Abstract: A method is presented for mixed-precision matrix multiplication. The method comprises decomposing an n-bit weight matrix into a series of n one-bit matrices. An m-bit activation matrix is received, where m>n. A lookup table of partial matrix multiplication results is generated based on permutations of groups of the m-bit activation matrix and the n one-bit matrices. An input weight matrix is received, the input weight matrix having a depth of g bits. For each bit of the input weight matrix, partial matrix multiplication results are retrieved from the lookup table using g-bit groups of the input weight matrix as indices. The retrieved partial matrix multiplication results are aggregated into a final mixed-precision matrix multiplication result.

Abstract: A communication apparatus and a communication method are disclosed. According to the communication method, a first device sends a block acknowledgment request frame to a second device. The block acknowledgment request frame is used to request the second device to send a block acknowledgment frame indicating a receiving status of an access control layer protocol data unit in first transmission. The first transmission is suspended at a first moment that is earlier than a transmission end moment when the first transmission is not suspended. The first device receives the block acknowledgment frame. The first device starts, at a second moment, to perform second transmission.

Abstract: Embodiments of the present disclosure include systems and methods for providing sparsity for neural network models based on sparsity attributes. A first neural network model definition is received. The first neural network model definition specifies a neural network model comprising a set of tensors and a set of sparsity attribute values for elements of a tensor in the set of tensors. The set of sparsity attribute values for the tensor are propagated to elements of a subset of the set of tensors to form a second neural network model definition. The neural network model is generated based on the second neural network model definition.

Abstract: This application provides a channel sounding method and apparatus. The method includes: A first communication device sends a first frame to a second communication device, where the first frame is used to indicate the second communication device to perform channel sounding on a part of antennas of the first communication device. The first communication device sends a second frame to the second communication device, where the second frame is used by the second communication device to perform channel sounding on the part of antennas. The first communication device receives a third frame from the second communication device, where the third frame is used to indicate a result of performing channel sounding on the part of antennas.

Abstract: Embodiments of this application provide a communication method between multi-link devices and an apparatus, and relate to the field of communications technologies, to reduce power consumption of a station multi-link device. The method includes: a multi-link device includes a plurality of STAs, one STA operates on one of a plurality of links, and the STA determines a primary link in the plurality of links. A STA operating on the primary link in the multi-link device may receive information sent by the AP indicating whether the STA operating on the primary link has a downlink traffic, and/or information indicating whether a STA operating on a link other than the primary link has a downlink traffic. Therefore, the STA operating on the primary link may determine a link on which the STA having a downlink traffic operates in the multi-link device.

Abstract: The present disclosure describes an approach that may be applied to a transmission opportunity (TXOP) of a first access point. The first access point notifies a first station by sending first indication information that the first station can perform reporting if there is currently to-be-transmitted low-latency data. In this way, when the first station has an urgent low-latency data, the first station may have an opportunity to feed back this case to the first access point. Then, the first access point may determine, based on feedback of the first station, whether to enable the first station to perform preemption transmission in a subsequent procedure. The approach can improve performance of a low-latency data and avoid a waste of channel resources.

Abstract: This application provides a channel sounding method and apparatus. The method includes: A first communication device sends a first frame to a second communication device, where the first frame is used to indicate the second communication device to perform channel sounding on a part of antennas of the first communication device. The first communication device sends a second frame to the second communication device, where the second frame is used by the second communication device to perform channel sounding on the part of antennas. The first communication device receives a third frame from the second communication device, where the third frame is used to indicate a result of performing channel sounding on the part of antennas.

Abstract: The embodiments of the application provides a resource allocation method. In an example method, a trigger frame is sent by an access point to a station. The trigger frame includes one user information field, and the user information field may indicate a plurality of resource units allocated to the station. The technical solutions provided in this application may be applied to Wi-Fi systems such as 802.11ax and 802.11be.

Abstract: An electronic expansion valve and a refrigeration device are provided. The electronic expansion valve has a valve seat, a nut and a connection piece. The valve seat is provided with a valve cavity. One end of the nut is installed in the valve cavity. An outer wall of the nut is provided with an annular position part that extends from the outer wall. The position part is positioned to be in contact with a wall of the valve cavity. One end of the connection piece is fixedly connected to the nut. Another end of the connection piece is extended away from the nut and fixedly connected to an axial end surface of the valve seat.

Abstract: A communication apparatus and a communication method are disclosed. According to the communication method, a first device sends a block acknowledgment request frame to a second device. The block acknowledgment request frame is used to request the second device to send a block acknowledgment frame indicating a receiving status of an access control layer protocol data unit in first transmission. The first transmission is suspended at a first moment that is earlier than a transmission end moment when the first transmission is not suspended. The first device receives the block acknowledgment frame. The first device starts, at a second moment, to perform second transmission.

