I have trained model and generate etlt and trt. I want to test this weight on video without using deepstream sdk because getting problem with DS if I stuck at some point in DS so it take 2-3 or more days to resolve it, so is there any sample code where i can test this weights normally like loading weights through tensorrt and detection on each frame and get b-box and draw b-box on frame with opencv and implement my custom logic.
Please help me out.
The tlt-infer tool can do inference against one frame and draw bbox.
From Where I can download this ? can you provide me the link.
It is a build-in tool inside tlt docker.
Can I use it without TLT docker ?
No, it only works inside TLT docker.