Aac Microtech (changzhou) Co, , Ltd.

Tencent Technology (shenzhen) Company Limited

Aac Technologies Pte. Ltd.

Aac Acoustic Technologies (shenzhen) Co., Ltd.

Huawei Technologies Co., Limited

Shenzhen Tatfook Wisdom Health Technology Co., Ltd.

Honor Device Co., Ltd.

Tong Zhang

Bo Xiao

Ronglin Linghu

Chenliang Kong

Zhaopeng Tu

Yan Zhang

Baoyuan Wu

Shuming Shi

Yanbo Fan

Weidong Chen

Wei Song

Yong Zhang

Peng Liu

Jian Li

Yu Luo

Zequn Jie

Tiecai Li

Guoming Zhou

Xiang Kong

Baosong Yang

Chengming Cao

Fan Jiang

Wenjing Li

Qiyun Yuan

Chunyan Xu

Li Wei

Zhen Cui

Zhao Peng Tu

Bao Song Yang

Xianrui Liang

Jianhao Liu

Libao Zhang

Yating Li

Jihai Che

A neural network training method, apparatus, a storage medium, and a computer device are provided. The method includes: obtaining a training sample set, each training sample including a standard label; inputting the each training sample into a neural network model including n attention networks, the n attention networks respectively mapping the each training sample to n subspaces, each of the n subspaces including a query vector sequence, a key vector sequence, and a value vector sequence; calculating a space difference degree between the n subspaces by using the neural network model; calculating an output similarity degree according to an output of the neural network model and the standard label corresponding to the each training sample; and adjusting a model parameter of the neural network model according to the space difference degree and the output similarity degree until a convergence condition is satisfied to obtain a target neural network model.

Neural network training method and apparatus, computer device, and storage medium

This application discloses a pair of glasses having a wireless charging receiver circuit, a glasses case for glasses, and a system, and relates to the field of terminal technologies. The glasses may automatically perform hardware reset without disposing a reset button. The glasses case includes a housing having a cavity configured to accommodate the glasses, a processor, and a wireless charging transmitter circuit coupled to the wireless charging receiver circuit. The processor is configured to determine that the glasses need to perform power-on reset; and the wireless charging transmitter circuit is configured to send a first signal to the glasses in response to a case in which the processor determines that the glasses need to perform power-on reset, to trigger at least one processor of the glasses to perform power-on reset.

Glasses having wireless charging receiver circuit, glasses case for glasses, and system

A model training method and apparatus for image recognition, and a non-transitory storage medium are provided. The model training method includes: obtaining a multi-label image training set including a plurality of training images each annotated with a plurality of sample labels; selecting target training images from the multi-label image training set for training a current model; performing label prediction on each target training image using the current model, to obtain a plurality of predicted labels of the each target training image; obtaining a cross-entropy loss function corresponding to the plurality of sample labels of the each target training image, a positive label loss being greater than a negative label loss and having a weight greater than 1; converging the predicted labels and the sample labels of the each target training image according to the cross-entropy loss function, and updating parameters of the current model, to obtain a trained model.

Model training method and apparatus for image recognition, network device, and storage medium

Provided is a multifunctional sounding device and an electronic device. The multifunctional sounding device includes a casing, a sounding body accommodated in the casing, a vibrator assembly and an elastic supporting member for suspending the vibrator assembly in the accommodating space and providing elastic restoring force. The sounding body includes a vibration system and a magnetic circuit system for driving the vibration system to sound. The elastic supporting member includes a first supporting arm fixed to the vibrator assembly, bent portions bent and extended from both ends of the first supporting arm, and second supporting arms extended away from the first supporting arm and both fixed to the casing. A thickening member is arranged on each bent portion, and a sum of a thicknesses of each bent portion and the thickening member is greater than a thicknesses of the first supporting arm and the second supporting arm.

Multifunctional sounding device and electronic device

A video target tracking method is provided to a computing device, the method includes: obtaining a partial detection map corresponding to a target image frame in a to-be-detected video; obtaining a relative motion saliency map corresponding to the target image frame; determining constraint information corresponding to the target image frame according to the partial detection map and the relative motion saliency map; adjusting a parameter of an image segmentation model by using the constraint information, to obtain an adjusted image segmentation model; and extracting a target object from the target image frame by using the adjusted image segmentation model.

Video target tracking method and apparatus, computer device, and storage medium

The present invention provides a multifunctional sounding device including a housing, a first vibration system, a second vibration system, a magnetic circuit system, an elastic component and a flexible circuit board. The elastic component includes a deformation part fixed to the second vibration system and an installation part extended from both ends of the deformation part and fixed to the housing. The second vibration system is suspended in the containment cavity through the elastic component. The flexible circuit board is connected with the elastic component, so that the vibration of the flexible circuit board is coupled with the vibration of the elastic component. Therefore, the damping of the elastic component is reduced, the force received by the elastic component during vibration is reduced, and the service life of the elastic component is extended.

