Abstract: Provided is a coil assembly and a coil assembly and a coil module. The coil assembly includes a coil, a supporting member, and a circuit board. The supporting member includes a supporting ring and an extending board. The circuit board is arranged on an inner side of the supporting ring. The extending board includes a first end connected to the supporting ring and a second end extends towards the circuit board. The coil is arranged on the supporting ring and has a first end segment and a second end segment which are arranged on the extending board and extend to the circuit board. A first end of the coil is connected to a first pad of the circuit board, and a second end of the coil is connected to a second pad of the circuit board.

Abstract: A model fusion method includes: calling a main process during a pre-training process of a large model, to cache intermediate model parameters obtained during the pre-training process into a main buffer; and calling a sub-process via the main process to read the intermediate model parameters from the main buffer and perform a parameter fusion process based on the intermediate model parameters.

Abstract: A data processing method, and a data processing model and a training method therefor are provided, and relate to the field of artificial intelligence, and specifically, to natural language processing, deep learning technologies, and large model technologies. An implementation solution includes: determining input data, where the input data includes a plurality of tokens; determining a correlation between each of the plurality of tokens and each of a plurality of expert networks based on a gating matrix, where the plurality of expert networks are used to reinforce the plurality of tokens; allocating the plurality of tokens to the plurality of expert networks in a uniform manner based on the correlation and a preset capacity of each expert network, to reinforce the plurality of tokens; and determining a data processing result based on the plurality of reinforced tokens.

Abstract: A cable assembly includes a number of wires, a covering layer, a braiding layer and a fixing block. Each wire includes a core for conducting electricity and an insulating layer wrapped on the core. The covering layer is wrapped on the plurality of wires. The braiding layer is wrapped on the covering layer. The fixing block is injection-molded on the braiding layer. A material of the fixing block at least partially penetrates into the braiding layer so as to integrate the braiding layer and the covering layer as a whole. With such arrangement, the structural reliability of the cable assembly of the present disclosure is relatively high. A cable connector having the cable assembly and a method of manufacturing the cable assembly are also disclosed.

Abstract: A wearable device, where the wearable device includes at least two working modes. The at least two working modes include a first working mode and a second working mode. The wearable device receives a user operation, the wearable device starts the first working mode in response to a first operation of a user to collect first activity data, where the first activity data is daily activity data of the user, and the wearable device starts the second working mode in response to a second operation of the user to collect second activity data, where the second activity data is professional exercise data of the user.

Abstract: A task execution method for a large model, an electronic device, and a storage medium are provided, which relate to a field of artificial intelligence technology, particularly to fields of deep learning technology and large model technology.

Abstract: A cable assembly includes a number of wires, a covering layer, a braiding layer and a fixing block. Each wire includes a core for conducting electricity and an insulating layer wrapped on the core. The covering layer is wrapped on the plurality of wires. The braiding layer is wrapped on the covering layer. The fixing block is injection-molded on the braiding layer. A material of the fixing block at least partially penetrates into the braiding layer so as to integrate the braiding layer and the covering layer as a whole. With such arrangement, the structural reliability of the cable assembly of the present disclosure is relatively high. A cable connector having the cable assembly and a method of manufacturing the cable assembly are also disclosed.

Abstract: A data processing method, and a data processing model and a training method therefor are provided, and relate to the field of artificial intelligence, and specifically, to natural language processing, deep learning technologies, and large model technologies. An implementation solution includes: determining input data, where the input data includes a plurality of tokens; determining a correlation between each of the plurality of tokens and each of a plurality of expert networks based on a gating matrix, where the plurality of expert networks are used to reinforce the plurality of tokens; allocating the plurality of tokens to the plurality of expert networks in a uniform manner based on the correlation and a preset capacity of each expert network, to reinforce the plurality of tokens; and determining a data processing result based on the plurality of reinforced tokens.

Abstract: A wearable device, where the wearable device includes at least two working modes. The at least two working modes include a first working mode and a second working mode. The wearable device receives a user operation, the wearable device starts the first working mode in response to a first operation of a user to collect first activity data, where the first activity data is daily activity data of the user, and the wearable device starts the second working mode in response to a second operation of the user to collect second activity data, where the second activity data is professional exercise data of the user.

Abstract: A communications system, a method for managing a communications system, and a controller, where the communications system includes a front-end device and a remote device, where the remote device includes at least one coaxial media converter (CMC), where each CMC includes a data over cable service interface specification (DOCSIS) processor supporting a broadband access service and an edge quadrature amplitude modulation (EQAM) processor supporting a video service, and the at least one CMC and the front-end device are connected using a digital fiber, and the front-end device includes a controller, where the controller is configured to manage the at least one CMC.

Abstract: Embodiments of the present invention disclose a method, relay agent and system for acquiring an IP address in a network. According to the method, the relay agent configures multiple gateway IP addresses corresponding to a type of a client, adds the multiple gateway IP addresses to a DHCP Discover packet of the client, and sends the DHCP Discover packet to a DHCP server, so that the DHCP server allocates an IP address to the client, and sends a DHCP Offer packet to the client after determining that one gateway IP address in the multiple gateway IP addresses belongs to a same network segment as the allocated IP address, so that the DHCP server allocates the IP address to the client, thereby implementing that multiple IPs in different IP address segments are allocated to a same DHCP client type.

