NVIDIA researchers' landmark achievement in machine learning uses multiresolution hash encoding

Researchers from NVIDIA have developed a method for very quickly training neural graphics primitives using a single GPU. Neural graphic primitives have traditionally required multiple, fully connected neural networks and are challenging, time-consuming and expensive to train and evaluate. The research team, comprised of Thomas Müller, Alex Evans, Christoph Schied, and Alexander Keller, has created a new input encoding method that significantly reduces the number of floating point and memory acc...