Segmentation fault while building an engine from an ONNX model

AI & Data Science Deep Learning (Training & Inference) TensorRT
1

Description

When using trtexec to convert onnx model to tensorrt engine, I encountered seg-fault.

Environment

TensorRT Version: 10.0.1
GPU Type: A100
Nvidia Driver Version: 535.104.12
CUDA Version: 12.2
CUDNN Version:
Operating System + Version: Ubuntu22.04
Python Version (if applicable): python3.10
TensorFlow Version (if applicable):
PyTorch Version (if applicable):
Baremetal or Container (if container which image + tag): doing every thing in a docker container which is based on nvcr.io/nvidia/cuda:12.2.2-runtime-ubuntu22.04

Relevant Files

Steps To Reproduce

  • I use this command to build tensorrt engine: trtexec --onnx=model.onnx --saveEngine=model.trt --verbose
  • trtexec is manually compiled from: /usr/src/tensorrt/samples/trtexec
2

Besides, my onnx model passed check_model:

In [4]: import onnx
   ...: import tensorrt as trt
   ...:
   ...: onnx_model = onnx.load("model.onnx")
   ...: model = onnx.checker.check_model(onnx_model, full_check=True)

In [5]:
3

Hi @runqiu.bao
Trying a repro for this.

Shall update you.