知搜-全球专利检索

  1. Login
  2. Sign Up

Search Results

27 records (took 0.016 seconds)

    WEIGHT DEMODULATION FOR A GENERATIVE NEURAL NETWORK
    2.
    发明申请
    WEIGHT DEMODULATION FOR A GENERATIVE NEURAL NETWORK 有权

    公开(公告)号：US20210150369A1

    公开(公告)日：2021-05-20

    申请号：US17160585

    申请日：2021-01-28

    Applicant: NVIDIA Corporation

    Inventor： Tero Tapani Karras ,  Samuli Matias Laine ,  Jaakko T. Lehtinen ,  Miika Samuli Aittala ,  Janne Johannes Hellsten ,  Timo Oskari Aila

    IPC: G06N3/08 ,  G06N3/04

    Abstract: A style-based generative network architecture enables scale-specific control of synthesized output data, such as images. During training, the style-based generative neural network (generator neural network) includes a mapping network and a synthesis network. During prediction, the mapping network may be omitted, replicated, or evaluated several times. The synthesis network may be used to generate highly varied, high-quality output data with a wide variety of attributes. For example, when used to generate images of people's faces, the attributes that may vary are age, ethnicity, camera viewpoint, pose, face shape, eyeglasses, colors (eyes, hair, etc.), hair style, lighting, background, etc. Depending on the task, generated output data may include images, audio, video, three-dimensional (3D) objects, text, etc.

    Smoothing regularization for a generative neural network
    3.
    发明授权
    Smoothing regularization for a generative neural network 有权

    公开(公告)号：US11620521B2

    公开(公告)日：2023-04-04

    申请号：US17160648

    申请日：2021-01-28

    Applicant: NVIDIA Corporation

    Inventor： Tero Tapani Karras ,  Samuli Matias Laine ,  Jaakko T. Lehtinen ,  Miika Samuli Aittala ,  Janne Johannes Hellsten ,  Timo Oskari Aila

    IPC: G06N3/08 ,  G06N3/04

    Abstract: A style-based generative network architecture enables scale-specific control of synthesized output data, such as images. During training, the style-based generative neural network (generator neural network) includes a mapping network and a synthesis network. During prediction, the mapping network may be omitted, replicated, or evaluated several times. The synthesis network may be used to generate highly varied, high-quality output data with a wide variety of attributes. For example, when used to generate images of people's faces, the attributes that may vary are age, ethnicity, camera viewpoint, pose, face shape, eyeglasses, colors (eyes, hair, etc.), hair style, lighting, background, etc. Depending on the task, generated output data may include images, audio, video, three-dimensional (3D) objects, text, etc.

    FUSED PROCESSING OF A CONTINUOUS MATHEMATICAL OPERATOR
    5.
    发明申请
    FUSED PROCESSING OF A CONTINUOUS MATHEMATICAL OPERATOR 有权

    公开(公告)号：US20220405980A1

    公开(公告)日：2022-12-22

    申请号：US17562521

    申请日：2021-12-27

    Applicant: NVIDIA Corporation

    Inventor： Tero Tapani Karras ,  Miika Samuli Aittala ,  Samuli Matias Laine ,  Erik Andreas Härkönen ,  Janne Johannes Hellsten ,  Jaakko T. Lehtinen ,  Timo Oskari Aila

    IPC: G06T9/00 ,  G06F13/16

    Abstract: Systems and methods are disclosed for fused processing of a continuous mathematical operator. Fused processing of continuous mathematical operations, such as pointwise non-linear functions without storing intermediate results to memory improves performance when the memory bus bandwidth is limited. In an embodiment, a continuous mathematical operation including at least two of convolution, upsampling, pointwise non-linear function, and downsampling is executed to process input data and generate alias-free output data. In an embodiment, the input data is spatially tiled for processing in parallel such that the intermediate results generated during processing of the input data for each tile may be stored in a shared memory within the processor. Storing the intermediate data in the shared memory improves performance compared with storing the intermediate data to the external memory and loading the intermediate data from the external memory.

    Fused processing of a continuous mathematical operator
    6.
    发明授权
    Fused processing of a continuous mathematical operator 有权

    公开(公告)号：US12142016B2

    公开(公告)日：2024-11-12

    申请号：US17562521

    申请日：2021-12-27

    Applicant: NVIDIA Corporation

    Inventor： Tero Tapani Karras ,  Miika Samuli Aittala ,  Samuli Matias Laine ,  Erik Andreas Härkönen ,  Janne Johannes Hellsten ,  Jaakko T. Lehtinen ,  Timo Oskari Aila

    IPC: G06T9/00 ,  G06F13/16

    Abstract: Systems and methods are disclosed for fused processing of a continuous mathematical operator. Fused processing of continuous mathematical operations, such as pointwise non-linear functions without storing intermediate results to memory improves performance when the memory bus bandwidth is limited. In an embodiment, a continuous mathematical operation including at least two of convolution, upsampling, pointwise non-linear function, and downsampling is executed to process input data and generate alias-free output data. In an embodiment, the input data is spatially tiled for processing in parallel such that the intermediate results generated during processing of the input data for each tile may be stored in a shared memory within the processor. Storing the intermediate data in the shared memory improves performance compared with storing the intermediate data to the external memory and loading the intermediate data from the external memory.

    TRAINING NEURAL NETWORKS WITH LIMITED DATA USING INVERTIBLE AUGMENTATION OPERATORS
    8.
    发明申请
    TRAINING NEURAL NETWORKS WITH LIMITED DATA USING INVERTIBLE AUGMENTATION OPERATORS 有权

    公开(公告)号：US20210383241A1

    公开(公告)日：2021-12-09

    申请号：US17210934

    申请日：2021-03-24

    Applicant: NVIDIA Corporation

    Inventor： Tero Tapani Karras ,  Miika Samuli Aittala ,  Janne Johannes Hellsten ,  Samuli Matias Laine ,  Jaakko T. Lehtinen ,  Timo Oskari Aila

    IPC: G06N3/08 ,  G06N3/04

    Abstract: Embodiments of the present disclosure relate to a technique for training neural networks, such as a generative adversarial neural network (GAN), using a limited amount of data. Training GANs using too little example data typically leads to discriminator overfitting, causing training to diverge and produce poor results. An adaptive discriminator augmentation mechanism is used that significantly stabilizes training with limited data providing the ability to train high-quality GANs. An augmentation operator is applied to the distribution of inputs to a discriminator used to train a generator, representing a transformation that is invertible to ensure there is no leakage of the augmentations into the images generated by the generator. Reducing the amount of training data that is needed to achieve convergence has the potential to considerably help many applications and may the increase use of generative models in fields such as medicine.

Patent Agency Ranking