Open c1aude opened 1 month ago
Hi @c1aude could you also share the model/the requirements.txt of your python env/test image to repro this issue?
Hi @c1aude could you also share the model/the requirements.txt of your python env/test image to repro this issue?
Hello @yf711 The env and model files used for inference.
image used for inference :
env requirements.txt : https://drive.google.com/file/d/16wj3sa0JFyBOTpPit_2iqBU0trZtITSw/view?usp=sharing
test model: https://drive.google.com/file/d/12GZKzMf5Pq1_qgKiwPF9HOCBwWM_jvsz/view?usp=sharing
This issue has been automatically marked as stale due to inactivity and will be closed in 30 days if no further activity occurs. If further support is needed, please provide an update and/or more details.
adding a note to further debug this issue
Describe the issue
Inference results are outputting abnormally when using YOLOv7 models with TensorRT EP.
We have confirmed that the results are normal when using CPU and CUDA.
The issue was reproducible in versions 1.18.0 to 1.18.1 using TensorRT 10, and did not occur in versions 1.17.3 and earlier using TensorRT 8.6.1.6.
When using TensorRT 10, are there any other actions required when converting pytorch models to onnx as opposed to using TensorRT8?
Tensor RT result:
CPU or CUDA Result:
To reproduce
The code we used for testing is shown below.
Urgency
No response
Platform
Windows
OS Version
Windows 11
ONNX Runtime Installation
Built from Source
ONNX Runtime Version or Commit ID
1.18.1
ONNX Runtime API
Python
Architecture
X64
Execution Provider
TensorRT
Execution Provider Library Version
CUDA 11.8, Cudnn 8.9.7, TensorRT 10.2.0.19