Per kernel kmeans compression for neural networks
AssigneeIntel Corporation
InventorsYonatan Glesner, Gal Novik, Dmitri Vainbrand, Gal Leibovich
Life of this patent
Estimating the term…
Abstract
Methods and apparatus relating to techniques for incremental network quantization. In an example, an apparatus comprises logic, at least partially comprising hardware logic to determine a plurality of weights for a layer of a convolutional neural network (CNN) comprising a plurality of kernels; organize the plurality of weights into a plurality of clusters for the plurality of kernels; and apply a K-means compression algorithm to each of the plurality of clusters. Other embodiments are also disclosed and claimed.
Claims17
1-20. (canceled)
An apparatus comprising a processor to:
- determine a plurality of weights for a layer of a convolutional neural network (CNN) comprising a plurality of kernels;
- organize the plurality of weights into a plurality of clusters on a per-kernel basis for the plurality of kernels;
- compute a center point for the plurality of weights on a per-kernel basis in each cluster;
- and
- apply a K-means compression algorithm to each of the plurality of clusters.
A processor-implemented method, comprising:
- determining a plurality of weights for a layer of a convolutional neural network (CNN) comprising a plurality of kernels;
- organizing the plurality of weights into a plurality of clusters on a per-kernel basis for the plurality of kernels;
- computing a center point for the plurality of weights on a per-kernel basis in each cluster;
- and
- applying a K-means compression algorithm to each of the plurality of clusters.
One or more non-transitory computer-readable medium comprising one or more instructions that when executed on at least one processor configure the at least one processor to perform one or more operations to:
- determine a plurality of weights for a layer of a convolutional neural network (CNN) comprising a plurality of kernels;
- organize the plurality of weights into a plurality of clusters on a per-kernel basis for the plurality of kernels;
- compute a center point for the plurality of weights on a per-kernel basis in each cluster;
- and
- apply a K-means compression algorithm to each of the plurality of clusters.
Dependencies are inferred from the claim text.