Using model on outside DS and/or python


Say I managed to train a detection model on my classes and data (Res1X for example).

Now, I wish to integrate it to my flow.Since I am using JP4.2, deepstream is currently not an option (not supported yet), and my live camera works now only with Python code.

Are there any options:
– To use the trained model with Python ? If so - is there a sample code for preparing the data, submitting it to the model and parsing (interpreting) the results?
– If not - is there a way I can run this on videos in c/c++, with my “Standalone” exe without DS wrappers ?



Sorry, but currently our workflow is only compatible with DeepStream.