search

open-mmlab / mmdeploy

OpenMMLab Model Deployment Framework
https://mmdeploy.readthedocs.io/en/latest/
Apache License 2.0
2.74k stars 629 forks source link

TRTNet: 1: [graphContext.h::MyelinGraphContext::24] Error Code 1: Myelin (Compiled against cuDNN 10.2.1.0 but running against cuDNN 11.2.1.0.) #266

Closed vansin closed 2 years ago

vansin commented 2 years ago

When I use mmdeploy_python package to run, follow errors occur

[2022-03-24 00:18:28.732] [mmdeploy] [info] Default CUDA allocator initialized
[2022-03-24 00:18:45.801] [mmdeploy] [warning] TRTNet: TensorRT was linked against cuBLAS/cuBLAS LT 11.6.5 but loaded cuBLAS/cuBLAS LT 11.2.1
[2022-03-24 00:18:49.157] [mmdeploy] [warning] TRTNet: TensorRT was linked against cuBLAS/cuBLAS LT 11.6.5 but loaded cuBLAS/cuBLAS LT 11.2.1
[2022-03-24 00:18:49.177] [mmdeploy] [error] TRTNet: 1: [graphContext.h::MyelinGraphContext::24] Error Code 1: Myelin (Compiled against cuDNN 10.2.1.0 but running against cuDNN 11.2.1.0.)
[2022-03-24 00:18:49.182] [mmdeploy] [error] TRTNet: 1: [graphContext.h::MyelinGraphContext::24] Error Code 1: Myelin (Compiled against cuDNN 10.2.1.0 but running against cuDNN 11.2.1.0.)
lzhangzz commented 2 years ago

This is most likely caused by version mismatch between CUDA, cuDNN and TensorRT.

lvhan028 commented 2 years ago

A similar issue: https://github.com/NVIDIA/TensorRT/issues/317

Cushgx commented 7 months ago

I met the same error, have you solved it yet? thx