GENERALIZED EVOLUTIONARY TRAINING FRAMEWORKS FOR DEEP NEURAL NETWORKS

    公开(公告)号:US20240403647A1

    公开(公告)日:2024-12-05

    申请号:US18327707

    申请日:2023-06-01

    Abstract: Embodiments of a methodology for generalized evolutionary training of a neural network may comprise: (i) obtaining a set of model snapshots by training a set of input models until at least one snapshot condition is satisfied for each input model from the set of input models, wherein each model snapshot comprises values of model components of its respective input model when the at least one snapshot condition was satisfied; (ii) generating model snapshot evaluation results by evaluating performance of each model snapshot; (iii) based upon the model snapshot evaluation results, selecting one or more parent models from the set of model snapshots; (iv) generating one or more child models by perturbing at least one or more model components of a parent model from the one or more parent models; and (v) setting the one or more child models as the set of input models for a subsequent training iteration.

    DESIGNING HIGH TDP MXM MODULE POWER SOLUTION FOR AUTONOMOUS VEHICLES

    公开(公告)号:US20250123662A1

    公开(公告)日:2025-04-17

    申请号:US18486870

    申请日:2023-10-13

    Abstract: In one embodiment, a power circuit for an autonomous driving vehicle (ADV) includes a first power source having a first power rating. The circuit includes a second power source having a second power rating. The circuit includes a switch coupled to the first and second power sources, the switch being configured to determining a power operating mode of an accelerator device, where the power operating mode is one of a primary mode, a secondary mode, and a dual mode. The circuit includes a first power transistor coupled between the first power source and the accelerator device and a second power transistor coupled between the second power source and the accelerator device, where the first and second power transistors are operable according to the power operating mode.

    HIGH SPEED CAMERA INTERFACE PCB FLOOR PLAN FOR AUTONOMOUS VEHICLES

    公开(公告)号:US20250081336A1

    公开(公告)日:2025-03-06

    申请号:US18026794

    申请日:2022-12-29

    Abstract: A printed circuit board (PCB) includes one or more signal layers in between a first surface and a second surface, the one or more signal layers including a first signal layer. The PCB includes a connector interface disposed on the first surface and a first signal via electrically coupled to the connector interface. The PCB includes first transmission line disposed at the first signal layer electrically coupling a first location to a second location of the first signal layer and a second signal via disposed at the second location. The PCB includes a power-over-cable circuit disposed on the first surface and electrically coupled to the second signal via and a receiver circuit disposed on the first surface adjacent to the power-over-cable circuit. The PCB includes a second transmission line disposed on the first surface layer electrically coupling the power-over-cable circuit to the receiver circuit.

    CPLD FIRMWARE OVER THE AIR UPDATES FOR AUTONOMOUS VEHICLES

    公开(公告)号:US20250021323A1

    公开(公告)日:2025-01-16

    申请号:US18220470

    申请日:2023-07-11

    Inventor: QIANG WANG

    Abstract: In one embodiment, a system determines a version identifier of a first firmware version at a complex programmable logic device (CPLD) executing the first firmware version. The system determines a second firmware version for the CPLD is available to the CPLD via a FOTA process, where a version identifier of the second firmware version is different from the version identifier of the first firmware version. The system performs a write operation to write data blocks corresponding to the second firmware version of the CPLD to a memory address of a data buffer of the CPLD in a block-by-block manner. The system instructs the CPLD to transfer the data blocks corresponding to the second firmware version from the data buffer to a configuration memory of the CPLD. The system causes the second firmware version at the configuration memory of the CPLD to load onto the CPLD.

    GPU STRESS TEST WITH DYNAMIC LOADS
    40.
    发明公开

    公开(公告)号:US20240211363A1

    公开(公告)日:2024-06-27

    申请号:US18069635

    申请日:2022-12-21

    Inventor: Congshi Huang

    CPC classification number: G06F11/263 G06F11/273

    Abstract: A data processing system of an autonomous driving vehicle (ADV) may include a central processing unit (CPU) and a graphics processing unit (GPU). The CPU may be configured to monitor a behavior of a graphics processing unit (GPU), operate the GPU to perform a randomly controlled operation using a random number generator, operate the GPU to perform an idle period, and operate the GPU to repeat the randomly controlled operation. The idle period may be performed between each repetition of the randomly controlled operation, and each repetition may utilize the random number generator.

Patent Agency Ranking