Abstract: According to implementations of the subject matter described herein, there is provided a solution for predicting the resource usage of the deep learning model. In this solution, information about a deep learning model is obtained, the information comprising first information for describing the deep learning model and second information about an operating environment of a job associated with the deep learning model. The static resource usage of the job is determined based on the first information and a strategy of the job during runtime in the operating environment is determined. Afterwards, resource usage of the job during runtime in the operating environment is predicted based on the strategy and the static resource usage. With this solution, the usage of various resources of the deep learning model, such as computation power consumption, memory consumption, execution time, and the like, under a specific runtime strategy can be accurately predicted.

Abstract: Embodiments of the present disclosure include systems and methods for compiling tensor operators for neural network models based on tensor tile configurations. A tensor expression for a neural network model is received. A tensor tile configuration for the tensor expression is determined based on specifications associated with a set of hardware devices. Based on the tensor tile configuration, a set of code for implementing the tensor expression for the neural network model is generated. The set of code is for execution on the set of hardware.

Abstract: A communication method and apparatus are provided, are applied to the field of communication technologies, and may be applied to a wireless local area network system based on 802.11 series protocols such as a next-generation Wi-Fi protocol of the IEEE 802.11ax, for example, 802.11be, Wi-Fi 7, or EHT, and a next-generation protocol of the 802.11be, Wi-Fi 8, UHR, or Wi-Fi AI, and may be further applied to a UWB-supported WPAN system, a sensing system, and the like. The method includes: A first communication apparatus generates a first communication frame, and transmits the first communication frame; and correspondingly, a second communication apparatus receives the first communication frame, and parses the first communication frame. The first communication frame includes duration information and first NSS information, and the first NSS information indicates a maximum number of receive spatial streams of the first communication apparatus within a target time.

Abstract: A ML entity performs a backoff procedure of the primary link based on the backoff counter of the primary link; and when a count value of the backoff counter is 0, the ML entity sends a first PPDU on each first link in K first links, where the K first links include the primary link and K?1 first nonprimary links, and the first nonprimary link is in an idle state in a first inter-frame space before a time point at which the count value of the backoff counter decreases to 0.

Abstract: Flexible scheduling of a slave AP in a unified frame format in a multi-AP coordinated transmission scenario provides an information transmission method. A first access point (AP) generates a slave trigger frame, where the slave trigger frame includes trigger type information and AP indication information. The trigger type information is used to indicate a type of the slave trigger frame. The AP indication information includes an identifier of a second AP. The identifier of the second AP is used to indicate the second AP. The slave trigger frame is used to trigger the second AP to send a physical protocol data unit (PPDU). The PPDU carries a frame corresponding to a type of the slave trigger frame. The first AP sends the slave trigger frame to the second AP.

Abstract: Embodiments of this application disclose a communication method between multi-link devices and an apparatus, and relate to the field of communications technologies, to reduce power consumption of a station multi-link device. A specific solution is as follows: A multi-link device includes a plurality of STAs, one STA operates on one of a plurality of links, and the STA determines a primary link in the plurality of links. A STA operating on the primary link in the multi-link device may receive information sent by the AP indicating whether the STA operating on the primary link has a downlink traffic, and/or information indicating whether a STA operating on a link other than the primary link has a downlink traffic. Therefore, the STA operating on the primary link may determine a link on which the STA having a downlink traffic operates in the multi-link device.

Abstract: A valve needle assembly, an electronic expansion valve and a refrigeration device are provided. The valve needle assembly has a valve needle sleeve, a valve rod, a valve needle, a buffer slider and a buffer spring. The valve needle sleeve has a first open end and a second open end opposite to each other. The valve rod has a driving end movably matched with the valve needle sleeve through the first open end to drive the valve needle provided on the second open end. The buffer slider is provided in the valve needle sleeve and abuts against the valve needle. The buffer spring is provided in the valve needle sleeve, through which the driving end is connected to the buffer slider.

Abstract: Embodiments of this application provide a communication method between multi-link devices and an apparatus, and relate to the field of communications technologies, to reduce power consumption of a station multi-link device. The method includes: a multi-link device includes a plurality of STAs, one STA operates on one of a plurality of links, and the STA determines a primary link in the plurality of links. A STA operating on the primary link in the multi-link device may receive information sent by the AP indicating whether the STA operating on the primary link has a downlink traffic, and/or information indicating whether a STA operating on a link other than the primary link has a downlink traffic. Therefore, the STA operating on the primary link may determine a link on which the STA having a downlink traffic operates in the multi-link device.

Abstract: This application provides a beacon frame transmission control method and a related apparatus. The method includes: generating, by a master access point AP, indication information used to indicate a transmission time period of a beacon frame of at least one secondary AP; and sending, by the master AP, the indication information to the at least one secondary AP, where the master AP and the at least one secondary AP belong to a same AP collaboration group, and transmission time periods of beacon frames of APs in the AP collaboration group are different. According to technical solutions provided in this application, transmission robustness of the beacon frame can be improved.

Abstract: A ML entity performs a backoff procedure of the primary link based on the backoff counter of the primary link; and when a count value of the backoff counter is 0, the ML entity sends a first PPDU on each first link in K first links, where the K first links include the primary link and K?1 first nonprimary links, and the first nonprimary link is in an idle state in a first inter-frame space before a time point at which the count value of the backoff counter decreases to 0.