Segmentation Fault errors and some other (Deserialize the cuda engine failed)

Description

Hello, I changed the config and everything else according to instructions from linghu8812/tensorrt_inference (github.com). Next, I ran commands for yolov7 and errors started popping up at startup.

If you run without sudo, Segmentation Fault (core dumped) pops up.

And if you run with sudo, then another error pops up.

I attached a photo to demonstrate the errors.

I would be grateful for help with solving the problem!

I use a custom yolov7 model with its own weights.

Environment

TensorRT Version: 7.2.2.3
GPU Type: NVIDIA RTX 2070
CUDA Version: 11.1
Operating System + Version: Ubuntu 20.04

Hi,

It looks like you’re using a very old version of TensorRT. We highly recommend that you try the latest TensorRT version 8.6 and let us know if you still face the same issue.

Thank you.

1 Like