Multifunctional sounding device

The application discloses a vibration sounding device including a housing body with a containment cavity, a sounding unit placed in the containment cavity, a weight, a drive coil and a spring plate. The spring plate includes a first fixed section for fixed connection with the weight and a second fixed section for fixed connection with the housing body. The first fixed section and the second fixed section are fixed by laser spot welding. Compared with the prior art, the first fixed section and the second fixed section are fixed by laser spot welding, which can improve the bonding strength between the first fixed section and the second fixed section, and ensures the strength of the spring plate when the vibration sounding device is working.

Vibration sounding device

The application relates to a vibration sounding device including: a housing body with a containment cavity and a sounding unit arranged in the containment cavity. The sounding unit includes a vibration system and a magnetic circuit system. The magnetic circuit system includes a main magnet and a secondary magnet assembly. The vibration sounding device also includes an elastic assembly, a weight assembly, and a drive coil. The second direction is perpendicular to the first direction. By virtue of the configuration, it is beneficial to simplify the forming process of the weight assembly, reduce the production cost, and also facilitate the replacement of the drive coil.

A pulse diagnostic device and a system of pulse diagnosis are provided. The pulse diagnostic device includes a cuff and a main monitor. The cuff includes a tube and an air bag arranged with an air path interface. A gas medium is received in the air bag. The tube is connected to the air path interface. The main monitor includes a pressure sensor and a controller. The tube extends to reach the main monitor and is connected to the pressure sensor, and the pressure sensor further connects to the controller. When the pulse diagnostic device is working, the cuff contacts an artery, the pressure sensor senses a pressure of the gas medium in the tube. The pressure sensor transmits the pressure of the gas medium to the controller. The controller obtains the pulse information and the blood pressure information based on the pressure of the gas medium.

Pulse diagnostic device and system of pulse diagnosis

The present disclosure describes a method, apparatus, and storage medium for generating network representation for a neural network. The method includes obtaining a source-side vector sequence corresponding to an input sequence. The method further includes performing linear transformation on the source-side vector sequence, to obtain a request vector sequence, a key vector sequence, and a value vector sequence corresponding to the source-side vector sequence; calculating a logical similarity between the request vector sequence and the key vector sequence; constructing a local strength matrix according to the request vector sequence; performing nonlinear transformation based on the logical similarity and the local strength matrix, to obtain a local strength attention weight distribution corresponding to elements in the input sequence; and fusing value vectors in the value vector sequence according to the local strength attention weight distribution, to obtain a network representation sequence corresponding to the input sequence.

Method, apparatus, and storage medium for generating network representation for neural network

The present disclosure provides a speaker including a housing and a vibration system and a magnetic circuit unit that are received in the housing. The magnetic circuit unit includes lower clamp plate and magnetic circuit assembly arranged on lower clamp plate. A central area of lower clamp plate abuts against bottom wall of the housing. An edge area of lower clamp plate is spaced from the bottom wall. The edge area is provided with flange portion bending and extending from lower clamp plate toward the bottom wall. The flange portion abuts against the bottom wall. The flange portion abuts against the bottom wall. In this way, when speaker falls off, flange portions come in contact with bottom wall, such that stability of magnetic circuit unit can be improved, then operational performance and service life can be improved.

Speaker

A method of establishing an image set for image recognition includes obtaining a single-label image set comprising an image annotated with a single label, and a multi-label image set comprising an image annotated with a plurality of labels; converting content of each label into a corresponding word identifier according to a semantic network, to obtain a word identifier set, a converted single-label image set, and a converted multi-label image set; and constructing a hierarchical semantic structure according to the word identifier set and the semantic network. The method also includes performing label supplementation on the image in the converted single-label image set to obtain a supplemented single-label image set; performing label supplementation on the supplemented single-label image set to obtain a final supplemented image set; and establishing a target multi-label image set to train an image recognition model by using the target multi-label image set.

Method and apparatus for establishing image set for image recognition, network device, and storage medium

Provided is a speaker, including a frame; a vibration unit fixed to the frame; and a magnetic circuit unit driving the vibration unit to vibrate and produce sound. The vibration unit includes a diaphragm and a voice coil; the diaphragm includes a first suspension, a second suspension and a vibrating portion; the magnetic circuit unit is provided with a magnetic gap, and the voice coil is inserted in the magnetic gap; the magnetic circuit unit includes a yoke; and the yoke and the diaphragm are arranged at two opposite sides of the frame; and the vibration unit further includes a fixing ring connected to the magnetic circuit unit; the second suspension surrounds the fixing ring; and the inner peripheral side of the second suspension is fixed to an end of the fixing ring close to the yoke. With this structure, the acoustic performance of the speaker is improved.

Provided is a sounding device, including frame and vibration system and magnetic circuit unit fixed to the frame. The vibration system includes diaphragm, voice coil driving the diaphragm to vibrate and sound, holder fixed to diaphragm, and auxiliary diaphragm spaced from diaphragm. The holder includes holder main body, first connection portion formed from holder main body to magnetic circuit unit and second connection portion spaced from first connection portion. The voice coil is fixed to second connection portion. One end of the auxiliary diaphragm is fixed to frame. One end of first connection portion away from holder main body bends and extends away from second connection portion to form flange. One end of auxiliary diaphragm away from frame is fixed to flange. The sounding device may reduce accuracy deviation and hidden dangers of gluing failure caused by rear assembling of the auxiliary diaphragm.

Growing community of inventors

Shenzhen, China

Tong Zhang