Tlt-convert for custom trained YoloV4 model failed on Jetson Nano 4G

[1733208392]

TensorRT cannot parse some ONNX models transferred from Paddle

Hi， I have checked the ONNX file has no problems, we use same way to transfer Paddle models to ONNX models, and the all of the ONNX files can be read by some other tools like Netron. Do you have some method to check whether the ONNX file match TensorRT Parser format?

Is it because of the protobuf version mismatch? How to check the protobuf version in my env?