Communication method and communication apparatus

Abstract

The present application provides a communication method and a communication apparatus. The method is applied to a first communication apparatus and comprises: determining a nested model, wherein the nested model is obtained by connecting a plurality of sub-models, and each sub-model among the plurality of sub-models can be used for implementing the function of the nested model; and sending first indication information to a second communication apparatus, wherein the first indication information indicates the information of the nested model. In this way, the plurality of sub-models can be transmitted together by means of one nested model, thereby reducing the number of independent models in a transmission process, and improving the transmission efficiency. In addition, each sub-model in the nested model can be independently used so as to implement the function of the nested model, and according to actual requirements and communication scenarios, the functions of nested models having different precision can be implemented on the basis of different quantities of functional modules comprised in the sub-models.

Description

A communication method and a communication apparatus

[0001] This application claims priority from the Chinese patent application No. 202411098381.1 filed on August 9, 2024, and entitled "A communication method and a communication apparatus", the content of which is incorporated herein by reference in its entirety. TECHNICAL FIELD

[0002] The present application relates to the field of wireless communication, and more particularly, to a communication method and a communication apparatus. BACKGROUND

[0003] In the intelligent wireless network architecture, new requirements and new scenarios will bring changes to the wireless network architecture and communication mode. In addition, the model structure and the model parameter quantity have a great influence on performance, computation, power consumption, etc. Considering the scenarios of limited computing power and performance changes, models with different structures and parameter quantities are needed. However, the current 3rd generation partnership project (3GPP) only studies the model transmission mode (transmitting the model or the data set), and the transmission efficiency is low. SUMMARY

[0004] The present application provides a communication method and a communication apparatus, which can improve the transmission efficiency by configuring a nested model considering the relationship between different models.

[0005] In a first aspect, a communication method is provided. The method can be applied to a communication apparatus, which can be a communication device (such as a terminal device, a network device, or an artificial intelligence (AI) node), or a component (such as a chip or a chip system or a circuit or a communication module) in a communication device. Hereinafter, the first communication apparatus will be mainly taken as an example for illustration.

[0006] The method can include determining a nested model, the nested model being obtained by connecting a plurality of sub-models, each of the plurality of sub-models being capable of implementing a function of the nested model; and sending first indication information to a second communication apparatus, the first indication information indicating information of the nested model.

[0007] Optionally, the nested model further includes an input adaptation layer and / or an output adaptation layer, wherein the input adaptation layer is connected (such as in series) with at least one of the plurality of sub-models, and the input adaptation layer is used for inputting data; and the output adaptation layer is connected (such as in series) with at least one of the plurality of sub-models, and the output adaptation layer is used for outputting data.

[0008] Optionally, one of the sub-models comprises one or more functional modules.

[0009] Optionally, the information of the nested model comprises information related to the nested model.

[0010] According to the technical solution, the first communication device determines a nested model, wherein the nested model comprises a plurality of sub-models, and sends information of the determined nested model to the second communication device through an indication information (i.e., a first indication information). In this way, the plurality of sub-models can be transmitted together through one nested model, thereby reducing the number of independent models in the transmission process and improving the transmission efficiency. In addition, each sub-model in the nested model can be independently used to realize the function of the nested model, and different precision of the function of the nested model can be realized based on the number of functional modules contained in the sub-models according to actual requirements and communication scenarios.

[0011] With reference to the first aspect, in some implementations of the first aspect, the information of the nested model comprises model segmentation information, and the model segmentation information is used to distinguish one or more of the sub-models in the nested model.

[0012] According to the technical solution, the first communication device can carry the model segmentation information in the information of the nested model transmitted to the second communication device, so that the second communication device can distinguish the sub-models in the nested model based on the model segmentation information.

[0013] With reference to the first aspect, in some implementations of the first aspect, the model segmentation information comprises at least one of the following: a number of layers of the nested model, a number of heads, a linear layer dimension, a hidden layer dimension, and a parameter quantity.

[0014] With reference to the first aspect, in some implementations of the first aspect, the determination of the nested model comprises: receiving first training data; and determining the nested model according to the first training data.

[0015] According to the technical solution, when the first communication device determines the nested model, the nested model can be determined based on the training data sent by the second communication device.

[0016] With reference to the first aspect, in some implementations of the first aspect, the determination of the nested model according to the first training data comprises: determining the nested model according to the first training data and performance information; and the performance information comprises performance information of the first communication device and / or performance information of the second communication device.

[0017] Based on the above technical solution, when the first communication device determines the nested model, the performance information of the first communication device and / or the performance information of the second communication device can also be used to determine the nested model. In this way, the nested model determined by the first communication device can adapt to the performance requirements of the first communication device and / or the second communication device.

[0018] In combination with the first aspect, in some implementations of the first aspect, the method further includes receiving the performance information of the second communication device from the second communication device.

[0019] In combination with the first aspect, in some implementations of the first aspect, the method further includes receiving first request information from the second communication device, the first request information being used to request a first sub-model in the nested model, the first sub-model including one or more of the sub-models; and sending first indication information to the second communication device, the first indication information indicating the parameter information of the nested model, including sending the first indication information to the second communication device based on the first request information, the first indication information indicating the first sub-model.

[0020] Based on the above technical solution, when the first communication device receives the request information (first request information) of the second communication device requesting a specific sub-model (first sub-model) in the nested model, the first communication device sends the parameter information of the specific sub-model (first sub-model) requested by the second communication device to the second communication device through an indication information (i.e., first indication information).

[0021] In combination with the first aspect, in some implementations of the first aspect, the first indication information indicating the first sub-model includes the first indication information indicating at least one of the following: an index, an identifier, and model parameters of the first sub-model.

[0022] In combination with the first aspect, in some implementations of the first aspect, after sending the first indication information to the second communication device, the method further includes performing model inference and / or performance monitoring on the nested model.

[0023] Based on the above technical solution, after the first communication device sends the indication information (first indication information) carrying the information of the nested model to the second communication device, the first communication device can perform model inference and / or performance monitoring on the nested model.

[0024] With reference to the first aspect, in some implementations of the first aspect, the method further includes: updating a second sub-model in the nested model, the second sub-model including one or more of the sub-models, the second sub-model being deployed on the first communication device, the second sub-model being determined based on at least one of: a performance monitoring result of the second sub-model, second indication information from the second communication device, the second indication information indicating to update the second sub-model.

[0025] Based on the above technical solution, after the first communication device determines the nested model and sends information of the relevant nested model to the second communication device, the first communication device determines a sub-model from the nested model again (updates the second sub-model in the nested model) according to at least one of: a performance monitoring result of the determined sub-model (the second sub-model), and indication information (the second indication information) from the second communication device indicating the first communication device to determine a sub-model from the available determined sub-models (the first sub-model). In this way, the first communication device can first determine a sub-model, or determine a sub-model from the sub-model first determined by the second communication device according to the conditions of the first communication device itself.

[0026] With reference to the first aspect, in some implementations of the first aspect, the second indication information indicating to update the second sub-model includes: the second indication information indicating at least one of: an index, an identifier, a model parameter of the first sub-model in the nested model.

[0027] With reference to the first aspect, in some implementations of the first aspect, the method further includes: sending third indication information to the second communication device, the third indication information indicating the second sub-model.

[0028] Based on the above technical solution, the first communication device sends the sub-model (the second sub-model) determined from the nested model to the second communication device through another indication information (the third indication information). In this way, the second communication device can determine a sub-model from the sub-model determined by the first communication device.

[0029] With reference to the first aspect, in some implementations of the first aspect, the third indication information indicating the second sub-model includes: the third indication information indicating at least one of: an index, an identifier, a model parameter of the second sub-model.

[0030] With reference to the first aspect, in some implementations of the first aspect, the performance information includes local computing power and / or performance requirement information.

[0031] In a second aspect, a communication method is provided. The method can be applied to a communication device, which can be a communication apparatus (e.g., a terminal device, a network device, or an AI point) or a component (e.g., a chip or a chip system or a circuit or a communication module) of a communication apparatus. The following mainly takes the second communication device as an example for illustration.

[0032] The method can include: receiving first indication information from a first communication device, the first indication information indicating information of a nested model; and determining the nested model based on the first indication information, the nested model being obtained by connecting a plurality of sub-models, each of the plurality of sub-models being capable of implementing a function of the nested model.

[0033] Based on the above technical solution, the second communication device receives indication information (first indication information) containing information of a nested model from the first communication device, and determines the nested model according to the information of the nested model, wherein the nested model contains a plurality of sub-models, and one sub-model can include one or more functional modules.

