Deepstream 6.4 doc

Accelerated Computing Intelligent Video Analytics DeepStream SDK
1
  • dGPU
  • DS 6.4

While looking for recent docs I always land here Gst-nvmsgconv — DeepStream documentation 6.4 documentation

The doc states, that:

deepstream-test4-app

  • Use commandline option –msg2p-meta to choose the metadata type to create payload.
  • 0=Event Msg meta(default), Create NVDS_EVENT_MSG_META type of meta and attach to buffer
  • 1=nvdsmeta, Use the fields within NvDsFrameMeta/NvDsObjectMeta to populate payload

But there is no –msg2p-meta option.

2

And there is also no deepstream-test5-app

3
  • By modifying the payload generator library: To perform a simple customization of DeepStream schema fields, modify the low-level payload generation library files inside directory sources/libs/nvmsgconv. Payload generated using NVDS_EVENT_MSG_META (NvDsEventMsgMeta) type can be customized by modifying sources/libs/nvmsgconv/deepstream_schema/eventmsg_payload.cpp. If payload is generated using metadata within DS GST buffers(NVDS_FRAME_META (NvDsFrameMeta) type and NVDS_OBJECT_META (NvDsObjectMeta) type), you can customize payload genertion by modifying the file sources/libs/nvmsgconv/deepstream_schema/dsmeta_payload.cpp.

This in mind: is there any demo which shows how to use the provisioning other than NVDS_EVENT_MSG_META, as deepstream-test4-app demonstrates?

5

I am checking.

6

sorry for the late reply. nvmsgconv has msg2p-newapi property. you can set 0/1 to enable Event Msg meta/nvdsmeta. if you are testing deepstream-test4, you can use --msg2p-meta to choose the metadata type. please refer to opt\nvidia\deepstream\deepstream-6.4\sources\apps\sample_apps\deepstream-test4\README

7

Ah, ok. I was referring to the python apps, there is no such a flag

8

Anyway, no issue anymore. I have dropped nvmsgconv and nvmsgbroker and do my own MQTT publishing. The current nv* components produce 10 to 20 times more CPU load than doing it manually

9

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