# PyTorch to Onnx export fails when importing tensorRT **URL:** \ **Category:** TensorRT\ **Tags:** tensorrt, pytorch, python\ **Created:** [August 31, 2023, 1:37pm UTC](https://forums.developer.nvidia.com/t/pytorch-to-onnx-export-fails-when-importing-tensorrt/265054 "2023-08-31T13:37:10Z")\ **Posts on this page:** 6\ **Page:** 1
**Author:** ![olt38](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@olt38](https://forums.developer.nvidia.com/u/olt38)\ **Post date:** [August 31, 2023, 1:37pm UTC](https://forums.developer.nvidia.com/t/pytorch-to-onnx-export-fails-when-importing-tensorrt/265054/1 "2023-08-31T13:37:10Z")
Hi, I want to use the tensorRT library in Python to measure the inference time of a PyTorch model. I start by converting the Pytorch model to Onnx, then I build the tensorRT engine using trtexec and finally I measure the network’s inference latency using a custom function made using tensorRT Python API. When tensorRT is imported before torch in the script, everything works fine. However, when torch is imported before tensorRT, the conversion of the PyTorch model to Onnx using torch’s onnx.export() function crashes with a Windows access violation. How can I solve this issue? ## Environment **TensorRT Version** : **GPU Type** : GeForceRTX 3080 Ti Laptop GPU **Nvidia Driver Version** : 522.06 **CUDA Version** : 11.8 **CUDNN Version** : 8.6 **Operating System + Version** : Windows 11 **Python Version (if applicable)**: 3.10.10 **PyTorch Version (if applicable)**: 2.0 ## Relevant Files [reproduce.py](https://forums.developer.nvidia.com/uploads/short-url/1kvJFD2DphsN3iObJd3jDDleyz9.py) (739 Bytes) I attached a Python script reproducing the issue. ## Steps To Reproduce 1. Run the script. 2. When tensorRT is imported before torch, “done” prints and onnx file is generated. 3. When torch is imported before tensorRT, a Windows access violation happens. Error message: Windows fatal exception: access violation Current thread 0x0000305c (most recent call first): File “C:\Users\xxxx\anaconda3\envs\test\lib\site-packages\torch\onnx\utils.py”, line 993 in \_create\_jit\_graph File “C:\Users\xxxx\anaconda3\envs\test\lib\site-packages\torch\onnx\utils.py”, line 1113 in \_model\_to\_graph File “C:\Users\xxxx\anaconda3\envs\test\lib\site-packages\torch\onnx\utils.py”, line 1548 in \_export File “C:\Users\xxxx\anaconda3\envs\test\lib\site-packages\torch\onnx\utils.py”, line 506 in export File “C:\Users\xxxx\Documents\pyTorchAndtensorRT\reproduce.py”, line 24 in ---
**Author:** ![AakankshaS](https://sea2.discourse-cdn.com/nvidia/user_avatar/forums.developer.nvidia.com/aakankshas/32/14047_2.png) [@AakankshaS](https://forums.developer.nvidia.com/u/AakankshaS)\ **Post date:** [August 31, 2023, 2:37pm UTC](https://forums.developer.nvidia.com/t/pytorch-to-onnx-export-fails-when-importing-tensorrt/265054/2 "2023-08-31T14:37:36Z")
Hi, Request you to share the ONNX model and the script if not shared already so that we can assist you better. Alongside you can try few things: > **[Quick Start Guide — NVIDIA TensorRT](https://docs.nvidia.com/deeplearning/tensorrt/latest/getting-started/quick-start-guide.html)** 1. validating your model with the below snippet # check\_model.py import sys import onnx filename = yourONNXmodel model = onnx.load(filename) onnx.checker.check\_model(model). 2) Try running your model with trtexec command. > **[TensorRT/samples/trtexec at master · NVIDIA/TensorRT](https://github.com/NVIDIA/TensorRT/tree/master/samples/trtexec)** > > NVIDIA® TensorRT™ is an SDK for high-performance deep learning inference on NVIDIA GPUs. This repository contains the open source components of TensorRT. - NVIDIA/TensorRT In case you are still facing issue, request you to share the trtexec “”–verbose"" log for further debugging Thanks! ---
**Author:** ![olt38](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@olt38](https://forums.developer.nvidia.com/u/olt38)\ **Post date:** [August 31, 2023, 6:22pm UTC](https://forums.developer.nvidia.com/t/pytorch-to-onnx-export-fails-when-importing-tensorrt/265054/5 "2023-08-31T18:22:48Z")
Hi, The issue arises before I can call trtexec as it happens when converting from PyTorch to Onnx. I believe it’s an issue related to tensorRT because the order in which I import the Python package matters. ---
**Author:** ![spolisetty](https://sea2.discourse-cdn.com/nvidia/user_avatar/forums.developer.nvidia.com/spolisetty/32/14047_2.png) [@spolisetty](https://forums.developer.nvidia.com/u/spolisetty)\ **Post date:** [September 5, 2023, 1:15pm UTC](https://forums.developer.nvidia.com/t/pytorch-to-onnx-export-fails-when-importing-tensorrt/265054/6 "2023-09-05T13:15:18Z")
We are unable to reproduce this issue. We could successfully run the script in both cases (import torch before and after trt). Please check that conda setup has enough permissions and is correct. Thank you. ---
**Author:** ![olt38](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@olt38](https://forums.developer.nvidia.com/u/olt38)\ **Post date:** [September 6, 2023, 7:31pm UTC](https://forums.developer.nvidia.com/t/pytorch-to-onnx-export-fails-when-importing-tensorrt/265054/7 "2023-09-06T19:31:09Z")
Hi, thank you for your feedback! I realized that I forgot to add the TensorRT version of my environment. I am on TensorRT version 8.5.3.1. Can you confirm that you are unable to reproduce this behaviour with this additional information ? Thank you ---
**Author:** ![a2gdb](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@a2gdb](https://forums.developer.nvidia.com/u/a2gdb)\ **Post date:** [December 20, 2023, 8:56am UTC](https://forums.developer.nvidia.com/t/pytorch-to-onnx-export-fails-when-importing-tensorrt/265054/8 "2023-12-20T08:56:16Z")
I faced the same issue with 8.6.1, cuda 12.1, pytorch 2.1.2 (and 2.1.1), win11 (up to date), latest drivers (546.33), latest nvidia software, etc. on my side, my script was silently killed while running torch.onnx.export (no error message) whatever the device (cpu, cuda) and datatype (fp16, float32) Importing tensorrt before pytorch solved my problem. I became mad because I spent too much time before identifying the root cause. Thanks a lot.