[0034] In combination with the second aspect, in some implementations of the second aspect, the information of the nested model includes model segmentation information, the model segmentation information being used to distinguish one or more of the sub-models in the nested model.

[0035] Based on the above technical solution, the first communication device can carry model segmentation information in the information of the nested model transmitted to the second communication device, so that the second communication device can distinguish the sub-models in the nested model based on the model segmentation information.

[0036] In combination with the second aspect, in some implementations of the second aspect, the model segmentation information includes at least one of the following: a number of layers of the nested model, a number of heads, a linear layer dimension, a hidden layer dimension, and a parameter quantity.

[0037] In combination with the second aspect, in some implementations of the second aspect, after receiving the first indication information from the first communication device, the method further includes: performing model inference and / or performance monitoring according to performance information, wherein the performance information includes performance information of the first communication device and / or performance information of the second communication device.

[0038] Based on the above technical solution, when the second communication device performs model inference and / or performance monitoring on the nested model, the model inference and / or performance monitoring on the nested model can also be based on the performance information of the first communication device and / or the performance information of the second communication device.

[0039] With reference to the second aspect, in some implementations of the second aspect, the method further includes sending performance information of the second communication device to the first communication device.

[0040] With reference to the second aspect, in some implementations of the second aspect, the method further includes updating a first sub-model in the nested model, the first sub-model including one or more of the sub-models, the first sub-model being deployed at the second communication device, the first sub-model being determined based on at least one of: a performance monitoring result of the first sub-model, first indication information from the first communication device, third indication information from the first communication device, the third indication information indicating to update the first sub-model.

[0041] Based on the above technical solution, after the second communication device determines the nested model and sends information of the relevant nested model to the first communication device, the second communication device determines a sub-model (updates a first sub-model in the nested model) from the nested model according to at least one of: a performance monitoring result of the determined sub-model (the first sub-model), and indication information (third indication information) from the first communication device indicating the second communication device to determine a sub-model from the available determined sub-models (second sub-models). In this way, the second communication device can first determine a sub-model, or determine a sub-model from a sub-model first determined by the first communication device according to the conditions of the second communication device itself.

[0042] With reference to the second aspect, in some implementations of the second aspect, the third indication information indicates to update the first sub-model, including: the third indication information indicating at least one of: an index, an identifier, a model parameter of a second sub-model in the nested model.

[0043] With reference to the second aspect, in some implementations of the second aspect, the method further includes sending first request information to the first communication device based on the performance information, the first request information being used to request the first sub-model in the nested model.

[0044] Based on the above technical solution, the second communication device can determine a specific sub-model (the first sub-model) in the nested model used by the second communication device according to the performance information of the first communication device and / or the performance information of the second communication device, and request the specific sub-model (the first sub-model) in the nested model in a request information (the first request information) sent to the first communication device.

[0045] With reference to the second aspect, in some implementations of the second aspect, the method further includes sending second indication information to the first communication device, the second indication information indicating the first sub-model.

[0046] Based on the above technical solution, the first communication device sends a sub-model (first sub-model) determined from the nested model to the first communication device through another indication information (second indication information). In this way, the first communication device can determine a sub-model from the sub-model determined by the second communication device.

[0047] In combination with the second aspect, in some implementations of the second aspect, the second indication information indicates the first sub-model, including: the second indication information indicates at least one of the following of the first sub-model: an index, an identifier, a model parameter.

[0048] In combination with the second aspect, in some implementations of the second aspect, the performance information includes local computing power and / or performance requirement information.

[0049] The third aspect provides a communication method. The method can be applied to a communication device, i.e., the communication device can be a communication equipment (such as a terminal equipment, or a network equipment, or an AI point), or the communication device can be a component (such as a chip or a chip system or a circuit or a communication module) in the communication equipment. Hereinafter, the first communication device is mainly taken as an example for illustration.

[0050] The method can include: receiving first training data; determining a nested model according to the first training data, the nested model being obtained by connecting a plurality of sub-models, each of the plurality of sub-models being capable of being used to implement a function of the nested model.

[0051] In combination with the third aspect, in some implementations of the third aspect, the determining the nested model according to the first training data includes: determining the nested model according to the first training data and performance information; wherein the performance information includes performance information of the first communication device and / or performance information of the second communication device.

[0052] In combination with the third aspect, in some implementations of the third aspect, the method further includes: receiving the performance information of the second communication device from the second communication device.

[0053] The fourth aspect provides a communication device for executing the method in any of the first aspect to the third aspect and any possible implementation thereof. Specifically, the device can include units and / or modules for executing the method in any of the first aspect to the third aspect and any possible implementation thereof, such as a processing unit and / or a communication unit.

[0054] In an implementation form, the apparatus is a communication device (e.g., a terminal device, e.g., a network device, e.g., an AI node). When the apparatus is a communication device, the communication unit can be a transceiver, or an input / output interface; the processing unit can be at least one processor. Optionally, the transceiver can be a transceiver circuit. Optionally, the input / output interface can be an input / output circuit.

[0055] In another implementation form, the apparatus is a chip, chip system or circuit or communication module for a communication device (e.g., a terminal device, e.g., a network device, e.g., an AI node). When the apparatus is a chip, chip system or circuit for a communication device, the communication unit can be an input / output interface, interface circuit, output circuit, input circuit, pin or related circuitry, etc. on the chip, chip system or circuit; the processing unit can be at least one processor, processing circuit or logic circuit, etc.

[0056] In a fifth aspect, a communication apparatus is provided, the apparatus comprising: at least one processor configured to cause the apparatus to perform the method in any one of the first aspect to the third aspect and any possible implementation thereof.

[0057] Optionally, the at least one processor is configured to execute computer programs or instructions to perform the method in any one of the first aspect to the third aspect and any possible implementation thereof.

[0058] Optionally, the apparatus further comprises a memory configured to store the computer programs or instructions.

[0059] Optionally, the at least one processor is coupled with the memory configured to store the computer programs or instructions. The memory can be disposed outside the apparatus.

[0060] Optionally, the apparatus further comprises a communication interface through which the processor reads the instructions on the memory. It can be understood that the communication interface is coupled with the processor, and can be used to input the computer programs or instructions to the processor, or output the information in the processor.

[0061] For the sending and obtaining / receiving operations involved, if no special description is made, or if it does not contradict the actual role or inherent logic in the related description, it can be understood as output, input, etc. operations, or as sending and receiving operations performed by the radio frequency circuit and the antenna, which are not limited in the present application.

[0062] In an implementation form, the apparatus is a communication device (e.g., a terminal device, e.g., a network device).

[0063] In another implementation, the apparatus is a chip, chip system or circuit or communication module for a communication device (e.g., a terminal device, or a network device). Optionally, the chip is a modem chip, also known as a baseband chip, or a system on chip (SoC) chip or a system in package (SIP) chip containing a modem core.

[0064] In a sixth aspect, a computer readable storage medium is provided, the computer readable medium having stored thereon computer programs (e.g., program codes) or instructions which, when executed on a communication apparatus, cause the communication apparatus to perform the method according to any of the first aspect to the third aspect and any possible implementation thereof.

[0065] In a seventh aspect, a computer program product is provided, which contains instructions, when the computer program product is executed on a computer, cause the computer to perform the method according to any of the first aspect to the third aspect and any possible implementation thereof.

[0066] In an eighth aspect, a communication system is provided, which includes a first communication apparatus and a second communication apparatus. The first communication apparatus is configured to perform the method according to any implementation of the first aspect or the third aspect, and the second communication apparatus is configured to perform the method according to any implementation of the second aspect. BRIEF DESCRIPTION OF DRAWINGS

[0067] FIG. 1 is a schematic diagram of a wireless communication system suitable for embodiments of the present application.

[0068] FIG. 2 is another schematic diagram of a wireless communication system suitable for embodiments of the present application.

[0069] FIG. 3 is another schematic diagram of a wireless communication system suitable for embodiments of the present application.

[0070] FIG. 4 is a schematic diagram of a communication method 400 according to an embodiment of the present application.

[0071] FIG. 5 is a schematic diagram of a nested model splitting manner according to an embodiment of the present application.

[0072] FIG. 6 is a schematic diagram of another nested model splitting manner according to an embodiment of the present application.

[0073] FIG. 7 is a schematic diagram of a communication method 700 according to an embodiment of the present application.

[0074] FIG. 8 is a schematic diagram of a communication method 800 according to an embodiment of the present application.

[0075] FIG. 9 is a schematic diagram of a communication method 900 according to an embodiment of the present application.

