ANNOUNCING NVIDIA® cuDNN EMBEDDED – GPU Accelerated Machine Learning for Jetson TK1
NVIDIA cuDNN is a GPU-accelerated library of primitives for deep neural networks. It emphasizes performance, ease-of-use, and low memory overhead. Integrated into higher-level machine learning frameworks, such as UC Berkeley’s popular Caffe software, it’s now available for the wildly popular Jetson TK1. Perfect for low-power computer vision and machine learning applications, like robotics and autonomous vehicles, this platform provides an amazing level of performance packed into a low-power board.
- Forward and backward convolution routines, tuned for NVIDIA GPUs
- Always optimized for latest NVIDIA GPU architectures
- Arbitrary dimension ordering, striding, and subregions for 4d tensors
- Forward and backward paths for common layer types (ReLU, Sigmoid, Tanh, pooling, softmax)
- Context-based API allows for easy multithreading
Product Manager – Strategic Alliances