Abstract: A method and apparatus for audio processing, an electronic device and a storage medium are provided. The method includes: obtaining an audio encoding result, wherein each element in the audio encoding result has a coordinate in an audio frame number dimension and a coordinate in a text label sequence dimension; in response to an output result of an ith frame in a decoding path being a non-null character, respectively increasing the coordinate in the audio frame number dimension and the coordinate in the text label sequence dimension corresponding to an output position of the ith frame by 1 to obtain an output position of a (i+1)th frame in the decoding path; and determining an output result corresponding to the output position of the (i+1)th frame according to the output result of the ith frame in the decoding path and an element of the (i+1)th frame in the audio encoding result.

Abstract: A method of training a speech recognition model is provided. The method includes that: speech data of each of a plurality of training samples is inputted into a teacher model and a to-be-trained speech recognition model separately. Additionally, an embedding outputted by the teacher model and encoded data outputted by the to-be-trained speech recognition model are obtained. Furthermore, quantized codebook data is obtained by performing a multi-codebook quantization on the embedding. A loss is calculated based on the encoded data, the quantized codebook data, and text data in the training sample. Moreover, a trained speech recognition model is obtained by stopping training the to-be-trained speech recognition model when the loss is less than or equal to a preset loss threshold and/or trained times is greater than preset trained times.

Abstract: A pixel structure includes a substrate body having a light entrance surface, a plurality of first photodiodes formed in the substrate body at a first depth with respect to the light entrance surface, and a second photodiode formed in the substrate body at a second depth with respect to the light entrance surface. The first depth corresponds to a photon absorption length in a material of the substrate body at a first wavelength range, and the second depth corresponds to a photon absorption length in the material of the substrate body at a second wavelength range that is different from the first wavelength range.

Abstract: In a method for image transmitting executed in a transmitting device, three data transmitting channels are established, the three data transmitting channels are a first channel, a second channel and a third channel. An image of a video is obtained, and the image is divided into a region of interest and a background region. A first data of the region of interest and a second data of the background region are obtained, and the first data is encoded through fountain coding to obtain a third data. The first data, the second data, and the third data are respectively transmitted through the first channel, the second channel, and the third channel to a receiving device. A network condition is received, and whether the network condition matches a preset condition is determined. When the network condition matches the preset condition, the first data is compensated according to a first preset algorithm.

Abstract: This application provides a login authentication method, apparatus, and system, and relates to the field of communication technologies. The method includes: a first terminal obtains a short-distance communication identifier of a second terminal; when the second terminal is determined to be a trusted device based on the short-distance communication identifier, a second token of the second terminal is obtained, and validity of a first token and the second token is verified by using a server; and when the verification of both the first token and the second token succeeds, a target account is automatically logged into, where the second token is applied for by the second terminal from the server by logging into the target account in advance.

Abstract: A narrow border reflective display device includes an driving circuit substrate, a TFT substrate, a front plane laminate, multiple conductive wires, a cover, and a glue. The TFT substrate is located on the driving circuit substrate. The TFT substrate is located between the driving circuit substrate and the front plane laminate. The conductive wires are electrically connected with the driving circuit substrate and the TFT substrate. The cover is located on the front plane laminate. The glue surrounds the driving circuit substrate, the TFT substrate, the front plane laminate, the front plane laminate, and the conductive wires.

Abstract: A feature map caching method of a convolutional neural network includes a connection analyzing step and a plurality of layer operation steps. The connection analyzing step is for analyzing a network to establish a convolutional neural network connection list. The convolutional neural network connection list includes a plurality of tensors and a plurality of layer operation coefficients. Each of the layer operation coefficients includes a step index, at least one input operand label and an output operand label. The step index as a processing order for the layer operation step. At least one of the layer operation steps is for flushing at least one of the tensors in a cache according to a distance between the at least one of the layer operation steps and a future layer operation step of the layer operation steps. The distance is calculated according to the convolutional neural network connection list.

Abstract: An integrated latch apparatus for a securing and releasing computer equipment from a chassis includes a support bracket, a rotating shaft, a sliding plate having a latching extension, and a handle having a rotating end. The handle is secured at the rotating end to the support bracket about the rotating shaft. The rotating end of the handle includes nudging extensions. A torsion spring is disposed on the rotating shaft such that the torsion spring biases the support bracket against the handle such that the handle is in an open position when the torsion spring is extended towards an equilibrium state and in a closed position when the torsion spring is biased toward a fully compressed state. The nudging extensions engage and laterally move a sliding plate when the handle is rotated between the open position and the closed position.

