The United States Postal Service is adopting end-to-end AI technology from NVIDIA to improve its package data processing efficiency.
The new system starts with high-performance servers powered by NVIDIA V100 Tensor Core GPUs and deep learning software to train multiple AI algorithms. The trained models are then deployed to NVIDIA EGX edge computing systems at close to 200 Postal Service facilities throughout the U.S. to enable more efficient package data processing. The NVIDIA-powered systems are being purchased by the Postal Service under contract with Hewlett Packard Enterprise.
The Postal Service operates the world’s highest volume logistics operation, processing and delivering some 146 billion pieces of mail annually, including more than 6 billion packages. NVIDIA claims that the The new AI system will process package data 10x faster and with higher accuracy.
Engineering teams from the Postal Service and NVIDIA have been collaborating for several months to develop AI models, using NVIDIA software including TensorRT for inference optimization; automatic mixed precision in PyTorch to accelerate training while maintaining model accuracy; NGC containers, which are GPU-optimized for streamlining software deployment; and DeepOps tools for optimizing GPU clusters.
Delivery and testing of the system will start this year and it is expected to be fully operational by spring of 2020.