Abstract: A system obtains a performance profile corresponding to times taken to perform an inferencing by a machine learning (ML) model using a different number of processing resources from a plurality of processing resources. The system determines one or more groupings of processing resources from the plurality of processing resources, each grouping includes one or more partitions. The system calculates performance speeds corresponding to each grouping based on the performance profile. The system determines a grouping having a best performance speed from the calculated performance speeds. The system partitions the processing resources based on the determined grouping to perform the inferencing.

Abstract: A system determines a timing matrix corresponding to inference times taken for a number of machine learning (ML) models to be executed by a number of processing resources of a computing device. The processing resources includes at least a first and a second type of processing resources. The system applies a service-specific model-first scheduling scheme or a service-specific hardware-first scheduling scheme to obtain corresponding service-specific mappings. The system determines a best mapping from the corresponding service-specific mappings. The system schedules each of the ML models to a corresponding processing resource from the processing resources according to the best mapping. The system executes the ML models using corresponding mapped processing resources.

Abstract: The embodiment of the present disclosure provides a cleaning device, which includes a power module, a vacuum cleaner module, and a surface wet cleaner module. The power module includes a power assembly and a motor assembly that are connected to each other. The vacuum cleaner module includes an air suction assembly, a dust collection assembly, and a filter assembly that are sequentially communicated, and the vacuum cleaner module has a first mounting position that is configured to be connected and matched with the power module. The surface wet cleaner module includes a floor brush and a body, and the body has a second mounting position that is configured to be connected and matched with the power module.

Abstract: Modem deep neural network (DNN) models have many layers with a single layer potentially involving large matrix multiplications. Such heavy calculation brings challenges to deploy such DNN models on a single edge device, which has relatively limited computation resources. Therefore, multiple and even heterogeneous edge devices may be required for applications with stringent latency requirements. Disclosed in the present patent documents are embodiments of a model scheduling framework that schedules multiple models on a heterogeneous platform. Two different approaches, model first scheduling (MFS) and hardware first scheduling (HFS), are presented to allocate a group of models for a service into corresponding heterogeneous edge devices, including CPU, VPU and GPU. Experimental results prove the effectiveness of the MFS and HFS methods for improving the inference speed of single and multiple AI-based services.

Abstract: The present disclosure provides eyeglasses including: an eyeglass rim; an eyeglass temple, the eyeglass temple comprising a control circuit or a battery; a rotating shaft, the rotating shaft being configured to connect the eyeglass rim and the eyeglass temple, so that the eyeglass rim and the eyeglass temple are relatively rotated around the rotating shaft, and the rotating shaft being disposed with a rotating shaft wiring channel along an axial direction; and a speaker, the speaker comprising an earphone core, the speaker being connected to the eyeglass temple, the control circuit or battery in the eyeglass temple driving the earphone core to vibrate through the connection wire, wherein the earphone core vibrates to generate a driving force to drive a housing panel of the speaker to vibrate, and a straight line of the driving force being not parallel to a normal line of the housing panel.

Abstract: The present disclosure embodiment may disclose a glasses including a glasses frame and two speakers. The glasses frame may include a glasses rim and two glasses temples. The two glasses temples may be rotatably connected to the glasses rim, respectively. The two speakers may include an earphone core. The two speakers may be connected to the two glasses temples via hinge components of the two glasses temples, respectively. The hinge components may be rotatable to change a position of each of the speakers relative to connected one of the glasses temples. The glasses temple may include a circuit component or a battery to drive the earphone core to vibrate to generate a sound. In the present disclosure, the glasses rim and the glasses temple may be connected through a rotating shaft, thereby protecting the connection wire in the glasses and extending the life of the connection wire.

Abstract: The present disclosure relates to a loudspeaker. The loudspeaker may include a loudspeaker mechanism, a fixing mechanism, and a connector. The loudspeaker mechanism may be configured to generate a vibration signal and transmit the vibration signal to the human body. The fixing mechanism may be configured to support and maintain the position of the loudspeaker mechanism. The connector may be configured to connect the loudspeaker mechanism with the fixing mechanism. The loudspeaker mechanism may at least include a first fixed position and a second fixed position. The first fixed position may be a fixed position of the loudspeaker when the loudspeaker is in a non-working state. The second fixed position may be a fixed position of the loudspeaker when the loudspeaker is in a working state. The connector may be configured to switch the loudspeaker mechanism between the first fixed position and the second fixed position.

