Description
After run the tensorrt model for some minutes, the error appears:
"
[TensorRT] ERROR: …/rtExt/cuda/cudaGatherRunner.cpp (111) - Cuda Error in execute: 700 (an illegal memory access was encountered)
[TensorRT] ERROR: FAILED_EXECUTION: std::exception
"
Environment
NGC 20.10
Hi @593904440,
Can you please try running your model with trtexec command, and share the verbose logs with us?
e.g.,
trtexec --onnx=your_model.onnx --verbose --explicitBatch --shapes=name:1x3x1216x800
Thanks!
Sry, I figured out that, I used the trt7 to transfer onnx->trt, but inference with execute_async, after i changed the model in dynamic and inference with execute_async_v2, the error was gone.
Best wishes!
1 Like