Skip to main content

Onnxruntime

Segmentation fault (core dumped)​

➜ python3 test.py
Traceback (most recent call last):
  File "test.py", line 32, in <module>
    yolov5 = YOLOv5(model_path=model, drawing=True, save_image=None, show_kpts=False)
  File "~/onnxruntime/inference.py", line 89, in __init__
    self.ort_inference = ort.InferenceSession(model_path)
  File "/usr/local/lib/python3.8/dist-packages/onnxruntime/capi/onnxruntime_inference_collection.py", line 283, in __init__
    self._create_inference_session(providers, provider_options, disabled_optimizers)
  File "/usr/local/lib/python3.8/dist-packages/onnxruntime/capi/onnxruntime_inference_collection.py", line 310, in _create_inference_session
[1]    2428621 segmentation fault (core dumped)  python3 test.py

Already found the issue that it caused by docker-compose if we set the CPU cores limitation.

It can be fixed by adding these lines.

Solution​

options = ort.SessionOptions()
options.intra_op_num_threads = 1

sess = ort.InferenceSession(model, options)

Output too many messages in the beginning​

Onnxruntime will output a lot of messages in the beginning when we start it.

Solution​

Add this line after import onnxruntime.

import onnxruntime as ort
ort.set_default_logger_severity(3)