Abstract: An apparatus to facilitate optimization of a convolutional neural network (CNN) is disclosed. The apparatus includes optimization logic to receive a CNN model having a list of instructions and including pruning logic to optimize the list of instructions by eliminating branches in the list of instructions that comprise a weight value of 0.

Abstract: A mechanism is described for facilitating smart convolution in machine learning environments. An apparatus of embodiments, as described herein, includes one or more processors including one or more graphics processors, and detection and selection logic to detect and select input images having a plurality of geometric shapes associated with an object for which a neural network is to be trained. The apparatus further includes filter generation and storage logic (“filter logic”) to generate weights providing filters based on the plurality of geometric shapes, where the filter logic is further to sort the filters in filter groups based on common geometric shapes of the plurality of geographic shapes, and where the filter logic is further to store the filter groups in bins based on the common geometric shapes, wherein each bin corresponds to a geometric shape.

Abstract: Embodiments provide systems and methods which facilitate optimization of a convolutional neural network (CNN). One embodiment provides for a non-transitory machine-readable medium storing instructions that cause one or more processors to perform operations comprising processing a trained convolutional neural network (CNN) to generate a processed CNN, the trained CNN having weights in a floating-point format. Processing the trained CNN includes quantizing the weights in the floating-point format to generate weights in an integer format. Quantizing the weights includes generating a quantization table to enable non-uniform quantization of the weights and quantizing the weights from the floating-point format to the integer format using the quantization table. The operations additionally comprise performing an inference operation utilizing the processed CNN with the integer format weights.

Abstract: A mechanism is described for facilitating smart convolution in machine learning environments. An apparatus of embodiments, as described herein, includes one or more processors including one or more graphics processors, and detection and selection logic to detect and select input images having a plurality of geometric shapes associated with an object for which a neural network is to be trained. The apparatus further includes filter generation and storage logic (“filter logic”) to generate weights providing filters based on the plurality of geometric shapes, where the filter logic is further to sort the filters in filter groups based on common geometric shapes of the plurality of geographic shapes, and where the filter logic is further to store the filter groups in bins based on the common geometric shapes, wherein each bin corresponds to a geometric shape.

Abstract: An apparatus to facilitate acceleration of machine learning operations is disclosed. The apparatus comprises at least one processor to perform operations to implement a neural network and accelerator logic to perform communicatively coupled to the processor to perform compute operations for the neural network.

Abstract: One embodiment provides a graphics processor comprising an instruction cache to store an instruction and a compute block configured to perform multiply-accumulate operations in response to execution of the instruction. The compute block includes a scheduler to schedule a plurality of threads for execution of the instruction and multiply-accumulate circuitry configured to execute the instruction via the plurality of threads, wherein the multiply-accumulate circuitry includes a plurality of functional units configured to process, in parallel via the plurality of threads, a corresponding plurality of matrix elements to multiply a first matrix and a second matrix, and to multiply the first matrix and the second matrix includes to multiply data elements in a row of the first matrix by corresponding data elements in a column of the second matrix to generate a plurality of products.

Abstract: Input images are partitioned into non-overlapping segments perpendicular to a disparity dimension of the input images. Each segment includes a contiguous region of pixels spanning from a first edge to a second edge of the image, with the two edges parallel to the disparity dimension. In some aspects, contiguous input image segments are assigned in a “round robin” manner to a set of sub-images. Each pair of input images generates a corresponding pair of sub-image sets. Semi-global matching processes are then performed on pairs of corresponding sub-images generated from each input image. The SGM processes may be run in parallel, reducing an elapsed time to generate respective disparity sub-maps. The disparity sub-maps are then combined to provide a single disparity map of equivalent size to the original two input images.

Abstract: One embodiment provides an apparatus comprising an instruction cache to store a plurality of instructions, a scheduler unit coupled to the instruction cache, the scheduler unit to schedule the plurality of instructions for execution, an instruction fetch and decode unit to decode the plurality of instructions to determine a set of operations to perform in response, one or more compute blocks to perform parallel multiply-accumulate operations based on the instruction fetch and decode unit decoding a first instruction of the plurality of instructions, and matrix multiplication logic to perform matrix multiplication operations based on the instruction fetch and decode unit decoding a second instruction of the plurality of instructions.

Abstract: An apparatus to facilitate acceleration of machine learning operations is disclosed. The apparatus comprises at least one processor to perform operations to implement a neural network and accelerator logic to perform communicatively coupled to the processor to perform compute operations for the neural network.

