Trtexec error converting onnx

Description

running trtexec on onnx model reults in the following error:
log_info.txt (5.1 KB)
log_warnings_errors.txt (19.8 KB)

A c[07/31/2022-16:42:26] [W] [TRT] Skipping tactic 0 due to Myelin error: myelinTargetSetPropertyMemorySize called with invalid memory size (0).
[07/31/2022-16:42:26] [E] Error[10]: [optimizer.cpp::nvinfer1::builder::`anonymous-namespace’::LeafCNode::computeCosts::1981] Error Code 10: Internal Error (Could not find any implementation for node {ForeignNode[102…Transpose_2732 + Reshape_2738]}.)
[07/31/2022-16:42:26] [E] Error[2]: [builder.cpp::nvinfer1::builder::Builder::buildSerializedNetwork::561] Error Code 2: Internal Error (Assertion enginePtr != nullptr failed. )lear and concise description of the bug or issue.

Environment

TensorRT Version: 8.4.1.5
GPU Type: RTX A3000
Nvidia Driver Version: R511.65 (r511_37-13) / 30.0.15.1165 (1-28-2022)
CUDA Version: 11.6
CUDNN Version: 8.4.1
Operating System + Version: win10
Python Version (if applicable): 3.9
TensorFlow Version (if applicable): 8.4.1.5
PyTorch Version (if applicable): 1.12.0
model.onnx (91.2 MB)

Baremetal or Container (if container which image + tag):

Relevant Files

Please attach or include links to any models, data, files, or scripts necessary to reproduce your issue. (Github repo, Google Drive, Dropbox, etc.)

Steps To Reproduce

Please include:

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

Hi,

Could you please try increasing the workspace using trtexec --workspace option.
Above error looks like due to insufficient workspace. We could build the engine successfully by increasing the workspace.

Thank you.

Thanks,
I have increased the workspase to 16384.
same error.

Boris

Could you please share with us complete logs with the --verbose option.

the complete log with --verbose option was attached to the original post.
please check log_info.txt and log_wanings_errors.txt

The logs you’ve shared don’t contain command output with the --verbose option.