[0076] FIG. 10 is a schematic block diagram of a communication apparatus 1000 according to an embodiment of the present application.

[0077] FIG. 11 is a schematic diagram of another communication apparatus 1100 according to an embodiment of the present application.

[0078] FIG. 12 is a schematic diagram of a chip system 1200 according to an embodiment of the present application. DETAILED DESCRIPTION

[0079] The technical solutions in the present application will be described below with reference to the drawings.

[0080] Before introducing the solutions of the present application, the following points are explained.

[0081] (1) In the present application, "indication" can include direct indication, indirect indication, explicit indication, implicit indication, etc. When describing that certain indication information indicates A, it can be understood that the indication information carries A, carries an identifier of A, carries B having a correlation relationship with A, carries an identifier of B having a correlation relationship with A, etc. In other words, if the receiving side of certain indication information can determine A according to the indication information, it can be described that the indication information indicates A, and the specific determination manner is not limited. When it is understood that the indication information carries A, "indication" can be replaced by "includes", and at this time, similar to the expression "sending / receiving indication information, the indication information indicates A", it can be replaced by "sending / receiving A".

[0082] In the present application, the information indicated by the indication information is referred to as to-be-indicated information. In the specific implementation process, there are many ways to indicate the to-be-indicated information, for example but not limited to, the to-be-indicated information can be directly indicated, such as the to-be-indicated information itself or an index of the to-be-indicated information, etc. The to-be-indicated information can also be indirectly indicated by indicating other information, where the other information has a correlation relationship with the to-be-indicated information. The to-be-indicated information can also be only indicated in part, and the other part of the to-be-indicated information is known or agreed in advance. For example, the indication of specific information can also be achieved by means of the arrangement order of each information agreed in advance (for example, a protocol stipulates), thereby reducing the indication overhead to a certain extent. In addition, the to-be-indicated information can be sent as a whole, or can be sent separately into multiple sub-information, and the sending period and / or sending occasion of these sub-information can be the same or different.

[0083] (2) In the present application, the expression " / " is used to represent that the objects associated before and after are in an "or" relationship; for example, A / B can represent: A or B. The expression "and / or" is used to represent that the objects associated before and after can be in an "and" relationship or an "or" relationship; for example, A and / or B can represent the following cases: A exists alone, B exists alone, A and B exist together, wherein A, B can be single or multiple. "At least one of the following" or similar expressions are used to represent any combination of the listed items; for example, at least one of A, B and (or) C can represent the following cases: A exists alone, B exists alone, C exists alone, A and B exist together, B and C exist together, A and C exist together, A, B and C exist together, wherein A, B, C can be single or multiple.

[0084] (3) In the present application, "sending" and "receiving" represent the direction of signal transmission. For example, "sending information to XX" can be understood as the destination of the information being XX, which can include direct sending through the air interface, or indirect sending through the air interface by other units or modules. "Receiving information from YY" can be understood as the source of the information being YY, which can include direct reception from YY through the air interface, or indirect reception from YY through the air interface from other units or modules. "Sending" can also be understood as the "output" of the chip interface, and "receiving" can also be understood as the "input" of the chip interface. In other words, sending and receiving can be between devices, such as between network devices and terminal devices, or within a device, such as between components, modules, chips, software modules or hardware modules within a device through buses, wires or interfaces.

[0085] (4) In various embodiments of the present application, the terms and / or descriptions of different embodiments are consistent and can be mutually referenced if there is no special description and no logical conflict. The technical features in different embodiments can be combined to form new embodiments according to their inherent logical relationship.

[0086] (5) In the present application, "predefined" can mean standard protocol predefined, or can also mean pre-agreed or pre-negotiated between devices. Among them, "protocol" can refer to standard protocols in the communication field, which can include fourth generation (4 th generation, 4G) network, fifth generation (5 th generation, 5G) network protocol, new radio (NR) protocol, 5.5G network protocol, and related protocols applied to future communication networks, which are not limited in the present application.

[0087] (6) In this application, the words "example", "such as", and the like are used for illustrative purposes, and any embodiment or design described as "example" in this application should not be interpreted as more preferred or having more advantages than other embodiments or design solutions. Rather, the word "example" is used to present the concept in a specific manner. In this application, "of", "corresponding", "relevant", "corresponding", and "associated" are sometimes used interchangeably, and it should be pointed out that when their differences are not emphasized, their meanings are consistent.

[0088] First, introduce the communication system applicable to this application.

[0089] The technical solutions provided by this application can be applied to various communication systems, such as: 5th generation (5G) or new radio (NR) system, long term evolution (LTE) system, LTE frequency division duplex (FDD) system, LTE time division duplex (TDD) system, etc. The technical solutions provided by this application can also be applied to future communication networks. The technical solutions provided by this application can also be applied to device to device (D2D) communication, vehicle-to-everything (V2X) communication, machine to machine (M2M) communication, machine type communication (MTC), and internet of things (IoT) communication system. The technical solutions provided by this application can also be applied to non-terrestrial network (NTN) system such as inter-satellite communication and satellite communication.

[0090] As an example, the satellite communication system includes a satellite base station and a terminal device. The satellite base station provides communication services for the terminal device. The satellite base station can also communicate with the base station. The satellite can act as a base station, and also as a terminal device. Among them, the satellite can refer to unmanned aerial vehicle, hot air balloon, low earth orbit satellite, medium earth orbit satellite, high earth orbit satellite, etc. The satellite can also refer to non-ground base station or non-ground device, etc.

[0091] As an example, V2X communication can include vehicle-to-vehicle (V2V) communication, vehicle-to-infrastructure (V2I) communication, vehicle-to-pedestrian (V2P) communication, vehicle-to-network (V2N) communication.

[0092] A device in a communication system can send a signal to another device or receive a signal from another device. Wherein the signal can include information, signaling or data, etc. Wherein the device can also be replaced by an entity, a network entity, a communication device, a communication module, a node, a communication node, etc. In the embodiments of the present application, the device is taken as an example for description.

[0093] The terminal device in the embodiments of the present application can be a device or module with corresponding communication functions for accessing the above-mentioned communication system. The terminal device can include various devices with wireless communication functions, which can be used to connect people, things, machines, etc. The terminal device can be widely used in various scenarios, such as cellular communication, D2D, V2X, peer to peer, M2M, MTC, IoT, virtual reality (VR), augmented reality (AR), industrial control, autonomous driving, remote medical treatment, smart power grid, smart furniture, smart office, smart wear, smart transportation, smart city UAV, robot, remote sensing, passive sensing, positioning, navigation and tracking, autonomous delivery, etc. The terminal device can be a terminal in any of the above scenarios, such as an MTC terminal, an IoT terminal, etc. The terminal device can be a 3GPP standard user equipment (UE), terminal, fixed device, mobile station device or mobile device, subscriber unit, handheld device, vehicle-mounted device, wearable device, cellular phone, smart phone, session initiation protocol (SIP) phone, wireless data card, personal digital assistant (PDA), computer, tablet computer, notebook computer, wireless modem, handset, laptop computer, computer with wireless transceiver function, smart book, vehicle, satellite, global positioning system (GPS) device, target tracking device, aircraft (such as a drone, helicopter, multi-helicopter, quad-helicopter, or airplane, etc.), ship, remote control device smart home device, industrial device, transport vehicle with wireless communication function, communication module, road side unit (RSU) with terminal function, or device built-in in the above-mentioned device (such as a communication module, modem or chip in the above-mentioned device, etc.), or other processing device connected to the wireless modem.

[0094] It should be understood that in some scenarios, the UE can also be used as a base station. For example, the UE can act as a scheduling entity that provides sidelink signals between UEs in V2X, D2D or peer to peer scenarios, etc.

[0095] In the embodiments of the present application, the device for implementing the function of the terminal device, i.e., the terminal device, can be a terminal device or a device capable of supporting the terminal device to implement the function, such as a chip system or a chip or a circuit or a communication module (i.e., a communication module performing a communication function), which can be installed in the terminal device. In the embodiments of the present application, the chip system can be composed of a chip or can include a chip and other discrete devices. In addition, the device can also be configured with program instructions for performing corresponding communication functions.