Abstract: A ventilator mask and a joint thereof are provided. The ventilator mask includes a ventilator mask body and the joint. The joint includes a joint body, an upper cover, and a perforated cover. The joint body includes a first end portion, a second end portion, and a third end portion. The first end portion is configured to be movably connected to the ventilator mask body. The third end portion is configured to communicate with an oxygen source. An opening is disposed between the first and second end portions. The upper cover is pivotally connected to the joint body for closing or opening the opening. The perforated cover is disposed on the second end portion. The perforated cover has a through hole for insertion of a tube. The ventilator mask can be airtightly fitted to the patient’s face. There is no need to remove the ventilator mask for performing sputum suction.

Abstract: Aiming at insufficient drying seafood on-board caused by hull swings, the present invention involves the establishment of location correction system for processing seafood transportation displaced by wind waves and anti-accumulation drying processing method. This invention employs CFD-DEM method to simulate the state and distribution of the material particles modulated by the deflector angle and wind speed. Finally, the optimized rotation angle of deflector and wind speed are obtained where material particles are equally distributed. Meanwhile, the uniform and fast drying of the marine products in the swinging hull are achieved. This invention shows the great advantages of high efficiency, automation and continuity.

Abstract: A computing system and a compressing method for neural network parameters are provided. In the method, multiple neural network parameters are obtained. The neural network parameters are used for a neural network algorithm. Every at least two neural network parameters are grouped into an encoding combination. The number of neural network parameters in each encoding combination is the same. The encoding combinations are compressed with the same compression target bit number. Each encoding combination is compressed independently. The compression target bit number is not larger than a bit number of each encoding combination. Thereby, the storage space can be saved and excessive power consumption for accessing the parameters can be prevented.

Abstract: In a method for image transmitting executed in a transmitting device, three data transmitting channels are established, the three data transmitting channels are a first channel, a second channel and a third channel. An image of a video is obtained, and the image is divided into a region of interest and a background region. A first data of the region of interest and a second data of the background region are obtained, and the first data is encoded through fountain coding to obtain a third data. The first data, the second data, and the third data are respectively transmitted through the first channel, the second channel, and the third channel to a receiving device. A network condition is received, and whether the network condition matches a preset condition is determined. When the network condition matches the preset condition, the first data is compensated according to a first preset algorithm.

Abstract: A front plate laminate structure includes a display medium layer, a top adhesive layer, a transparent substrate, a transparent conductive film, and a color filter layer. The top adhesive layer is located on the display medium layer. The transparent substrate is located on the top adhesive layer. The transparent conductive film is located between the transparent substrate and the top adhesive layer. The transparent conductive film includes a bottom surface facing the top adhesive layer. The color filter layer is located between the transparent substrate and the display medium layer.

Abstract: In a device with adjusted pause-type hinge, a holding portion and a fixing axis are combined to form a pause-type hinge, the holding portion has a shaft joint portion and an extending portion, the fixing axis passes through the shaft joint portion and the extending portion in shaft manner for rotating and combined to the holding portion, the holding portion has a protruding portion, the fixing axis has at least one fixing hole, and the protruding portion is elastically stretched and engaged with the fixing hole to form a pausing position of a corresponding angle.

Abstract: In a method for image transmitting executed in a transmitting device, three data transmitting channels are established, the three data transmitting channels are a first channel, a second channel and a third channel. An image of a video is obtained, and the image is divided into a region of interest and a background region. A first data of the region of interest and a second data of the background region are obtained, and the first data is encoded through fountain coding to obtain a third data. The first data, the second data, and the third data are respectively transmitted through the first channel, the second channel, and the third channel to a receiving device. A network condition is received, and whether the network condition matches a preset condition is determined. When the network condition matches the preset condition, the first data is compensated according to a first preset algorithm.

Abstract: A matrix multiplier and an operation method thereof are provided. The matrix multiplier includes a plurality of first input lines, a plurality of second input lines and a computing array. The computing array includes a plurality of multiplication accumulation (MAC) cells. A first MAC cell of the plurality of MAC cells is coupled to a first corresponding input line of the plurality of first input lines and a second corresponding input line of the plurality of second input lines to receive a first input value and a second input value to perform a multiplication accumulation operation. When at least one of the first input value and the second input value is a specified value, the multiplication accumulation operation of the first MAC cell is disabled.