Running re_identification_net in deepstream

Accelerated Computing Intelligent Video Analytics DeepStream SDK
23

Yeah, I have tried and got this error,

image
image1415×206 13.3 KB

Found similar forum mentioning a solution like this,
image

But the location to be changed didn’t specified there.

24

Is there a physical monitor connected with your board? What happened after you changing the sink to fakesink?

25

No, I don’t have a monitor.

I didn’t change the sink and where can I change that?

26

Please open the /opt/nvidia/deepstream/deepstream/sources/apps/sample_apps/deepstream-test1/deepstream_test1_app.c and change the following code:

if(prop.integrated) {
    sink = gst_element_factory_make("nv3dsink", "nv3d-sink");
  } else {

TO

if(prop.integrated) {
    sink = gst_element_factory_make("fakesink", "nv3d-sink");
  } else {

Then you can build the sample app and use it.

27

Yeah, Now I can run the sample app

28

Then you can follow the NVIDIA-AI-IOT/deepstream_tao_apps at release/tao4.0_ds6.3ga (github.com) README to prepare the MDX sample.

29

I was able to run the MDX and got the output video.

But I want to get the embeddings out by a message broker and the kafka message broker consumer part is missing in the README which is available for the latest MDX release.

30

The message broker part was added since DeepStream 6.4. You can port the message broker code by yourself.

31

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

32

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.