- Allows to inject Tensorflow Models Zoo into Gstreamer Pipeline in Python
- COCO Labels
python3 -m venv venv
source venv/bin/activate
pip install --upgrade wheel pip setuptools
pip install --upgrade --requirement requirements.txt
export GOOGLE_APPLICATION_CREDENTIALS=$PWD/credentials/gs_viewer.json
dvc pull
- Tested on TF-GPU==1.5
pip install tensorflow==1.15
pip install tensorflow-gpu==1.15
./run_example.sh
export GST_PLUGIN_PATH=$GST_PLUGIN_PATH:$PWD/venv/lib/gstreamer-1.0/:$PWD/gst/
gst-launch-1.0 filesrc location=video.mp4 ! decodebin ! videoconvert ! \
video/x-raw,format=RGB ! gst_tf_detection config=data/tf_object_api_cfg.yml ! videoconvert ! gtksink sync=False
- config: path to filename of Config Format
- model: instance of object TFObjectDetectionModel
gst-launch-1.0 filesrc location=video.mp4 ! decodebin ! videoconvert ! \
video/x-raw,format=RGB ! gst_tf_detection config=data/tf_object_api_cfg.yml ! videoconvert ! \
gst_detection_overlay ! videoconvert ! gtksink sync=False
-
python convert_labels_pbtxt_to_yml.py -f mscoco_label_map.pbtxt
export TF_CPP_MIN_LOG_LEVEL={0,1,2,3,4,5 ...}
export GST_DEBUG=python:{0,1,2,3,4,5 ...}
export GST_PYTHON_LOG_LEVEL={0,1,2,3,4,5 ...}
MIT License