Please provide complete information as applicable to your setup.
• Hardware Platform (Jetson / GPU) Jetson xavier NX• DeepStream Version 6.1.1• JetPack Version (valid for Jetson only) 5.0.2• TensorRT Version 8.4• Issue Type( questions, new requirements, bugs) questions
I have a pipeline like this
I am trying to access the tensormeta of the sgie which gives binary classiification results. when i use the above pipeline means pgie->tracker->sgie i am not getting any results of sgie (few rarely) but when i use pgie->sgie->tracker i am getting classification results from sgie.
i wonder what is causing the issue. if i use sgie before tracker i am not able to use appsink.
kesong
October 26, 2023, 10:18am
3
Can you share your config file of PGIE/SGIE/tracker? Please check if you set “classifier-async-mode=1” in SGIE config file: Gst-nvinfer — DeepStream 6.3 Release documentation
I have commented this line classifier-async-mode=1
pgie_config.txt (3.1 KB)sgie_config.txt (3.0 KB)
I think i am not getting any sgie tensormeta data in appsink prob
@kesong any update on this.
sorry i have shared the other pipeline, please find the pipeline which i am using appsink.
“gst-launch-1.0 filesrc location=/home/sensormatic/Videos/pvh-site-15min.mp4 ! decodebin ! nvvideoconvert ! m.sink_0 nvstreammux name=m width=1280 height=720 batch-size=1 batched-push-timeout=40000 ! queue ! nvinfer config-file-path=/home/sensormatic/vision-works-ds/cvads/configurations/occupancy/pgie_config.txt name=pgie ! queue ! nvtracker tracker-width=640 tracker-height=384 ll-lib-file=/opt/nvidia/deepstream/deepstream-6.1//lib/libnvds_nvmultiobjecttracker.so ll-config-file=/home/sensormatic/vision-works-ds/cvads/configurations/occupancy/config_tracker_NvDCF_perf.yml enable-past-frame=0 enable-batch-process=1 ! queue ! nvinfer config-file-path=/home/sensormatic/vision-works-ds/cvads/configurations/occupancy/sgie_config.txt name=sgie ! queue ! appsink name=appsink emit-signals=True sync=false”
kesong
October 30, 2023, 8:26am
6
As you have the nvinfer source code, can you check if the object has been processed in function should_infer_object().
I am using python3 for pipeline and the same giving the results with nvosd but not in appsink.
kesong
October 31, 2023, 8:56am
8
Python pipeline also use nvinfer plugin, it will print the log if you add debug log in nvinfer source code and compile it.
Hi @kesong can you help me to resolve the issue when i use pgie->tracker->sgie i am not getting sgie results.
Meanwhile i will look into the nvinfer for appsink issue.
kesong
November 1, 2023, 10:12am
10
There is no update from you for a period, assuming this is not an issue anymore. Hence we are closing this topic. If need further support, please open a new one. Thanks
I need you add log in should_infer_object() and build nvinfer based on: /opt/nvidia/deepstream/deepstream-6.3/sources/gst-plugins/gst-nvinfer/README
system
November 20, 2023, 6:51am
11
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.