The system and methods disclosed herein include a runtime architecture that takes a
nonspecific set of systems of differential equations, distributes them across the network, and …

One embodiment provides for a system to configure distributed training of a neural network.
The system includes memory to store a library to facilitate transmission of data during …

Apparatuses, systems, and techniques to improve neural network computations. In at least
one embodiment, a deep neural network library receives computation descriptors from one …

Apparatuses, systems, and techniques to automatically generate a reduced number of
compute kernels for performing operations of one or more neural networks. In at least one …

A method and system for computing one or more outputs of a neural network having a
plurality of layers is provided. The method and system can include determining a plurality of …

Abstract Systems and methods for a scalable artificial neural network, wherein the
architecture includes: an input layer; at least one hidden layer; an output layer; and a …

For one embodiment, a hardware accelerator with a heterogenous architecture for training
quantized neural networks is described. In one example, a hardware accelerator for training …

The disclosed embodiments relate to a system that facilitates dynamic runtime execution of a
deep neural network (DNN). During operation, the system receives a model, a set of weights …

Apparatuses, systems, and techniques to transform input data for training neural networks. In
at least one embodiment, one or more data transforms are identified in a sequence of data …

(57) ABSTRACT A system receives and executes a sequence oftensor instructions, for
example, instructions for performing a neural network computation. The system may be …