Home > Servers > PowerEdge Components > White Papers > Developing and Deploying Vision AI with Dell and NVIDIA Metropolis > Software stack
The following table specifies the software stack, versions, and useful resources.
Software stack, tools, and GitHub repos | |
OS | 18.04.4 LTS (GNU/Linux 4.15.0-154-generic x86_64) |
Software stack | |
Pre-trained Models | TAO Pretrained Object Detection |
Framework Back end | TensorFlow |
Training dataset | KITTI |
Delivery method | Container |
RetinaNet Jupyter Notebook retinanet.ipynb | |
Python | 3.6.9 |
TensorRT | 7.2.2 |
CUDA | 11.1 |
cuDNN | 8.0.0+ |
NVIDIA driver | 470.57.02 |
Tools | |
TAO Toolkit 3.21.08 | |
For model conversion | TAO converter for x86 + GPU platform with CUDA 11.1 / cuDNN 8.0 / TensorRT 7.2 |
For model visualization | inspection | NVIDIA Polygraphy |
For model deployment at scale | Docker image deepstream:6.0-ea-21.06-Triton |
Links to computer vision models and container collection | |
TAO pre-trained models | https://ngc.nvidia.com/catalog/models/nvidia:tao:pretrained_object_detection/version |
TAO converter | |
Jupyter notebooks collection | https://ngc.nvidia.com/catalog/resources/nvidia:tao:cv_samples |