[0096] The network device in the embodiments of the present application can be a device or a module with a corresponding communication function. The network device can be a device for communicating with the terminal device, and the network device can also be referred to as an access network device or a radio access network device, such as a network device, which can be a base station. The network device in the embodiments of the present application can refer to a radio access network (RAN) node (or device) for accessing the terminal device to a wireless network. The base station can broadly cover various names in the following or be replaced by the following names, such as: Node B (NodeB), evolved Node B (eNB), next generation Node B (gNB), relay station, access point, transmitting and receiving point (Transmit / Receive Point, TRP), transmission point, primary station, secondary station, multi-standard radio (MSR) node, home base station, network controller, access node, wireless node, access point (AP), transmission node, transceiver node, baseband unit (BBU), remote radio unit (RRU), active antenna unit (AAU), remote radio head (RRH), central unit (CU), distributed unit (DU), positioning node, etc. The base station can be a macro base station, a micro base station, a relay node, a donor node, or the like, or a combination thereof. The base station can also refer to a communication module, a modem, or a chip for being disposed in the foregoing device or apparatus. The base station can also be a mobile switching center and a device assuming a base station function in D2D, V2X, M2M communication, a device assuming a base station function in a future communication system, etc. The base station can support networks of the same or different access technologies. The embodiments of the present application do not limit the specific technology and specific device form adopted by the network device.

[0097] A base station can be fixed, or mobile. For example, a helicopter or drone can be configured to act as a mobile base station, with one or more cells moving according to the location of the mobile base station. In other examples, a helicopter or drone can be configured to act as a device that communicates with another base station.

[0098] In some deployments, the network device mentioned in embodiments of the present application can be a device including a CU, or a DU, or a device including a CU and a DU, or a control plane CU node (central unit-control plane (CU-CP)) and a user plane CU node (central unit-user plane (CU-UP)), and a DU node.

[0099] In some deployments, a plurality of RAN nodes cooperate to assist a terminal device to implement wireless access, and different RAN nodes respectively implement part of the functions of a base station. For example, a RAN node can be a CU, a DU, a CU-CP, a CU-UP, or a radio unit (RU), etc. The CU and the DU can be separately arranged, or can also be included in the same network element, such as a BBU. The RU can be included in a radio frequency device or a radio frequency unit, such as an RRU, an AAU, or an RRH.

[0100] In different systems, the CU (or CU-CP and CU-UP), DU, or RU can also have different names, but those skilled in the art can understand their meanings. For example, the wireless access network can also be an open radio access network (O-RAN) architecture, in which the CU can also be referred to as an open CU (O-CU), the DU can also be referred to as an open DU (O-DU), the CU-CP can also be referred to as an open CU-CP (O-CU-CP), the CU-UP can also be referred to as an open CU-UP (O-CU-UP), and the RU can also be referred to as an open RU (O-RU). Any of the CU (or CU-CP, CU-UP), DU, and RU in the present application can be implemented by a software module, a hardware module, or a combination of a software module and a hardware module.

[0101] In an embodiment of the present application, the apparatus for implementing the function of the network device can be a network device, or can be an apparatus capable of supporting the network device to implement the function, such as a chip system or a chip or a circuit or a communication module (i.e., a communication module performing a communication function), which can be installed in the network device. In an embodiment of the present application, the chip system can be composed of a chip, or can include a chip and other discrete devices. In addition, program instructions for performing corresponding communication functions can also be configured in the apparatus. In an embodiment of the present application, only the apparatus for implementing the function of the network device is taken as an example of the network device, and the scheme of the embodiment of the present application is not limited.

[0102] The network device and the terminal device can be deployed on land, including indoors or outdoors, handheld or vehicle-mounted; can also be deployed on water; and can also be deployed on airplanes, balloons and satellites in the air. The scenario in which the network device and the terminal device are located is not limited in the embodiments of the present application.

[0103] The following briefly introduces a communication system suitable for the embodiments of the present application as follows.

[0104] Referring to FIG. 1, FIG. 1 is a schematic diagram of a wireless communication system suitable for the embodiments of the present application.

[0105] As shown in FIG. 1, the wireless communication system includes a radio access network 100. The radio access network 100 can be a next-generation (for example, 6G or higher version) radio access network, or a traditional (for example, 5G, 4G, 3G or 2G) radio access network. One or more terminal devices (120a-120j, collectively referred to as 12) can be connected to each other or connected to one or more network devices (110a, 110b, collectively referred to as 110) in the radio access network 100. The network elements in the wireless communication system are connected through an interface (such as NG, Xn), or connected through an air interface.

[0106] FIG. 1 is only a schematic diagram, and the wireless communication system can also include other devices, such as core network (CN) devices, wireless relay devices and / or wireless backhaul devices, etc., which are not shown in FIG. 1.

[0107] Referring to FIG. 2, as an example, FIG. 2 is a schematic diagram of an O-RAN system suitable for the embodiments of the present application. The O-RAN system includes a core network, an access network device and a UE. As an example, the O-RAN system can also include other components in addition to the components shown in FIG. 2, which are not limited in the present application.

[0108] The access network device can communicate with a core network (CN) through a backhaul. The access network device can communicate with a UE through an air interface. Specifically, a BBU in the access network device communicates with the core network through the backhaul. An RU in the access network device communicates with at least one UE through the air interface. The BBU communicates with at least one RU through a fronthaul, and the BBU and the RU can be co-located or not. The BBU includes at least one CU and at least one DU, and the CU and the DU can communicate through at least one midhaul. The BBU includes at least one CU and at least one DU, and the CU and the DU can communicate through at least one midhaul.

[0109] Referring to FIG. 3, as an example, FIG. 3 is a schematic diagram of an access network device suitable for embodiments of the present application.

[0110] Optionally, the access network device includes a CU. The CU is a logical node that carries radio resource control (RRC), service data adaptation protocol (SDAP) layer, packet data convergence protocol (PDCP) layer, and other control functions of the access network device. The CU can be connected to network nodes such as core networks through some interfaces. For example, E2 interface. The CU can have part of the functions of the core network. The CU (e.g., the PDCP layer and / or higher layers of the CU) is connected to the DU (e.g., the radio link control (RLC) layer and lower layers of the DU) through some interfaces. For example, F1 interface. Optionally, the F1 interface can provide control plane (C-Plane) and user plane (U-Plane) functions (e.g., interface management, system information management, UE context management, RRC message transmission, etc.). F1AP is the application protocol of the F1 interface, which defines the signaling procedures of F1 in some examples. The F1 interface supports control plane F1-C and user plane F1-U.

[0111] As an example, the CU includes a CU-CP and a CU-UP. Among them, the CU-CP is a logical node carrying the control plane part of PDCP (PDCP-C) layer of RRC layer and packet data convergence protocol layer, used to implement the control plane function of the CU. The CU-CP can interact with a network element in the core network for implementing the control plane function. The network element in the core network for implementing the control plane function can be an access and mobility function network element, such as an access and mobility management (AMF) in a 5G system. The AMF network element is used to be responsible for the mobility management in the mobile network, such as location update of the terminal device, registration network of the terminal device, handover of the terminal device, etc. The CU-UP is a logical node carrying the user plane part of PDCP (PDCP-U) layer of SDAP layer and packet data convergence protocol layer, used to implement the user plane function of the CU. The CU-UP can interact with a network element in the core network for implementing the user plane function. The network element in the core network for implementing the user plane function, for example, a user plane function (UPF) in a 5G system, is used to be responsible for the forwarding and receiving of data in the terminal device. The above configuration of the CU and the DU is only an example, and in actual application, the CU and the DU can also be configured to have functions according to needs. For example, the CU or the DU can be configured to have more functions of protocol layers, or the CU or the DU can be configured to have partial processing functions of protocol layers. For example, part of the functions of the RLC layer and the functions of the protocol layers above the RLC layer are arranged in the CU, and the remaining functions of the RLC layer and the functions of the protocol layers below the RLC layer are arranged in the DU. For another example, the functions of the CU or the DU can be divided according to the service type or other system requirements, for example, according to the delay, the functions that need to meet the delay requirement are arranged in the DU, and the functions that do not need to meet the delay requirement are arranged in the CU.

[0112] Optionally, the access network device includes a DU. As shown in FIG. 3, the DU is a logical node carrying an RLC layer, a medium access control (MAC) layer, a higher physical (Higher PHY) layer and other functions. In some examples, the DU can control at least one RU. The DU is connected to the RU through some interfaces, which can be a front interface. In some examples, the Higher PHY layer includes part of the PHY layer processing, such as forward error correction (FEC) encoding and decoding, scrambling, modulation and demodulation, etc.

[0113] Optionally, the access network device includes a RU. As shown in FIG. 3, the RU is a logical node that carries lower physical layer (Lower PHY) and radio frequency (RF) processing. In some examples, the RU can be a 3GPP transmission reception point (TRP) or a remote radio head (RRH) or other similar functional entity. In some examples, the Low-PHY includes portions of PHY processing such as fast fourier transform (FFT), inverse fast fourier transformation (IFFT), digital beamforming and filtering, and other processing functions. The RU communicates with one or more UEs over a wireless link.