Abstract: The present disclosure provides eyeglasses including an eyeglass rim; an eyeglass temple, the eyeglass temple comprising a control circuit or a battery; a rotating shaft, the rotating shaft being configured to connect the eyeglass rim and the eyeglass temple, so that the eyeglass rim and the eyeglass temple are relatively rotated around the rotating shaft, and the rotating shaft being disposed with a rotating shaft wiring channel along an axial direction; a connection wire, the connection wire passing through the rotating shaft wiring channel and extending to the eyeglass rim and the eyeglass temple, respectively; and a speaker, the speaker comprising an earphone core, and a core housing for accommodating the earphone core, the speaker being connected to the eyeglass temple.

Abstract: The present disclosure provides glasses including: a glasses rim; a glasses temple; a rotating shaft, the rotating shaft being configured to connect the glasses rim and the glasses temple, so that the glasses rim and the glasses temple are able to relatively rotate around the rotating shaft, and the rotating shaft being disposed with a rotating shaft wiring channel along an axial direction; a connection wire, the connection wire passing through the rotating shaft wiring channel and extending to the glasses rim and the glasses temple, respectively; and a speaker, the speaker comprising an earphone core, the speaker being connected to the glasses temple, the control circuit or the battery in the glasses temple driving the earphone core to vibrate to generate a sound through the connection wire.

Abstract: A method for 3D object detection is described. The method includes detecting semantic keypoints from monocular images of a video stream capturing a 3D object. The method also includes inferring a 3D bounding box of the 3D object corresponding to the detected semantic vehicle keypoints. The method further includes scoring the inferred 3D bounding box of the 3D object. The method also includes detecting the 3D object according to a final 3D bounding box generated based on the scoring of the inferred 3D bounding box.

Abstract: A method for 2D semantic keypoint detection and tracking is described. The method includes learning embedded descriptors of salient object keypoints detected in previous images according to a descriptor embedding space model. The method also includes predicting, using a shared image encoder backbone, salient object keypoints within a current image of a video stream. The method further includes inferring an object represented by the predicted, salient object keypoints within the current image of the video stream. The method also includes tracking the inferred object by matching embedded descriptors of the predicted, salient object keypoints representing the inferred object within the previous images of the video stream based on the descriptor embedding space model.

Abstract: A method for testing a mechanical performance of a lithium ion battery electrode based on a nano-indentation technology includes following steps: connecting an assembled lithium ion battery with an electrochemical test device and setting different test working conditions, so that cyclic charge and discharge experiments are performed on the battery to obtain an attenuation curve of a battery capacity; disassembling the battery and taking out the electrode; scraping some powder from a surface of the cyclic electrode plate and an initial uncyclic electrode plate, and laying down the powder in cold mounting molds separately, pouring the cold mounting solution into the molds; taking out the samples from the molds respectively after the liquid is completely cured and cooled; detecting a mechanical performance after polishing the samples surfaces and analyzing the mechanical performance decay rule of the electrodes.

Abstract: A method for 3D object detection is described. The method includes detecting semantic keypoints from monocular images of a video stream capturing a 3D object. The method also includes inferring a 3D bounding box of the 3D object corresponding to the detected semantic vehicle keypoints. The method further includes scoring the inferred 3D bounding box of the 3D object. The method also includes detecting the 3D object according to a final 3D bounding box generated based on the scoring of the inferred 3D bounding box.

Abstract: A method for semantic keypoint detection is described. The method includes linking, using a keypoint graph neural network (KGNN), semantic keypoints of an object within a first image of a video stream into a 2D graph structure corresponding to a category of the object. The method also includes embedding descriptors within the semantic keypoints of the 2D graph structure corresponding to the category of the object. The method further includes tracking the object within subsequent images of the video stream using the embedded descriptors within the semantic keypoints of the 2D graph structure corresponding to the category of the object.

Abstract: A method for testing a mechanical performance of a lithium ion battery electrode based on a nano-indentation technology includes following steps: connecting an assembled lithium ion battery with an electrochemical test device and setting different test working conditions, so that cyclic charge and discharge experiments are performed on the battery to obtain an attenuation curve of a battery capacity; disassembling the battery and taking out the electrode; scraping some powder from a surface of the cyclic electrode plate and an initial uncyclic electrode plate, and laying down the powder in cold mounting molds separately, pouring the cold mounting solution into the molds; taking out the samples from the molds respectively after the liquid is completely cured and cooled; detecting a mechanical performance after polishing the samples surfaces and analyzing the mechanical performance decay rule of the electrodes.

Abstract: A joint includes a male module and a female module capable of docking with each other. The male module includes a core rod and a sliding cylinder that is slidably sleeved on an outer periphery of the core rod. A first valve port is disposed in the sliding cylinder, and the core rod is provided with a piston head that is in hermetic fit with the first valve port. The female module includes a valve body and a piston. A first channel is disposed in the valve body, the piston is slidably disposed in the first channel, and a second valve port that is in hermetic fit with the piston is disposed in the first channel.