Segmentation fault (core dumped) while trying to convert to TensorRT on Jetson Nano

[cvpfus]

Hi, I'm trying to export my custom model to TensorRT on Jetson Nano, but when I tried this command:

python3 tools/trt.py -f ~/project/YOLOX/exps/default/yolox_s.py -c best_ckpt.pth.tar

It throws an error:
Segmentation fault (core dumped)

[GOATmessi8]

Could you provide more information about the error? like the tensorrt log

[cvpfus]

Could you provide more information about the error? like the tensorrt log

After I ran trt.py it only shows the segmentation fault error

[cvpfus]

How to add tensorrt log?

[zxy-zzz]

How to add tensorrt log?

in trt.py change max_workspace_size=(1 << 32) into max_workspace_size=(1 << 28), or smaller, in order to reduce memory usage.

[cvpfus]

How to add tensorrt log?

in trt.py change max_workspace_size=(1 << 32) into max_workspace_size=(1 << 28), or smaller, in order to reduce memory usage.

Thank you. Reduced the max workspace size and it works.

[hzlijianjun]

More matched issue is:
[TensorRT] ERROR: ../rtSafe/safeRuntime.cpp (25) - Cuda Error in allocate: 2 (out of memory)
[TensorRT] ERROR: ../rtSafe/safeRuntime.cpp (25) - Cuda Error in allocate: 2 (out of memory)

[tulbureandreit]

@yusufpraditya could you provide more details please?