Dustin Franklin

Can you run `video-viewer rtsp://192.168.1.82:8554/live.ts` and confirm that you can view your RTSP stream with a display attached to your Jetson? If not, you could try rebuilding jetson-inference with cmake...

`net.GetNetworkFPS()` returns the FPS for just the DNN network. `videoOutput.GetFrameRate()` would return the frame rate that images are rendered to the display.

Hi @stekla79, pytorch-ssd will sample from negative areas to build up the negative examples, but I think each image requires at least one positive bounding box.

Perhaps the code could also be modified to work with images that had 0 annotations. I had to manually disable images with 0 annotations for VOC-type datasets, because it would...

> `/usr/lib/x86_64-linux-gnu/libpython3.8.so` You are building this on x86, but this project is only officially supported on Jetson. There are a number of Jetson-specific things, mostly in the jetson-utils module (such...

> I have the same issue when trying to build this inside the Jetpack 5.0 l4t-ml:r34.1.1-py3 container Have you tried building it like this: ``` bash $ cd jetson-inference $...

Hi @chris-volley, when you reflashed, did you do it with SDK Manager and did it install the NVIDIA Container Runtime for you, or did you do that manually? Do you...

Are you able to run this? ``` bash $ sudo docker run -it --rm --net=host --runtime nvidia nvcr.io/nvidia/l4t-base:r34.1 # python3 -c 'import tensorrt' ``` I wonder if you had done...

Hmm, ok. Sorry there was a typo in the command above: ``` bash $ sudo docker run -it --rm --net=host --runtime nvidia nvcr.io/nvidia/l4t-base:r32.7.1 # python3 -c 'import tensorrt' ```

Unfortunately I'm at a loss for why it would be failing to mount things inside container after freshly re-flashing it with SDK Manager, sorry about that. I assume that you...