This browser is no longer supported.
Upgrade to Microsoft Edge to take advantage of the latest features, security updates, and technical support.
Answer the following questions to check your learning.
The DeepStream Graph Composer can be described as a collection of the following features and functionalities:
An SDK that contains optimized libraries for GPU accelerated inference tasks
Tooling for drawing charts that display inference detections in DeepStream apps
A visual environment for developing IVA applications that make use of the NVIDIA DeepStream SDK, a registry tool that provides extensions and components for use in the visual environment, and tooling for packaging NVIDIA DeepStream Graphs into containerized applications
When you run a container application that requires access to the X11 environment for drawing a user interface, the following details must be accounted for:
The container runtime will configure itself and just work.
Before you run the workload in question, the host machine needs to enable access to the X11 environment, which can be accomplished with sudo xhost +. The container must also mount in the system X11 socket at runtime.
sudo xhost +
Enable the option for X11 access when you build the container.
Distributing a GPU accelerated workload as a container provides the following benefits:
Containerized workloads can be deployed across various architectures without the need for repackaging.
Containerized workloads are guaranteed to exhibit similar runtime behavior when deployed to homogenous hardware. Time to deployment is potentially reduced because a full download of incremental image updates isn't required. Easy-to-understand required solution dependencies are created by inspecting the associated Docker file of a given container image.
GPU acceleration isn't possible because of access limitations when you run workloads as a container.
You must answer all questions before checking your work.
Continue
Was this page helpful?