Inference (raw output)

I would like to get the raw output of the bounding box coordinates when I run inference in the docker container. Currently, it only overlays the bounding box of the image; then saves the image. How would it be possible to get access to that raw model output ?

I currently do not have access to a Nvidia Jetson device.

Hi ishan,
The labels will be saved when you run tlt-infer against images.
Take detectnet_v2 jupyter notebook as an example, it says

  1. Overlain images in $USER_EXPERIMENT_DIR/tlt_infer_testing/images_annotated
  2. Frame by frame bbox labels in kitti format located in $USER_EXPERIMENT_DIR/tlt_infer_testing/labels

Is there a pointer to the detectnet_v2 jupyter notebook; the one i followed: https://github.com/NVIDIA-AI-IOT/tlt-iva-examples/blob/master/TLT_notebook.ipynb

The one above does not save any labels under a folder called “labels” only creates annotated images.

Refer to https://docs.nvidia.com/metropolis/TLT/tlt-getting-started-guide/index.html#building_docker