TensorRT 8.2 convert art model for onnx failed

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

Description

I get the following error when trying to run trtexec with my onnx model and can not get a try model on GPU L20. But it works well on T40 GPU.

[10/22/2024-17:19:26] [W] [TRT] Skipping tactic 0 due to Myelin error: myelinTargetSetPropertySm called with invalid sm
[10/22/2024-17:19:26] [E] Error[10]: [optimizer.cpp::computeCosts::2011] Error Code 10: Internal Error (Could not find any implementation for node {ForeignNode[Func/StatefulPartitionedCall/input/_28:0…StatefulPartitionedCall/Reshape_4 + StatefulPartitionedCall/conv1a/Conv1D/ExpandDims + StatefulPartitionedCall/conv1a/Conv1D__236]}.)

Environment

TensorRT Version: 8.2.5.1
GPU Type: NVIDIA L20
Nvidia Driver Version: 535.183.01
CUDA Version: 12.2
CUDNN Version:
Operating System + Version: CentOS Linux 7 (Core)
Python Version (if applicable):
TensorFlow Version (if applicable):
PyTorch Version (if applicable):
Baremetal or Container (if container which image + tag):

Steps To Reproduce

/usr/local/TensorRT/bin/trtexec --onnx=20241022105939451.onnx --saveEngine=20241022105939451.engine --workspace=1024

Please include:

  • Exact steps/commands to build your repro
  • Exact steps/commands to run your repro
  • Full traceback of errors encountered
5

Hi @baiyuxin ,
Can you pls try upgrading the trt , the issue should be resolved there.

Thanks