Get multiple classification results

Accelerated Computing Intelligent Video Analytics DeepStream SDK
#1

• Hardware Platform (Jetson / GPU) : GPU
• DeepStream Version : 5.1
• TensorRT Version : 7.2.1-1+cuda11.1
• NVIDIA GPU Driver Version (valid for GPU only) : 455.23.05
• Issue Type(questions, new requirements, bugs) : Question

Hi,

I´m using an image classifier as primary gie in Deepstream and I would like to get more than one prediction for each frame of the analyzed video but right now I´m able to get only one prediction for each frame.
I tried to do what it is explained in this topic: Multiple classification results.
However, it didn´t work for me as I said there: Multiple classification results - #9 by alexmoreno98.

Is it another way I could achieve my objective?

Thanks in advance.

#3

Sorry! I forgot to mention that the config needs to enable two options below

#0=Detection 1=Classifier 2=Segmentation 100=other
network-type=100
# Enable tensor metadata output
output-tensor-meta=1

did you also enable “network-type=100” ?

How to convert Pytorch model probabilities in Deepstream to probabilities between 0 and 1?
#4

I’ve tried setting “network-type=100” and “output-tensor-meta=1” but it remains the same.
User metadata structure is still empty because the execution doenst´t enter into the loop.

#5

is it possible to sahre a repo?

#6

Yes, is there a way to share it to you in private?

#7

Any update?

#8

Hi @alexmoreno98 ,
could you update the repo in Google drive,by which, I can untar it in deepstream6.0 docker (nvcr.io/nvidia/deepstream:6.0-devel), and then run command to repo it directly, and share me the Google drive link through private message ?

#9

Ok, thank you. I will send you the Google drive link through private message.