Abstract: One embodiment provides an apparatus comprising an instruction cache to store a plurality of instructions, a scheduler unit coupled to the instruction cache, the scheduler unit to schedule the plurality of instructions for execution, an instruction fetch and decode unit to decode the plurality of instructions to determine a set of operations to perform in response, one or more compute blocks to perform parallel multiply-accumulate operations based on the instruction fetch and decode unit decoding a first instruction of the plurality of instructions, and matrix multiplication logic to perform matrix multiplication operations based on the instruction fetch and decode unit decoding a second instruction of the plurality of instructions.

Abstract: A library of machine learning primitives is provided to optimize a machine learning model to improve the efficiency of inference operations. In one embodiment a trained convolutional neural network (CNN) model is processed into a trained CNN model via pruning, convolution window optimization, and quantization.

Abstract: One embodiment provides for a general-purpose graphics processing unit including a scheduler to schedule multiple matrix operations for execution by a general-purpose graphics processing unit. The multiple matrix operations are determined based on a single machine learning compute instruction. The single machine learning compute instruction is a convolution instruction and the multiple matrix operations are associated with a convolution operation.

Abstract: One embodiment provides for a compute apparatus to perform machine learning operations, the apparatus comprising a decode unit to decode a single instruction into a decoded instruction, the decoded instruction to perform one or more machine learning operations, wherein the decode unit, based on parameters of the one or more machine learning operations, is to request a scheduler to schedule the one or more machine learning operations to one of an array of programmable compute units and a fixed function compute unit.

Abstract: Input images are partitioned into non-overlapping segments perpendicular to a disparity dimension of the input images. Each segment includes a contiguous region of pixels spanning from a first edge to a second edge of the image, with the two edges parallel to the disparity dimension. In some aspects, contiguous input image segments are assigned in a “round robin” manner to a set of sub-images. Each pair of input images generates a corresponding pair of sub-image sets. Semi-global matching processes are then performed on pairs of corresponding sub-images generated from each input image. The SGM processes may be run in parallel, reducing an elapsed time to generate respective disparity sub-maps. The disparity sub-maps are then combined to provide a single disparity map of equivalent size to the original two input images.

Abstract: An apparatus to facilitate acceleration of machine learning operations is disclosed. The apparatus comprises at least one processor to perform operations to implement a neural network and accelerator logic to perform communicatively coupled to the processor to perform compute operations for the neural network.

Abstract: A mechanism is described for facilitating smart convolution in machine learning environments. An apparatus of embodiments, as described herein, includes one or more processors including one or more graphics processors, and detection and selection logic to detect and select input images having a plurality of geometric shapes associated with an object for which a neural network is to be trained. The apparatus further includes filter generation and storage logic (“filter logic”) to generate weights providing filters based on the plurality of geometric shapes, where the filter logic is further to sort the filters in filter groups based on common geometric shapes of the plurality of geographic shapes, and where the filter logic is further to store the filter groups in bins based on the common geometric shapes, wherein each bin corresponds to a geometric shape.

Abstract: An apparatus to facilitate compute optimization is disclosed. The apparatus includes a at least one processor to perform operations to implement a neural network and compute logic to accelerate neural network computations.

Abstract: Embodiments provide systems and methods which facilitate optimization of a convolutional neural network (CNN). One embodiment provides for a non-transitory machine-readable medium storing instructions that cause one or more processors to perform operations comprising processing a trained convolutional neural network (CNN) to generate a processed CNN, the trained CNN having weights in a floating-point format. Processing the trained CNN includes quantizing the weights in the floating-point format to generate weights in an integer format. Quantizing the weights includes generating a quantization table to enable non-uniform quantization of the weights and quantizing the weights from the floating-point format to the integer format using the quantization table. The operations additionally comprise performing an inference operation utilizing the processed CNN with the integer format weights.

Abstract: One embodiment provides for a compute apparatus to perform machine learning operations, the apparatus comprising a decode unit to decode a single instruction into a decoded instruction, the decoded instruction to perform one or more machine learning operations, wherein the decode unit, based on parameters of the one or more machine learning operations, is to request a scheduler to schedule the one or more machine learning operations to one of an array of programmable compute units and a fixed function compute unit.

Abstract: An apparatus to facilitate optimization of a convolutional neural network (CNN) is disclosed. The apparatus includes optimization logic to receive a CNN model having a list of instructions and including pruning logic to optimize the list of instructions by eliminating branches in the list of instructions that comprise a weight value of 0.