I have trained centerpose end-to-end on my custom data using synthetic data generated by Isaac Sim (older version 2023.1.1) using TAO Toolkit docker 5.2 centerpose model card.
I am able to use tao toolkit 5.2 for inference on images in a folder. However, I need to get inference on live camera feed. I tried to modify the code inside the script/inference.py but authors have used pyarmor_runtime and have obfuscated the code.
Could you please provide a guideline to get inference on live camera feed? My camera is Intel RealSense D435.
(base) mona@ada:~$ tao info
Configuration of the TAO Toolkit Instance
task_group: ['model', 'dataset', 'deploy']
format_version: 3.0
toolkit_version: 5.2.0.1
published_date: 01/16/2024
(base) mona@ada:~$ docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
1568b9368e74 nvcr.io/nvidia/tao/tao-toolkit:5.2.0-pyt2.1.0 "/opt/nvidia/nvidia_…" 7 hours ago Up 7 hours
(base) mona@ada:~$ docker exec -it cranky_khayyam bash
groups: cannot find name for group ID 1002
I have no name!@ada:/opt/nvidia/tools$ ls
Jenkinsfile.develop Jenkinsfile.main README.md README.txt build.sh converter tao-converter
I have no name!@ada:/opt/nvidia/tools$
I have no name!@ada:/opt/nvidia/tools$ find / -name "centerpose" -type d 2> /dev/null
/usr/local/lib/python3.10/dist-packages/nvidia_tao_pytorch/cv/centerpose
/workspace/tao-experiments/centerpose
/workspace/tao-experiments/data/centerpose
I also wrote this as an issue in [QST] Modifying code inside tao docker · Issue #35 · NVIDIA/tao_pytorch_backend · GitHub but seems not maintained as regularly.
Please provide the following information when requesting support.
• Hardware (T4/V100/Xavier/Nano/etc)
• Network Type (Detectnet_v2/Faster_rcnn/Yolo_v4/LPRnet/Mask_rcnn/Classification/etc)
• TLT Version (Please run “tlt info --verbose” and share “docker_tag” here)
• Training spec file(If have, please share here)
• How to reproduce the issue ? (This is for errors. Please share the command line and the detailed log here.)
1 post - 1 participant