In this article (https://developer.nvidia.com/deepstream-sdk) it is mentioned that
“DeepStream application running on Jetson Nano with ResNet-based object
detector concurrently on eight independent 1080p 30fps video streams.”
Q1. Can you tell me which Resnet based detector was used for implementation ?
Resnet 18, Resnet 34, Resnet 50 and many others… ?
Q2. Is this model using SSD ?
Thanks.
Hi,
1.
It called resnet-10, a customized resnet model for detection from NVidia.
[url]https://ngc.nvidia.com/catalog/models/nvidia:tlt_iva_object_detection_resnet10[/url]
2.
No.
It’s customized model and doesn’t related to SSD.
Thanks.
Thanks for taking time and answering the question.
Can you please share the steps on how to train this model ?
Hi,
It’s recommended to use our Transfer Learning Toolkit:
[url]https://developer.nvidia.com/transfer-learning-toolkit[/url]
Thanks.
Hi,
wish to clarify: Is the RPN/localisation model faster-RCNN? Or it is a closed/undisclosed model?
Thanks