../rtSafe/cuda/cudaConvolutionRunner.cpp (483) - Cudnn Error in executeConv: 3 (CUDNN_STATUS_BAD_PARAM)

AI & Data Science Deep Learning (Training & Inference) cuDNN
1

Description

When the trt-model was in inference，I got the error like:
[E] [TRT] …/rtSafe/cuda/cudaConvolutionRunner.cpp (483) - Cudnn Error in executeConv: 3 (CUDNN_STATUS_BAD_PARAM)
[E] [TRT] FAILED_EXECUTION: std::exception
[E] [TRT] engine.cpp (1036) - Cuda Error in executeInternal: 700 (an illegal memory access was encountered)

Environment

TensorRT Version : 7.2.2.3
GPU Type : TITAN V
Nvidia Driver Version : 440.33.01
CUDA Version : 10.2
CUDNN Version : 8.0.2
Operating System + Version : ubuntu 16.04

2 Likes
2

Hi @654818923 ,
I see the error you have shared can be caused by the date type of the input data.
However to debug it further, we may request the verbose logs from you.

Thanks!