Abstract: A communications system, a method for managing a communications system, and a controller, where the communications system includes a front-end device and a remote device, where the remote device includes at least one coaxial media converter (CMC), where each CMC includes a data over cable service interface specification (DOCSIS) processor supporting a broadband access service and an edge quadrature amplitude modulation (EQAM) processor supporting a video service, and the at least one CMC and the front-end device are connected using a digital fiber, and the front-end device includes a controller, where the controller is configured to manage the at least one CMC.

Abstract: Embodiments of the present disclosure provide a service provisioning method in a coaxial cable system, the method comprises: a CMC receives a DHCP Ack message from a DHCP server and replaces an address of a TFTP server in the DHCP Ack message with an IP address of the CMC; the CMC sends the DHCP Ack message carrying the IP address of the CMC to the CM; and the CMC receives a request message from the CM for delivering a configuration file, acquires a configuration file corresponding to the CM, and after converting the configuration file corresponding to the CM into a format identifiable by the CM, sends the configuration file to the CM.

Abstract: Embodiments of the present invention disclose a method, relay agent and system for acquiring an IP address in a network. According to the method, the relay agent configures multiple gateway IP addresses corresponding to a type of a client, adds the multiple gateway IP addresses to a DHCP Discover packet of the client, and sends the DHCP Discover packet to a DHCP server, so that the DHCP server allocates an IP address to the client, and sends a DHCP Offer packet to the client after determining that one gateway IP address in the multiple gateway IP addresses belongs to a same network segment as the allocated IP address, so that the DHCP server allocates the IP address to the client, thereby implementing that multiple IPs in different IP address segments are allocated to a same DHCP client type.

Abstract: Embodiments of the present disclosure provide a service provisioning method in a coaxial cable system, the method comprises: a CMC receives a DHCP Ack message from a DHCP server and replaces an address of a TFTP server in the DHCP Ack message with an IP address of the CMC; the CMC sends the DHCP Ack message carrying the IP address of the CMC to the CM; and the CMC receives a request message from the CM for delivering a configuration file, acquires a configuration file corresponding to the CM, and after converting the configuration file corresponding to the CM into a format identifiable by the CM, sends the configuration file to the CM.

Abstract: The present invention provides a Layer 2 access method, device and system on a hybrid fiber-coaxial network, where the method includes: receiving a DHCP packet of a terminal device, extracting a source MAC address and a device type from the DHCP packet, and generating a correspondence between the source MAC address and an SVLAN according to a correspondence between the device type and the SVLAN; and when a data packet of the terminal device is received, extracting the source MAC address from the data packet, obtaining the corresponding SVLAN according to the stored correspondence between the MAC address and the SVLAN, and forwarding the data packet of the terminal device according to the corresponding SVLAN.

Abstract: A method and apparatus is provided for preventing the counterfeiting of a network-side Media Access Control (MAC) address. The method includes: receiving a message from a user equipment (UE) and resolving the message from the UE in order to obtain a MAC address of the UE; learning the MAC address of the UE if the MAC address of the UE is different from a known MAC address of a network-side equipment; learning the MAC address of the network-side equipment; and generating a MAC address learning table by utilizing the learned MAC address of the network-side equipment and setting the MAC address learning table to be a static address table, and/or filtering messages having source MAC addresses being the MAC address of the network-side equipment and from other user-side ports by utilizing the learned MAC address of the network-side equipment.

Abstract: Embodiments of the present invention provide a method and a system for implementing multicast using a slave block in DSL access technique. Several slave blocks are cascaded with a master block accessed to the DSL. The master block and slave block are respectively connected with DSL users via RTUs, and access an uplink via a network interface of the master block. The slave block is configured to support the IGMP proxy or the IGMP snooping so as to implement multicast function independently. This invention overcomes the shortcomings including the heavy burden on the master block, low efficiency and serious waste in bandwidth resulting from the fact that the operation of the slave block multicast member is implemented in the master block, thus increasing the efficiency of multicast and saving bandwidth of the master block.

Abstract: A method and apparatus is provided for preventing the counterfeiting of a network-side Media Access Control (MAC) address. The method includes: receiving a message from a user equipment (UE) and resolving the message from the UE in order to obtain a MAC address of the UE; learning the MAC address of the UE if the MAC address of the UE is different from a known MAC address of a network-side equipment; learning the MAC address of the network-side equipment; and generating a MAC address learning table by utilizing the learned MAC address of the network-side equipment and setting the MAC address learning table to be a static address table, and/or filtering messages having source MAC addresses being the MAC address of the network-side equipment and from other user-side ports by utilizing the learned MAC address of the network-side equipment.

Abstract: Embodiments of the present invention provide a method and a system for implementing multicast using a slave block in DSL access technique. Several slave blocks are cascaded with a master block accessed to the DSL. The master block and slave block are respectively connected with DSL users via RTUs, and access an uplink via a network interface of the master block. The slave block is configured to support the IGMP proxy or the IGMP snooping so as to implement multicast function independently. This invention overcomes the shortcomings including the heavy burden on the master block, low efficiency and serious waste in bandwidth resulting from the fact that the operation of the slave block multicast member is implemented in the master block, thus increasing the efficiency of multicast and saving bandwidth of the master block.