Neural network training technique
AssigneeNVIDIA Corporation
InventorsZhiding Yu, Wuyang Chen, Shalini De Mello, Sifei Liu, Jose Manuel Alvarez Lopez, Anima Anandkumar
Life of this patent
Estimating the term…
Abstract
Apparatuses, systems, and techniques to train neural networks to perform image processing tasks. In at least one embodiment, one or more second neural networks are used to train one or more first neural networks based, at least in part, on a first object type in one or more images and a second object type in the one or more images, in parallel.
Claims20
One or more processors, comprising circuitry to:
- apply one or more neural networks to input data comprising one or more images to detect one or more instances of an object type, wherein the one or more neural networks are configured to distinguish the one or more instances of the object type from one or more other object types, the one or more neural networks having been previously trained based, at least in part, on outputs of one or more pre-trained neural networks corresponding to other object types relative to the object type;
- and
- output a detection result indicating one or more identified instances of the object type in the input data.
A system, comprising:
- one or more processors to:
- apply one or more neural networks to input data comprising one or more images to detect one or more instances of an object type, wherein the one or more neural networks are configured to distinguish the one or more instances of the object type from one or more other object types, the one or more neural networks having been previously trained based, at least in part, on outputs of one or more pre-trained neural networks corresponding to other object types relative to the object type;
- and
- output a detection result indicating one or more identified instances of the object type in the input data.
A non-transitory machine-readable medium having stored thereon a set of instructions that, when executed by one or more processors, cause the one or more processors to:
- apply one or more neural networks to input data comprising one or more images to detect one or more instances of an object type, wherein the one or more neural networks are configured to distinguish one or more instances of the object type from one or more other object types, the one or more neural networks have been previously trained based, at least in part, on outputs of one or more pre-trained neural networks that correspond to other object types relative to the object type;
- and
- output a detection result indicating one or more identified instances of the object type in the input data.
Dependencies are inferred from the claim text.