[0114] The DU and the RU can or can not be co-located. The DU and the RU exchange control plane information and user plane information via a lower-layer split-control, user and synchronization (LLS-CUS) interface over a fronthaul link. The LLS-CUS can include a lower-layer split control (LLS-C) interface and a lower-layer split user (LLS-U) interface that provide control plane (C-Plane) and user plane (U-Plane), respectively. In some examples, the control plane (C-Plane) refers to real-time control between the DU and the RU. The DU and the RU have a LLS-M interface of the fronthaul link to exchange management information, and the management plane (M-Plane) refers to non-real-time management operations between the DU and the RU.

[0115] The DU and the RU can cooperate to collectively implement the functionality of the PHY layer. One DU can be connected to one or more RUs. The functionality of the DU and the RU can be configured in multiple ways according to design. For example, the DU is configured to implement baseband functionality, and the RU is configured to implement mid- RF functionality. As another example, the DU is configured to implement high-layer functionality in the PHY layer, and the RU is configured to implement low-layer functionality in the PHY layer or to implement the low-layer functionality and RF functionality. The high-layer functionality in the PHY layer can include a portion of the functionality of the PHY layer that is closer to the MAC layer, and the low-layer functionality in the PHY layer can include another portion of the functionality of the PHY layer that is closer to the mid-RF side.

[0116] The above Figs. 1 to 3 are illustrative, and embodiments of the present application are not limited thereto.

[0117] In order to better understand the technical solutions of the present application, some related technologies related to the technical solutions of the present application are introduced.

[0118] 1. Artificial intelligence: It is to make the machine have learning ability and can accumulate experience to solve the problems such as natural language understanding, image recognition and chess playing which can be solved by human experience. Artificial intelligence can be understood as the intelligence shown by the machine made by human. Artificial intelligence usually refers to the technology of presenting human intelligence through computer program. The goal of artificial intelligence includes understanding intelligence by constructing symbolic reasoning or reasoning computer program.

[0119] 2. AI model: It is an algorithm or computer program that can realize AI function. The AI model represents the mapping relationship between the input and output of the model, or the AI model is a function model that maps a certain dimension of input to a certain dimension of output. The parameters of the function model can be obtained by machine learning training. For example, f(x) = ax2 + b is a quadratic function model, which can be regarded as an AI model, and a and b are the parameters of the AI model, which can be obtained by machine learning training. Illustratively, the AI model mentioned in the embodiments of the present application is not limited to neural network, linear regression model, decision tree model, support vector machine (SVM), Bayesian network, Q learning model or other machine learning (ML) model.

[0120] The implementation of the AI model can be hardware circuit, software or combination of software and hardware, which is not limited. Non-limiting examples of software include program code, program, subprogram, instruction, instruction set, code, code segment, software module, application program or software application, etc.

[0121] 3. Model application: Using the trained model to solve practical problems.

[0122] 4. Antenna: Each communication device, such as network device 110 or terminal device 120 in Fig. 1, can be configured with multiple antennas (virtual antennas or physical antennas), and the communication devices can communicate through multiple antenna technology. The multiple antennas can include at least one transmitting antenna (also referred to as transmitting antenna) for transmitting signals and at least one receiving antenna for receiving signals.

[0123] To meet the vision of smart and inclusive in the future, intelligence will further evolve at the wireless network architecture level, AI will be further deeply integrated with wireless networks to achieve network-born intelligence, and the intelligence of terminals will also be included to meet the following new demands and new scenarios:

[0124] 1. Diversification of terminal types and more flexible and intelligent terminal connections: Diversification of terminal types, super internet of things (Supper IOT) (internet of things, vehicle networking, industry, medical treatment, etc.), massive connections, more flexible terminal connections, and terminals with certain AI capabilities.

[0125] 2. Network-born intelligence: In addition to providing traditional communication connection services, networks will further provide computing and AI services to better support inclusive, real-time, and high-security AI services.

[0126] These new demands and new scenarios will bring changes to the wireless network architecture and communication mode. In addition, model structure and parameter quantity have a great influence on performance, computation (power consumption), etc. Considering scenarios with limited computing power / performance changes, different models with different structures and parameter quantities are needed. The current 3GPP model transmission / update and model management only study the model delivery method (transmitting a model or a dataset), and do not study the structural relationship and activation / selection method between different models. One of the existing methods is relatively intuitive: when the computing power or performance changes, different models need to be trained / managed separately, and different model structures and parameters need to be transmitted separately, which makes the data transmission efficiency low and does not take advantage of the model structure characteristics.

[0127] Therefore, an embodiment of the present application proposes a way to utilize the model structure characteristics, design nested models, jointly manage each sub-model in the nested models, and reduce the number of independent models. In addition, signaling interaction can be used to select sub-models in the nested models, thereby achieving the effect that the model transmission process can adapt to the computing power and / or performance requirements of different communication devices, and further improving the transmission efficiency.

[0128] The method provided by the embodiments of the present application will be described in detail below with reference to the drawings. For ease of description, the first communication device and the second communication device are taken as examples for illustrative description. The first communication device can be a terminal device, or a component (for example, a chip or a chip system or a circuit or a communication module) of the terminal device; or the first communication device can be a network device, or a component (for example, a chip or a chip system or a circuit or a communication module) of the network device. The second communication device can be a terminal device, or a component (for example, a chip or a chip system or a circuit or a communication module) of the terminal device; or the second communication device can be a network device, or a component (for example, a chip or a chip system or a circuit or a communication module) of the network device.

[0129] Referring to FIG. 4, as an example, FIG. 4 is a schematic diagram of a communication method 400 provided by an embodiment of the present application. The method 400 shown in FIG. 4 can include the following steps.

[0130] S410, the first communication device determines a nested model.

[0131] The nested model is obtained by connecting a plurality of sub-models. As an example, the connection can also be referred to as combination or splicing, which can include series connection and / or parallel connection.

[0132] The one sub-model can include one or more functional modules (or functional layers), and therefore the nested model is obtained by connecting a plurality of sub-models, which can be alternatively replaced by: the nested model is obtained by connecting a plurality of functional modules, in other words, a model is obtained by connecting a plurality of functional module information, that is, the nested model. The nested model can include a plurality of independently usable models (that is, sub-models), and the sub-model is composed of one or more functional modules. Different sub-models have the same function, that is, each sub-model can be used independently and has the function of the nested model. As an example, different sub-models can include common functional modules, and the number of functional modules included in different sub-models can be the same or different. Different sub-models having common functional modules can achieve the same effect of the function of the nested model, which is not limited in the present application.

[0133] Each sub-model can realize the function of the nested model, in other words, each sub-model in the nested model can be independently used to realize the function of the nested model. As an example, based on the number of functional modules contained in the sub-model, the sub-model can realize the function of the nested model with different effects, that is, the accuracy of the sub-model is different. For example, if the number of functional modules contained in the sub-model is larger, the accuracy of the sub-model is higher; if the number of functional modules contained in the sub-model is smaller, the accuracy of the sub-model is lower. The sub-model can also be referred to as a nested sub-model or a functional module group. It can be understood that the nested model and the sub-model (nested sub-model) are only a possible naming manner, and the name does not limit the protection scope of the embodiments of the present application. For the sake of unity and convenience of description, the nested model and the sub-model are described below.

[0134] As an example, a suitable sub-model can be selected according to actual needs and communication scenarios. For example, for a scenario with high communication performance requirements (such as a scenario with high latency requirements and high accuracy requirements), a sub-model with more functional modules can be selected; for example, for a scenario with low communication performance requirements (such as a scenario with low latency requirements and low accuracy requirements), a sub-model with fewer functional modules can be selected.

[0135] As an example, each sub-model corresponds to different parameter requirements of different needs, and a sub-model with more functional modules can be selected, or a sub-model with a specific number of functional modules can be selected for specific needs.

[0136] Optionally, the nested model further includes an adaptation layer (or referred to as an adaptation model), such as an input adaptation layer and / or an output adaptation layer. As shown in FIG. 5, as an example, one nested model can include one input adaptation layer, a plurality of functional modules, and one output adaptation layer, wherein the input adaptation layer is connected (such as in series) with at least one functional module in the plurality of functional modules, and then the data input through the input adaptation layer can enter the plurality of functional modules; the plurality of functional modules are connected with each other (such as in series and / or in parallel), and then the functional modules can jointly process the data input through the input adaptation layer; the output adaptation layer is connected (such as in series) with at least one functional module in the plurality of functional modules, and then the data processed by the plurality of functional modules can be output through the output adaptation layer.

