Not getting classification result in image-data-multistream for vehicle Type

Hi Team,

I am using Nvidia Nx-Xavier system, jetpack-4.4-b186 and DS-5.0.
I am running python DS test app-2 and I am able to get the Vehicle type classification result. but with the deepstream-image-data multistream application I am able get the type classification result but only some frame (if I am getting result of type classification on 2000 frame in test-app-2 so for the same configuration I am getting result on 5 frame with deepstream-imagedata-multistream app).
The problem is only with the type classification other classifier are working on both application.

below I am attaching the result of classification and Codebase file too.

CodeBase : (18.3 KB) (14.6 KB)

Configuration file :

dstest2_pgie_config.txt (3.4 KB)
dstest2_sgie1_config.txt (3.5 KB)
dstest2_sgie2_config.txt (3.5 KB)
dstest2_sgie3_config.txt (3.5 KB)
dstest2_tracker_config.txt (1.8 KB)

Please help me out in finding why I am unable to get classification results on deepstream-imagedata-multistream.



Just check your and
The pipeline is different.



And for


The sgie3, which represents for the vehicle type classifier, is missing in the pipeline of
So please add the sgie3 back to the pipeline to fix this issue.


Thanks @AastaLLL for the rply.

Actually in deepstream-imagedata-multistream I am not using car color classifier.
In deepstream-imagedata-multistream sgie1 and sgie2 represent vehicle type and make.

sgie1.set_property(‘config-file-path’, “dstest2_sgie2_config.txt”)
sgie2.set_property(‘config-file-path’, “dstest2_sgie3_config.txt”)

please help where I am wrong.


There is no update from you for a period, assuming this is not an issue any more.
Hence we are closing this topic. If need further support, please open a new one.


Sorry that I don’t notice that.

Based on the pipeline, the behavior for sgie1 and sgie2 should be similar.
Could you get the expected classification result from sgie1?