Efficient machine learning model architectures for training and inference
AssigneeQUALCOMM Incorporated
InventorsJoseph Binamira SORIAGA, Hossein HOSSEINI
Life of this patent
Estimating the term…
Loading patent…
AssigneeQUALCOMM Incorporated
InventorsJoseph Binamira SORIAGA, Hossein HOSSEINI
Estimating the term…
Certain aspects of the present disclosure provide techniques for improved machine learning. A data tensor is generated as output from a layer of a neural network. A first subset of the first data tensor and a second subset of the first data tensor are generated using a tensor splitting operation. The second subset of the first data tensor is stored, and the first subset of the first data tensor is provided to a subsequent layer of the neural network. One or more parameters of the layer of the neural network are refined based at least in part on the stored second subset of the first data tensor.
A computer-implemented method, comprising:
A processing system, comprising:
A non-transitory computer-readable medium comprising computer-executable instructions that, when executed by one or more processors of a processing system, cause the processing system to perform an operation comprising:
A computer-implemented method, comprising:
Dependencies are inferred from the claim text.