Uncaught exception detected: Unable to open library: nvinfer_plugin.dll

[deeyonn]

I'm currently following a video guide and I'm getting this error after clicking on Convert ONNX to TensorRT.

Here's a snippet of the output:

[06/06/2023-01:57:53] [I] === Device Information ===
[06/06/2023-01:57:53] [I] Selected Device: NVIDIA GeForce RTX 4090
[06/06/2023-01:57:53] [I] Compute Capability: 8.9
[06/06/2023-01:57:53] [I] SMs: 128
[06/06/2023-01:57:53] [I] Device Global Memory: 24563 MiB
[06/06/2023-01:57:53] [I] Shared Memory per SM: 100 KiB
[06/06/2023-01:57:53] [I] Memory Bus Width: 384 bits (ECC disabled)
[06/06/2023-01:57:53] [I] Application Compute Clock Rate: 2.535 GHz
[06/06/2023-01:57:53] [I] Application Memory Clock Rate: 10.501 GHz
[06/06/2023-01:57:53] [I]
[06/06/2023-01:57:53] [I] Note: The application clock rates do not reflect the actual clock rates that the GPU is currently running at.
[06/06/2023-01:57:53] [I]
[06/06/2023-01:57:53] [I] TensorRT version: 8.6.1
[06/06/2023-01:57:53] [I] Loading standard plugins
[06/06/2023-01:57:56] [E] Uncaught exception detected: Unable to open library: nvinfer_plugin.dll
&&&& FAILED TensorRT.trtexec [TensorRT v8601] # C:\stable-diffusion-webui\extensions\stable-diffusion-webui-tensorrt\TensorRT-8.6.1.6\bin\trtexec.exe --onnx=C:\stable-diffusion-webui\models\Unet-onnx\model15.onnx --saveEngine=C:\stable-diffusion-webui\models\Unet-trt\model15.trt --minShapes=x:2x4x64x64,context:2x77x768,timesteps:2 --maxShapes=x:16x4x96x96,context:16x308x768,timesteps:16 --fp16
*** Error completing request

[deeyonn]

its probably because I'm on windows 11

[sunwukong0727]

i'm on windows 11 too , but ok

[deeyonn]

i'm on windows 11 too , but ok

did u use build tools 2019 or 2022

[sunwukong0727]

i use 2022

[deeyonn]

i use 2022

if you have the time could you share your step by step from a fresh dev branch for me to replicate?

[sunwukong0727]

today i make a vidio but i'm not sure can help you i'm a fresh bird

[ZeroinChina]

I encountered the same problem when converting onnx to tensorrt model. Has anyone solved it?