Cloud-AI Video-Streaming Platform
NVIDIA Maxine is a completely accelerated platform SDK for developers of movie conferencing services to create and deploy features that are AI-powered use state-of-the-art models within their cloud. Video conferencing applications based on Maxine can lessen movie bandwidth usage right down to one-tenth of H.264 utilizing AI video clip compression, considerably reducing expenses.
Maxine includes APIs for the latest innovations from NVIDIA research such as for instance face positioning, look modification, face re-lighting and real-time interpretation along with abilities such as for instance super-resolution, sound treatment, shut captioning and digital assistants.
These abilities are completely accelerated on NVIDIA GPUs to perform in real-time video clip applications that are streaming the cloud.
Maxine-based applications allow solution providers provide the exact same features to every individual on any unit, including computer systems, pills, and phones. Applications constructed with Maxine can effortlessly be implemented as microservices that scale to thousands of channels in a Kubernetes environment.
NVIDIA Maxine Features
User friendly SDK
Includes libraries, tools and pipelines that are example designers to quickly add AI features with their applications.
AI Video Compression utilizes one-tenth the bandwidth of H.264 video clip compression standard.
State-of-the-Art AI versions
Includes models that are pre-trained hundreds or even thousands of hours of training on NVIDIA DGXв„ў A100.
Optimizes end-to-end pipelines for the performance that is highest on NVIDIA Tensor Cores GPUs. (more…)