incubator#593 · Filter by author pranavm-nvidia opened on Mar 31, 2025
  

• Cannot use tensorrt.broadcast with int8 inputs

  tripy
  Bug

  NVIDIA/TensorRT-Incubator#580 · Filter by author pranavm-nvidia opened on Mar 18, 2025

• quantize to int4 does not work with constants

  tripy
  Bug

  NVIDIA/TensorRT-Incubator#579 · Filter by author pranavm-nvidia opened on Mar 18, 2025

• Add support for multiple devices

  Enhancement

  NVIDIA/TensorRT-Incubator#577 · Filter by author pranavm-nvidia opened on Mar 14, 2025

• allclose seems to incorrectly report mismatches

  Bug

  NVIDIA/TensorRT-Incubator#571 · Filter by author pranavm-nvidia opened on Mar 12, 2025
  1

• Enable tensorrt.assertion in Tripy

  Enhancement

  NVIDIA/TensorRT-Incubator#570 · Filter by author pranavm-nvidia opened on Mar 12, 2025
  1

• Some Conv tests failing with BF16

  Bug

  NVIDIA/TensorRT-Incubator#569 · Filter by author pranavm-nvidia opened on Mar 12, 2025

• ConvTranspose does not work with BF16

  Bug

  NVIDIA/TensorRT-Incubator#565 · Filter by author pranavm-nvidia opened on Mar 11, 2025

• [mlir-tensorrt] Enhancement: Support splat dense attributes for TRT plugins

  mlir-tensorrt
  NVIDIA/TensorRT-Incubator#491 · Filter by author samurdhikaru opened on Feb 7, 2025

• MHA kernels not triggered in nanoGPT sample

  mlir-tensorrtperformance
  Bug

  NVIDIA/TensorRT-Incubator#464 · Filter by author yizhuoz004 opened on Dec 18, 2024
  2

  

• Improve behavior of evaluation during compile

  Enhancement

  NVIDIA/TensorRT-Incubator#409 · Filter by author pranavm-nvidia opened on Nov 26, 2024
  2

• Stride validation fails for empty tensor input

  mlir-tensorrt
  Bug

  NVIDIA/TensorRT-Incubator#398 · Filter by author parthchadha opened on Nov 21, 2024
  2

  

• Incorrect quantize output with fp16 scales

  mlir-tensorrt
  Bug

  NVIDIA/TensorRT-Incubator#392 · Filter by author parthchadha opened on Nov 19, 2024
  1