Approximate data bus inversion technique for latency sensitive applications

    公开(公告)号:US12117960B2

    公开(公告)日:2024-10-15

    申请号:US17029288

    申请日:2020-09-23

    CPC classification number: G06F13/4291 G06F13/4068 G06F13/423

    Abstract: Systems, methods, and apparatuses associated with an approximate majority based data bus inversion technique are disclosed. A method comprises obtaining, at a first device connected by a plurality of lanes to a second device, original data comprising first bits and second bits, where the first bits are to be transmitted in a new clock cycle via first lanes of the plurality of lanes, and the second bits are to be transmitted in the new clock cycle via second lanes of the plurality of lanes. The method further includes determining whether a first criterion associated with the first bits is met, determining whether a second criterion associated with the second bits is met, and transmitting an inverted version of the original data via the plurality of lanes based, at least in part, on determining that the first criterion and the second criterion are met.

    CONTROL OF INPUT DIMENSIONS FOR COMPUTER VISION MODEL TRAINING

    公开(公告)号:US20240338789A1

    公开(公告)日:2024-10-10

    申请号:US18747926

    申请日:2024-06-19

    CPC classification number: G06T3/40 G06T7/62

    Abstract: Example apparatus disclosed herein determine an initial spatial input size for training a computer vision model, the initial spatial input size based on sizes of input training images, apply an adjustment to the initial spatial input size to determine an adjusted spatial input size, the adjustment based on sizes of objects in the input training images, and map the adjusted spatial input size to one of a set of available spatial input sizes to determine a final spatial input size for training the computer vision model. Some disclosed apparatus evaluates a linear model to determine a final batch size for training the computer vision model, the linear model based on first and second simulations of training the computer vision model, the first simulation based on the final spatial input size and a first batch size, the second simulation based on the final spatial input size and a second batch size.

Patent Agency Ranking