Open korkland opened 3 months ago
for issue that trtexec failed to build engine, it seems your model has ScatterND op which need reduction
attribute, but the latest tensorrt hasn't supported that yet: https://github.com/onnx/onnx-tensorrt/blob/10.2-GA/docs/operators.md
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.
Describe the issue
We are trying to quantize our proprietary model based on RetinaNet using TensorRT's model optimization library. The following warning was raised: "Please consider running pre-processing before quantization." Hoping for performance improvement, I tried running:
but faced the following error:
Here is our ONNX implementation for NMS:
Next, I tried running it without symbolic shape inference:
and it passed, allowing me to quantize the model (using '_quantizestatic'):
still getting the warning..
but when im trying to build the trt engine im getting
any help will be appreciated. Thanks
To reproduce
detailed in the description
Urgency
Blocking me from quantizing model as per ONNX recommendations
Platform
Linux
OS Version
Ubuntu 22.04
ONNX Runtime Installation
Released Package
ONNX Runtime Version or Commit ID
onnxruntime-gpu 1.18.1
ONNX Runtime API
Python
Architecture
X64
Execution Provider
CUDA
Execution Provider Library Version
CUDA 11.8