• Hardware Platform
GPU
• DeepStream Version
7.0
• TensorRT Version
provider by nvcr.io/nvidia/deepstream:7.0-gc-triton-devel
**• NVIDIA GPU Driver Version **
535.183.01
• Issue Type
bug
• How to reproduce the issue ?
Dockerfile:
FROM nvcr.io/nvidia/deepstream:7.0-gc-triton-devel
RUN chmod ag+rw /opt/nvidia/deepstream/deepstream/sources
RUN apt install -y apt-transport-https ca-certificates \
python3-gi python3-dev python3-gst-1.0 python-gi-dev git meson \
python3 python3-pip python3.10-dev cmake g++ build-essential libglib2.0-dev \
libglib2.0-dev-bin libgstreamer1.0-dev libtool m4 autoconf automake libgirepository1.0-dev libcairo2-dev
RUN update-ca-certificates
RUN cd /opt/nvidia/deepstream/deepstream/sources
RUN wget https://github.com/NVIDIA-AI-IOT/deepstream_python_apps/releases/download/v1.1.11/pyds-1.1.11-py3-none-linux_x86_64.whl && \
pip3 install ./pyds-1.1.*.whl && \
rm ./pyds-1.1.*
Build docker with:
docker build -t pyds .
Run docker with:
docker container run -it --gpus=all pyds bash
Inside docker, run following python script:
import gi
gi.require_version("Gst", "1.0")
from gi.repository import GLib, Gst
import pyds
def main():
Gst.init(None)
pipe_str = "videotestsrc name=videotestsrc1 ! fakesink"
pipe = Gst.parse_launch(pipe_str)
child = pipe.get_by_name("videotestsrc1")
if __name__ == "__main__":
main()
Program crashe with:
Fatal Python error: _PyGILState_NoteThreadState: Couldn't create autoTSSkey mapping
Python runtime state: initialized
Thread 0x00007ff26ad604c0 (most recent call first):
File "/opt/nvidia/deepstream/deepstream-7.0/main.py", line 16 in main
File "/opt/nvidia/deepstream/deepstream-7.0/main.py", line 20 in <module>
Extension modules: gi._gi, gi.overrides._gi_gst (total: 2)
Aborted (core dumped)
With
pipe_str = "nvinfer name=primary-inference ! fakesink"
program freezes on
child = pipe.get_by_name("primary-inference")
After deleting ‘pyds’ import, both versions works good.