Feature reordering based on similarity for improved memory compression transfers during machine learning jobs
AssigneeATI Technologies ULC
InventorsArash Hariri, Mehdi Saeedi, Boris Ivanovic, Gabor Sines
Life of this patent
Estimating the term…
Abstract
A processing device for executing a machine learning neural network operation includes memory and a processor. The processor is configured to receive input data at a layer of the machine learning neural network operation, receive a plurality of sorted filters to be applied to the input data, apply the plurality of sorted filters to the input data to produce a plurality of different feature maps, compress the plurality of different feature maps according to a similarity of the feature maps relative to each other and store the plurality of different feature maps in the memory.
Claims20
A processing device for executing a machine learning neural network operation comprising:
- memory;
- and
- a processor configured to:
- receive input data at a layer of a machine learning neural network;
- receive a plurality of sorted filters to be applied to the input data;
- apply the plurality of sorted filters to the input data to produce a plurality of different feature maps;
- compress the plurality of different feature maps according to a similarity of the different feature maps relative to each other;
- and
- store the plurality of different feature maps in the memory.
A machine learning processing method comprising:
- receiving input data at a layer of a machine learning neural network;
- receiving a plurality of sorted filters to be applied to the input data;
- applying the plurality of sorted filters to the input data to produce a plurality of different feature maps;
- compressing the plurality of different feature maps according to a similarity of the different feature maps relative to each other;
- and
- storing the plurality of different feature maps in memory.
A non-transitory computer readable medium having stored instructions for causing a computer to execute a machine learning processing method comprising:
- receiving input data at a layer of a machine learning neural network;
- receiving a plurality of sorted filters to be applied to the input data;
- decompressing the input data;
- applying the plurality of sorted filters to the input data to produce a plurality of different feature maps;
- compressing the plurality of different feature maps according to a similarity of the different feature maps relative to each other;
- and
- storing the plurality of different feature maps in memory.
Dependencies are inferred from the claim text.