• Hardware Platform (Jetson / GPU): Jetson NX Dev Kit
• DeepStream Version: 6.0
• JetPack Version (valid for Jetson only): 4.6 GA
Currently I am trying to run the Deepstream-python example: deepstream_test_2.py. I have not changed any codes, and run the script by
python3 deepstream_test_2.py /opt/nvidia/deepstream/deepstream-6.0/samples/streams/sample_qHD.h264 1 on the NX Dev machine.
I see the video display with bounding box detected. However,
null. I didn’t find the solution and questions in forum. And I am wondering whether you can help, for why it is null in this example, and how to use
Thank you so much.
Can you share more what do you want to do? Can you share more on you use case?
I just want to run this deepstream_test_2.py example and see results. And I found
null. I don’t quite understand why, or something I did wrong. I am wondering whether you can re-run this example, and see the value of
The reason why I am very interested in this result is because I want to know how to extract meta-data from multiple models by using python bindings. I don’t know whether some answers can be found in this example.
Thanks a lot.
Why you care about batch_user_meta_list? What do you want to get from it?
I don’t quite understand your comments.
So, the reason why I raised this question is because when I run deepstream_test_2.py example with official test video containing cars and human, I found
null. I just want to know whether this
null return is correct or not, and why.
I don’t quite understand the usage of
batch_meta.batch_user_meta_list, that’s the reason I asked this question. In my mind, since we already have batch meta, obj meta, display meta. I don’t quite understand the usage of batch user meta. And I didn’t find many explanations online.
On the other side, I want to know how to extract meta-data from multiple models by using python bindings. I don’t know whether some answers can be found in this example.
Maybe this question is too simple for you, haha. But if you don’t mind, can you help to rerun
deepstream_test_2.py on your side, and see whether you will get the same result as me? And if you want, can you help to give me a link to explain the usage of
Thanks a lot for your help.
batch_user_meta_list is for nvinfer output tensor or hold jpeg image if enabled. batch_user_meta_list will be null if you dont use those feature.
deepstream_test_2.py example, batch_user_meta_list should not return null, right? because obj_meta contains vehicle and human detected information.
I got what’s the problem. I need to set interval not equal to zero in dstest2_pgie_config.txt and enable-past-frame=1 in dstest2_tracker_config.txt.
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.