How to define the kafka message with my own classes-info format

environment: docker container with samples rev for deepstream5.1

I want to run the deepstream under python code to send my own kafka message with my own model. I have successfully sending the kafka message with my own model but I don’t know how to revise the message format by myself. Hope somebody help to show me the detail, thanks in advance!

Your topic was posted in the wrong category. I am moving this to the DeepStream SDK category for visibility.

What format do you want to revise? Could you refer our demo code below?
https://github.com/NVIDIA-AI-IOT/deepstream_python_apps/tree/master/apps/deepstream-test4

Thanks!
Yes, I had checked this code, it seems the message have already defined in the c code? I want to define the message by my own model with my class_names.

You can refer the user guide and the demo code below to learn how to use our msgconv plugin.
https://docs.nvidia.com/metropolis/deepstream/dev-guide/text/DS_plugin_gst-nvmsgbroker.html#nvds-kafka-proto-kafka-protocol-adapter

Thanks, I will check it.
Or, is it possible to replace the block of msg-converter and broker by pykafka?
I had tried to comment out the queue for msgconv and add the pykafka code in tiler_src_pad_buffer_probe function, and I find it works. But I’m not clear whether it may affect the program process?

We didn’t suggest you replace it. The two plugins are at the C layer actually. The framework will use the C plugins by default. If you want to replace it, you may need to change lots of codes.

I had replace the code, and found the message can be sent successfully.
As your suggestion, I want to redefine the message, but I’m not clear how to define the message in C layer… for the original is only suitable for the classes of CAR, PERSON and so on. But I want to change it to my own class names.

You don’g have to define your message in the code, just in the config file.You can refer the link below:
https://docs.nvidia.com/metropolis/deepstream/dev-guide/text/DS_plugin_gst-nvmsgconv.html
You can also refer our demo code below:
https://github.com/NVIDIA-AI-IOT/deepstream_python_apps/tree/master/apps/deepstream-test4

1 Like

Many thanks for your kind help, I will check the detail.

Do you still need support for this topic? Or should we close it? Thanks.

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