"DeprecationWarning: This process is multi-threaded, use of fork() may lead to deadlocks in the child." since OpenVINO 2025.0.0

[Dobiasd]

Calling openvino.convert_model results in having multiple lines of the following warning being printed:

/usr/local/lib/python3.12/multiprocessing/popen_fork.py:66: DeprecationWarning: This process (pid=8) is multi-threaded, use of fork() may lead to deadlocks in the child.
#13 2.534   self.pid = os.fork()

This started with OpenVINO version 2025.0.0 and was not yet the case with version 2024.6.0.

I'm unsure if this is a problem, but I thought at least reporting it might be a good idea.

Here is a minimal example to reproduce the warning (and the difference between the two versions):

FROM python:3.12.7

RUN apt-get update && apt-get install -y libgl1-mesa-glx

RUN pip install --no-cache-dir --upgrade numpy==1.26.4 tensorflow==2.18.0

RUN echo 'import keras\n\
import openvino as ov\n\
base = keras.applications.ResNet50(input_shape=(224, 224, 3), weights="imagenet", include_top=True)\n\
image_input = keras.layers.Input(shape=(224, 224, 3), name="input_layer")\n\
ov_model = ov.convert_model(keras.Model(inputs=image_input, outputs=base(image_input), name="image_model"))' >> main.py

ADD "https://www.random.org/cgi-bin/randbyte?nbytes=10&format=h" skipcache

RUN pip install --no-cache-dir openvino==2024.6.0
RUN python main.py

RUN pip install --no-cache-dir openvino==2025.0.0
RUN python main.py

Command:

docker build --rm --progress=plain .

[mlukasze]

we are aware of the problem and solution is under development
#28905

[Dobiasd]

Ah, great. Thanks a lot! 👍