[0137] As an example, the nested model can be an AI model or other model, which is not limited.

[0138] Optionally, the nested model is deployed in the first communication device and / or the second communication device.

[0139] As an example, the first communication apparatus can determine the nested model based on at least one of the following: model structure, training data, performance information. The performance information can include performance information of the first communication apparatus and / or performance information of the second communication apparatus. As an example of the performance information of the first communication apparatus, the performance information of the first communication apparatus includes local computing power and / or performance requirement information.

[0140] Optionally, the first communication apparatus determining the nested model comprises the following manner: the first communication apparatus determines the nested model based on the training data.

[0141] As an example, when determining the nested model, the first communication apparatus can train the model parameters of the training data by selecting a suitable loss function and using an optimization algorithm in the training data collection link, so that the value of the loss function is less than a threshold, or so that the value of the loss function meets a target requirement, to obtain the nested model.

[0142] S420, the first communication apparatus sends first indication information to the second communication apparatus, the first indication information indicating information of the nested model.

[0143] As an example, if the nested model is deployed in the second communication apparatus, the method 400 further comprises S420, i.e., the first communication apparatus indicates the nested model to the second communication apparatus.

[0144] In one possible case, the first communication apparatus actively sends the first indication information to the second communication apparatus after completing the determination of the nested model. The first indication information can indicate at least one of the following: index, identifier, model parameter of the sub-model, which is not limited herein.

[0145] In another possible case, the first communication apparatus can send the first indication information to the second communication apparatus based on a request of the second communication apparatus. For example, before S420, the method 400 further comprises: the second communication apparatus sends first request information to the first communication apparatus, the first request information being used to request information of the nested model; in S420, in response to the first request information, the first communication apparatus sends the first indication information to the second communication apparatus.

[0146] The first indication information indicates information of the nested model, in other words, the first indication information indicates information related to the nested model, or the related information of the nested model.

[0147] In a possible implementation, the information of the nested model comprises at least one of the following: nested model structure information, nested model parameter information, and nested model segmentation information. In other words, the first indication information can indicate at least one of the following: nested model structure information, nested model parameter information, and nested model segmentation information. The nested model structure information indicates the composition information of the adaptation layer and the functional layer of the nested model, and the nested model parameter information indicates the model parameters of the sub-models in the nested model.

[0148] The nested model segmentation manner information, or segmentation information for short, can be used by the second communication device to distinguish the sub-models in the nested model, or to identify each sub-model in the nested model, or to identify the segmentation manner of each sub-model in the nested model. Specifically, the communication device (such as the second communication device) can determine which functional modules are a group, that is, which functional modules are the functional modules in a sub-model, based on the nested model segmentation manner information.

[0149] As an example, the nested model segmentation manner information comprises at least one of the following: number of layers, number of multi-heads, linear layer dimension, hidden layer dimension, and parameter quantity. The number of layers is the number of layers of the Transformer (the number of Transformer blocks), the number of multi-heads is the number of multi-head attention mechanisms in the Transformer structure (the parameter quantity increases when dh=d and remains unchanged when dh=d / Nh), the linear layer dimension is the hidden_dim and embeding output dimension, the hidden layer dimension is the dimension of the features in the Transformer module, and the parameter quantity is the number of parameters of the model.

[0150] For ease of understanding, the following is described in conjunction with FIGS. 5 and 6. As an example, the encoder (encode) in FIG. 5 is, for example, a first communication device, and the decoder (decoder) is, for example, a second communication device, and the first communication device and the second communication device both deploy a nested model. The nested model comprises a plurality of sub-models, and different sub-models comprise different numbers of functional modules, and the multilayer perceptron (MLP) MLP-i in FIG. 5 represents the output adaptation layer of a new sub-model formed by segmentation at the i th functional module (or other sub-model segmentation point) of the nested model. FIG. 6 introduces the composition of a functional module by taking one functional module as an example. As an example, one functional module can comprise at least one of the following: multi-head attention mechanism, addition, normalization, feed forward network, linear layer, attention layer, and concatenation.

[0151] Specifically, as shown in FIG. 5 to FIG. 6, taking the Transformer structure as an example, the nested model can take at least one of the number of layers, the number of multi-heads, the linear layer dimension, and the hidden layer dimension as the splitting point, that is, each sub-model in the nested model can be distinguished based on the at least one, and then the first communication device can indicate the at least one to the second communication device. Two examples are introduced below.

[0152] Example 1, the nested model splitting manner information can be realized by containing the number of layers, the linear layer dimension, and the parameter amount.

[0153] The nested model splitting manner information is indicated by a serial number, indicating that the nested model is split by the number of layers, the linear layer dimension, and the parameter amount information corresponding to the serial number in Table 1.

[0154] For example, if the nested model splitting manner information is {5-0, 5-1, 5-2, 4-0, 4-1, 4-2…}, 5 in “5-0” indicates that the first communication device model adopts the 5th indication (the corresponding parameters are layer=12, dim=32, and the parameter amount is 0.22M), and 0 in “5-0” indicates that the first communication device model adopts the 0th indication (the corresponding parameters are layer=12, dim=128, and the parameter amount is 2.5M), and the meanings of other information are similar, which will not be repeated here.

[0155] Table 1

[0156] As an example, when the nested model splitting manner information indicates that the corresponding parameters are layer=12, dim=32, and the parameter amount is 0.22M, a sub-model with a splitting layer number of 12, a linear layer dimension of 32, and a parameter amount of 0.22M is indicated.

[0157] Example 2, the nested model splitting manner information can be realized by containing the number of layers, the linear layer dimension, and the hidden layer dimension.

[0158] The nested model splitting manner information splits the nested model by directly indicating the number of layers, the linear layer dimension, and the hidden layer dimension of the Transformer structure of the first communication device and the second communication device.

[0159] For example, if the nested model splitting manner information is {T(l1,d1,h1)-R(l2,d2,h2),…}, T(l1,d1,h1) represents that the number of layers of the first communication device is l1, the linear layer dimension is d1, and the hidden layer dimension is h1; and R(l2,d2,h2) represents that the number of layers of the second communication device is l2, the linear layer dimension is d2, and the hidden layer dimension is h2.

[0160] As an example, when the nested model splitting manner information indicates that the corresponding parameter is the number of layers l1, the linear layer dimension d1, and the hidden layer dimension h1, a sub-model with the number of split layers l1, the linear layer dimension l1, and the hidden layer dimension h1 is indicated.

[0161] Example 3, the nested model splitting manner information can be implemented by containing the number of layers.

[0162] As an example, when the nested model splitting manner information indicates that the corresponding parameter is the number of layers n, a sub-model with the number of split layers n is indicated.

[0163] Example 4, the nested model splitting manner information can be implemented by containing the number of heads.

[0164] As an example, when the nested model splitting manner information indicates that the corresponding parameter is the number of heads n, a sub-model with the number of split heads n is indicated.

[0165] Example 5, the nested model splitting manner information can be implemented by containing the linear layer dimension.

[0166] As an example, when the nested model splitting manner information indicates that the corresponding parameter is the linear layer dimension n, a sub-model with the number of split linear layers n is indicated.

[0167] Example 6, the nested model splitting manner information can be implemented by containing the hidden layer dimension.

[0168] As an example, when the nested model splitting manner information indicates that the corresponding parameter is the hidden layer dimension n, a sub-model with the number of split hidden layers n is indicated.

[0169] Example 7, the nested model splitting manner information can be implemented by containing the number of parameters.

[0170] As an example, when the nested model splitting manner information indicates that the corresponding parameter is the number of parameters n, a sub-model with the number of split parameters n is indicated.

[0171] Optionally, the number of layers, the number of heads, the linear layer dimension, the hidden layer dimension, and / or the number of parameters in the above examples can also vary at the same time, and the embodiments of the present application are not limited herein.

[0172] The above examples 1 to 7 are example illustrations, and the embodiments of the present application are not limited thereto.

[0173] As an example, based on the nested model cutting manner of the specific neural network structure parameter, an MLP / convolutional neural network (CNN) + iterative unfolding structure can also be used, and the number of iterations of unfolding can be used as a nested cutting point. The number of iterations of unfolding can be conveyed in the nested model cutting manner information. In addition, the nested model cutting manner information can also be other model cutting points agreed by the first communication device and the second communication device, and the embodiments of the present application are not limited thereto.

