Is there a way to use Conv3D in Deepstream SDK?

Accelerated Computing Intelligent Video Analytics DeepStream SDK
#1

Hi,
I want to use a DNN model using conv3d in Deepstream sdk.

I transformed the pytorch model (.pth) into TensorRT engine (.engine) successfully.

However, as you know, Conv3d requires the input image with 5 dim tensor. (batch, channel, sequence_length, width, height)

I am wondering is there an example test case using Conv3d at deepstream sdk??
I already checked C/C++ Sample Apps and python Apps in GitHub - NVIDIA-AI-IOT/deepstream_python_apps: A project demonstrating use of Python for DeepStream sample apps given as a part of SDK (that are currently in C,C++).

It will be very helpful!!

Thanks in advance,
Sincerely,
Joe Jang

#2

Currently, I am considering to attach image as an usermeta at the nvinfer plugin

#3

Moved to the Deepstream SDK forum

1 Like
#5

No. Current deepstream does not support cnv3d model inference.

#6

Thanks

#8

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.