Unable to convert onnx model to engine w/ trtexec. (Network creation failed)

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

Description

Unable to convert onnx model to engine. I want to be able to be more descriptive with my issue, but truthfully, I have no idea whats going on other than that trtexec.exe failed to create a network.

Environment

TensorRT Version: 8.6.1.6
GPU Type: Mobile GTX 1060 3GB
Nvidia Driver Version: 551.61
CUDA Version: 11.6
CUDNN Version: 8.9.7
Operating System + Version: Windows 11

Relevant Files

model.onnx (373.9 MB)
Alternative drive link: model.onnx - Google Drive

Steps To Reproduce

  1. Put model.onnx in the same directory as trtexec.exe. (I have a cudart64_110.dll file in there as well)

  2. Open directory in command prompt

  3. Enter: ./trtexec.exe --onnx=model.onnx --saveEngine=model.engine --fp16 (doesn’t work with or without the --fp16 flag)

  4. Command prompt output:
    onnxtoenginelog.txt (5.2 KB)

2

I’ve figured out the problem.
It was either my cuda cudnn or tensorrt installation. I reinstalled everything and it works now.

3

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.