[0174] Optionally, the method 400 further includes that the first communication device performs model inference and / or performance monitoring on the determined nested model.

[0175] As an example, the first communication device performs model inference and / or performance monitoring on the determined nested model. In the model inference link, the sub-model of the determined nested model is used to perform inference based on inference data provided by a data source to obtain an inference result. This link can also be understood as follows: the inference data is input into the sub-model, and the output of the sub-model is obtained, which is the inference result. The inference result can indicate the configuration parameters used (executed) by the first communication device and / or the operations executed by the first communication device. In the inference result application link, the inference result is published, for example, the inference result can be uniformly planned by the first communication device, for example, the first communication device can send the inference result to one or more execution objects (for example, core network equipment, access network equipment, or terminal equipment, etc.) to execute. The first communication device can also feed back the performance of the sub-model to the data source to facilitate subsequent implementation of sub-model updating.

[0176] It can be understood that the implementation of performing model inference and / or performance monitoring on the determined nested model can be hardware circuit, software, or a combination of software and hardware, which is not limited. Non-limiting examples of software include program code, programs, subprograms, instructions, instruction sets, codes, code segments, software modules, applications, or software applications, etc.

[0177] Further optionally, the first communication device updates the second sub-model after performing model inference and / or performance monitoring on the determined nested model, and notifies the second communication device of the second sub-model, so that the second communication device updates the first sub-model according to the second sub-model. The first sub-model and the second sub-model belong to one or more sub-models of the nested model, the first sub-model is deployed in the second communication device, and the second sub-model is deployed in the first communication device. After updating the first model and the second model, the first communication device loads the model parameters of the second sub-model, and the second communication device loads the model parameters of the first sub-model.

[0178] As an example, after the first communication device completes the performance monitoring on the determined nested model, the first communication device determines the sub-model of the nested model used by the first communication device as the updated second sub-model according to the performance monitoring result of the nested model (i.e., the performance monitoring result of the updated second sub-model) and the performance information of the first communication device, and notifies the second communication device of the second sub-model (for example, notifies the second communication device of at least one of the following information of the second sub-model: index, identifier, model parameter, which are not limited herein), so that the second communication device updates the first sub-model used by the second communication device according to the second sub-model.

[0179] Further optionally, the second communication device updates the first sub-model after performing model inference and / or performance monitoring on the determined nested model, and notifies the first communication device of the first sub-model, so that the first communication device updates the second sub-model according to the first sub-model. After updating the first model and the second model, the first communication device loads the model parameters of the second sub-model, and the second communication device loads the model parameters of the first sub-model.

[0180] As an example, after the second communication device completes the performance monitoring on the determined nested model, the second communication device determines the sub-model of the nested model used by the second communication device as the updated first sub-model according to the performance monitoring result of the nested model (i.e., the performance monitoring result of the updated first sub-model), and notifies the first communication device of the first sub-model (for example, notifies the second communication device of at least one of the following information of the second sub-model: index, identifier, model parameter, which are not limited herein), so that the first communication device updates the second sub-model used by the first communication device according to the first sub-model.

[0181] For ease of understanding, the following describes a specific process suitable for the embodiments of the present application. It can be understood that the process described below is only an example description, and the embodiments of the present application are not limited thereto. The content not described in detail below can refer to the description in the method 400, which is not described below.

[0182] As an example, FIG. 7 is a schematic diagram of a communication method 700 provided by the embodiments of the present application. The method 700 shown in FIG. 7 can include the following steps.

[0183] Optionally, the method 700 includes S710.

[0184] S710, the first communication device and the second communication device determine the model structure. In other words, the first communication device and the second communication device align the model structure used.

[0185] As an example, the first communication device and the second communication device can configure the delivery of the base model or the model structure online or offline, which is not limited herein.

[0186] S720, the second communication device reports training data to the first communication device, and the training data can be used to determine the nested model.

[0187] S730, the first communication device determines the nested model.

[0188] Specifically, the first communication device can train the model based on the training data received in S720, and then determine the nested model.

[0189] In a possible implementation, the first communication device determines the nested model based on the model structure and other performance information. For details, refer to the related description of the communication method 400, which is not repeated here.

[0190] As an example, the first communication device determines the model nested with different parameters and the model structure nesting based on the training data in S720 and the local computing power and / or performance requirement information of the first communication device, so as to reduce the number of independent models.

[0191] Optionally, before S730, the method 700 further includes S721: the second communication device sends performance information of the second communication device to the first communication device. In this way, the first communication device can determine the nested model and / or the sub-model based on the received performance information of the second communication device.

[0192] As an example, the first communication device determines the nested model according to the performance information of both sides, and directly sends the corresponding sub-model parameters through the information of the nested model in the first indication information, so that the second communication device does not need to further screen the selectable sub-models sent by the first communication device.

[0193] S740, the first communication device sends the first indication information.

[0194] As an example, the first communication device determines the sub-model available for the first communication device according to the local computing power and / or performance requirement information, and sends it to the second communication device.

[0195] Optionally, the information of the nested model can also include nested model splitting manner information, which is used by the second communication device to distinguish the sub-models in the nested model. The technical solutions related to the nested model splitting manner information have been described above, and are not repeated here.

[0196] Optionally, the method 700 further includes S731: the second communication device sends first request information to the first communication device.

[0197] As an example, the second communication device can filter the selectable sub-models in the information of the nested model in the first indication information sent by the first communication device according to the second communication device performance information, and send the first sub-model determined by the second communication device to the first communication device. The first communication device sends the model parameters of the first sub-model requested in the first request information to the second communication device through the information of the nested model according to the first request information sent by the second communication device, and the second communication device loads the model parameters of the first sub-model based on the model structure.

[0198] Optionally, the filtering of the selectable sub-models sent by the first communication device can be based on the nested model splitting mode information in the information of the nested model to distinguish each sub-model in the nested model. The technical solutions for distinguishing the sub-models have been described above and will not be repeated here.

[0199] Optionally, the method 700 further includes S721, the first communication device determines the nested model according to the performance information of the first communication device and the second communication device, and directly sends the sub-model parameters through the information of the nested model, so that the second communication device does not need to determine the first sub-model from the selectable sub-models sent by the first communication device.

[0200] Optionally, the information of the nested model includes nested model parameter information.

[0201] As an example, the first communication device sends the nested model parameter information to the second communication device through the information of the nested model. The technical solutions for sending the nested model parameter information have been described in the communication method 400 above and will not be repeated here.

[0202] S750, the first communication device and the second communication device perform model inference, computing power and / or performance monitoring.

[0203] As an example, the first communication device and the second communication device perform model inference and / or performance monitoring on the determined nested model. When the computing power of the first communication device and / or the second communication device is improved or the performance requirement is improved, more model parameters can be used to improve the performance.

[0204] S760, the first communication device incrementally sends the model parameters.

[0205] As an example, the first communication device does not need to transmit the model parameters that have been transmitted, but only needs to incrementally send them.

[0206] With reference to FIG. 4 to FIG. 7 and the communication method 400 and the communication method 700 in the above embodiments, in another embodiment, the first communication device can first send all model parameters of the determined nested model to the second communication device, and then select a sub-model from the nested model for inference when the model needs to be inferred.

[0207] With reference to FIG. 8, FIG. 8 is a schematic diagram of a communication method 800 provided by an embodiment of the present application. The communication method 800 shown in FIG. 8 can include the following steps.

[0208] S810, the first communication device and the second communication device determine a model structure. In other words, the first communication device and the second communication device align the model structure adopted.

[0209] As an example, the first communication device and the second communication device can configure the delivery of the basic model or the model structure online, or can offline agree on the basic model or the model structure, which is not limited herein.

[0210] S820, the second communication device reports training data to the first communication device, which can be used to determine the nested model.

[0211] S830, the first communication device determines the nested model.

[0212] Specifically, the first communication device can train the model based on the training data received in S820, and then determine the nested model. In one possible implementation, the first communication device determines the nested model based on the model structure and other performance information. For details, reference can be made to the related description in the communication method 400, which is not repeated here.

[0213] As an example, the first communication device trains models with different parameters and model structure nesting based on the training data in S820 and the local computing power and / or performance requirement information of the first communication device, so as to reduce the number of independent models.

[0214] S840, the first communication device sends first indication information.

[0215] As an example, the first communication device determines the sub-model available to the first communication device according to the local computing power and / or performance requirement information, and sends the model parameters of the sub-model available to the first communication device and the sub-model in the nested model to the second communication device through the information of the nested model.

