Add new class in Primary Detector NVIDIA deepstream sdk

Accelerated Computing Intelligent Video Analytics DeepStream SDK
1

Hello, I am new to the Deep Stream SDK, My problems are :

  1. I want to add new class for detection in Primary_Detector(nvidia deepstream sdk) and also want to build my own classifier like Secondary_CarColor Secondary_CarMake like that so how do I can add new class and train detection and classification model incrementally.

  2. I have pnet.caffe rnet.caffe onet.caffe and that is i am using for head detection is there any way to convert this model into deep stream weights like .caffemodel and .engine file because i want to run my detection model on NVIDIA deep stream sdk.

Please help me out Thanks in advanced.

2

For adding new class, you can refer https://docs.nvidia.com/metropolis/deepstream/plugin-manual/index.html#page/DeepStream_Plugin_Manual%2Fdeepstream_plugin_metadata_migration.html%23
For 2nd problem, NVIDIA does not own the 3rd-party model conversion.

3

Hi,
Regarding caffe to tensorrt please see https://devtalk.nvidia.com/default/topic/1032511/jetson-tx2/converting-caffe-model-to-tensorrt/

Caffe parser:
https://docs.nvidia.com/deeplearning/sdk/tensorrt-api/python_api/parsers/Caffe/pyCaffe.html

Thanks

4

Hi eyal.enav,

Thanks for sharing!