Training Code for License Plate Recognition and Detectition

Accelerated Computing Intelligent Video Analytics TAO Toolkit
1

Hi everyone,

I’ve known that NVIDIA provided the Transfer Learning Toolkit to re-train the License Plate Detection and License Plate Recognition with our custom datasets easily. I ran the training code easily without any error.

I would like to know more about the training code such as what is the framework (Keras + Tensorflow?) , is there any support for us to use another framework like Pytorch?

Furthermore, I want to integrate the trained model with my Python Application without applying the Deepstream. How can I do that? I mean I want to use the model independently.

Thank you in advance.

Reference:
https://developer.nvidia.com/blog/creating-a-real-time-license-plate-detection-and-recognition-app

2

It is not open source. What do you mean by “for us to use another framework like Pytorch” ?

FOr deploying the trained model or trt engine, you can leverage GitHub - NVIDIA-AI-IOT/deepstream_lpr_app: Sample app code for LPR deployment on DeepStream