any recognition model on deepstream (such as text recognition)?

Hi
is there any recognition model on deepstream that can detect text and recognize the sentence?
I have a model but i dont have the prototrxt, uff or onnx file in order to convert to tensorrt?
any suggestion?

Hi,

Sorry that Deepstream doesn’t have a model for text or sentence.
It’s recommended to check our NGC webpage first:
https://ngc.nvidia.com/catalog/models?orderBy=modifiedDESC&query=&quickFilter=models&filters=

There are lots of model and you might find one for your usecase.

By the way, could you share which frameworks do you use for the your model?
TensorRT can support various frameworks, ex. Caffe, TensorFlow, pyTorch.

Thanks.

hi Aasta
I am doing license plate recognition (LPR) on pyTourch. which I want to do on deepstream.

Hi,

pyTorch model is supported by the TensorRT.
Please check this page for the detail instructions:
https://github.com/NVIDIA-AI-IOT/torch2trt

Thanks.