At the 2018 GPU Technology Conference, NVIDIA announced TensorRT 4, a programmable inference accelerator that speeds up videos, recommender systems, and speech applications. Using TensorRT and NVIDIA DGX-1, California-based startup REELY achieved 100X faster inferencing over traditional CPU-based methods to generate real-time video highlights for 140 HD streams simultaneously.
- Category
- Computing
Sign in or sign up to post comments.
Be the first to comment