[0216] Optionally, the information of the nested model can also include nested model splitting manner information, which is used by the second communication device to distinguish the sub-models in the nested model. The technical solutions related to the nested model splitting manner information have been described above, and are not repeated here.

[0217] S850, the first communication device and the second communication device perform model inference, computing power and / or performance monitoring.

[0218] As an example, the first communication device and the second communication device perform model inference and / or performance monitoring on the determined nested model. Where the computing power of the first communication device and / or the second communication device is improved or the performance requirement is improved, more model parameters can be used to improve performance.

[0219] S860, the second communication device updates the first sub-model.

[0220] As an example, the second communication device updates the first sub-model to be selected by the second communication device according to the nested model parameter information, the available sub-models indicated by the first communication device, and the performance information of the second communication device.

[0221] Optionally, the screening of the selectable sub-models in the information of the nested model sent by the first communication device can be performed according to the nested model splitting manner information in the information of the nested model. The technical solutions for distinguishing the sub-models have been described above and will not be repeated here.

[0222] S861, the second communication device sends second indication information.

[0223] As an example, the second communication device sends the second indication information to the first communication device to indicate the first sub-model updated by the second communication device.

[0224] S862, the first communication device determines a second sub-model.

[0225] As an example, the first communication device updates the second sub-model to be determined by the first communication device according to the first sub-model updated in the second indication information in S761.

[0226] S863, the first communication device and the second communication device perform model inference.

[0227] As an example, the first communication device loads model parameters based on the model structure according to the second sub-model updated in S862, and the second communication device loads model parameters based on the model structure according to the first sub-model updated in S860.

[0228] In another implementation manner of the communication method 800, after S850, the first communication device can first update the second sub-model:

[0229] S870, the first communication device updates the second sub-model.

[0230] As an example, the first communication device updates the sub-model to be selected by the first communication device according to at least one of the following: the nested model parameter information, the available sub-models indicated by the second communication device, the local computing power of the first communication device, and the performance requirement information of the first communication device.

[0231] S871, the first communication device sends third indication information.

[0232] As an example, the first communication device sends the third indication information to the second communication device to indicate the second sub-model updated by the first communication device.

[0233] S872, the second communication device updates the first sub-model.

[0234] As an example, the second communication device updates the first sub-model to be updated by the second communication device according to the second sub-model updated in the third indication information in S871.

[0235] S873, the first communication device and the second communication device perform model inference.

[0236] As an example, the first communication device loads the model parameters based on the model structure according to the second sub-model updated in S870, and the second communication device loads the model parameters based on the model structure according to the first sub-model updated in S872.

[0237] In another embodiment, only the second communication device can update the sub-model unilaterally, in combination with the communication method 400, the communication method 700 and the communication method 800 in the above embodiments and FIGS. 4-8.

[0238] Referring to FIG. 9, FIG. 9 is a schematic diagram of a communication method 900 provided by an embodiment of the present application. The communication method 900 shown in FIG. 9 can include the following steps.

[0239] S910, the first communication device and the second communication device determine a model structure. In other words, the first communication device and the second communication device align the model structure adopted.

[0240] As an example, the first communication device and the second communication device can configure the basic model or the model structure online or offline, which is not limited herein.

[0241] S920, the second communication device reports training data to the first communication device, which can be used to determine a nested model.

[0242] S930, the first communication device determines the nested model.

[0243] Specifically, the first communication device can train the model based on the training data received in S920, and then determine the nested model.

[0244] In a possible implementation, the first communication apparatus determines the nested model based on the model structure and other performance information. For details, refer to the description of the communication method 400, which will not be repeated here.

[0245] As an example, the first communication apparatus trains the model with different parameters and the model structure nesting based on the training data in S920 and the local computing power and / or performance requirement information of the first communication apparatus, so as to reduce the number of independent models.

[0246] In S940, the first communication apparatus sends the first indication information.

[0247] As an example, the first communication apparatus determines the sub-model available to the first communication apparatus according to the local computing power and / or performance requirement information of the first communication apparatus, and sends the model parameters of the sub-model available to the first communication apparatus and the sub-model in the nested model to the second communication apparatus through the information of the nested model.

[0248] Optionally, the information of the nested model can also include nested model splitting manner information, which is used by the second communication apparatus to distinguish the sub-models in the nested model. The technical solutions related to the nested model splitting manner information have been described above and will not be repeated here.

[0249] In S950, the second communication apparatus performs model inference, computing power and / or performance monitoring.

[0250] As an example, the second communication apparatus performs model inference and / or performance monitoring on the determined nested model, and the first communication apparatus no longer performs model inference, computing power and / or performance monitoring.

[0251] In S960, the second communication apparatus updates the first sub-model.

[0252] As an example, the second communication apparatus updates the first sub-model to be selected by the second communication apparatus according to at least one of the following: the nested model parameter information, the indication of the available sub-model by the first communication apparatus, the local computing power of the second communication apparatus, and the performance requirement information of the second communication apparatus. Wherein, the first communication apparatus no longer updates the second sub-model.

[0253] Optionally, the screening of the selectable sub-models in the information of the nested model sent by the first communication apparatus can be performed according to the nested model splitting manner information in the information of the nested model to distinguish the sub-models in the nested model. The technical solutions related to the distinguishing of the sub-models have been described above and will not be repeated here.

[0254] In S970, the second communication apparatus performs model inference.

[0255] As an example, the second communication device loads the model parameters based on the model structure according to the updated first sub-model in S960.

[0256] The method provided by the embodiments of the present application is described in detail above in combination with FIG. 4 to FIG. 9. The device provided by the embodiments of the present application is described in detail below in combination with FIG. 10 to FIG. 12. It should be understood that the description of the device embodiments corresponds to the description of the method embodiments, and therefore, the content not described in detail can be referred to the method embodiments described above, which will not be described here for brevity.

[0257] As an example, FIG. 10 is a schematic diagram of a communication device 1000 provided by the embodiments of the present application. The communication device 1000 includes a transceiver unit 1010. The transceiver unit 1010 can be used to implement corresponding communication functions. The transceiver unit 1010 can also be referred to as a communication interface or a communication unit. Optionally, the device 1000 further includes a processing unit 1020. The processing unit 1020 can be used to process the nested model in the method embodiments described above.

[0258] Optionally, the device 1000 can further include a storage unit, which can be used to store instructions and / or data. The processing unit 1020 can read the instructions and / or data in the storage unit, so that the device implements the method embodiments described above.

[0259] In a first possible design, the device 1000 can be the first communication device in the embodiments described above, and the device 1000 can implement the steps or procedures corresponding to the steps or procedures performed by the first communication device in the method embodiments described above. Specifically, the transceiver unit 1010 can be used to perform the transceiving related operations (such as the operations of sending and / or receiving data or messages) of the first communication device in the method embodiments described above, and the processing unit 1020 can be used to perform the processing related operations or operations other than transceiving (such as operations other than sending and / or receiving data or messages) of the first communication device in the method embodiments described above.

[0260] In a second possible design, the device 1000 can be the second communication device in the embodiments described above, and the device 1000 can implement the steps or procedures corresponding to the steps or procedures performed by the second communication device in the method embodiments described above. Specifically, the transceiver unit 1010 can be used to perform the transceiving related operations (such as the operations of sending and / or receiving data or messages) of the second communication device in the method embodiments described above, and the processing unit 1020 can be used to perform the processing related operations or operations other than transceiving (such as operations other than sending and / or receiving data or messages) of the second communication device in the method embodiments described above.

[0261] It should be understood that the specific process of each unit performing the corresponding steps described above has been described in detail in the method embodiments described above, and therefore, will not be described here for brevity.

[0262] It should also be understood that the apparatus 1000 is embodied in the form of a functional block diagram. The terminology used herein, such as "unit", can refer to an application specific integrated circuit (ASIC), an electronic circuit, a processor (for example, a shared processor, a dedicated processor, or a group processor, etc.) and a memory for executing one or more software or firmware programs, a combination of logic circuitry and / or other suitable components that support the described functions. In an optional example, those skilled in the art can understand that the apparatus 1000 can be embodied as a communication device in the above-mentioned embodiments, and can be used to execute the processes and / or steps corresponding to the communication device in the above-mentioned method embodiments. To avoid repetition, details are not described here.

[0263] The apparatus 1000 of each of the above-mentioned schemes has a function of implementing the corresponding steps performed by the communication device (for example, the first communication device, or the second communication device) in the above-mentioned methods. The function can be implemented by hardware, or by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the above-mentioned functions; for example, the transceiver unit can be replaced by a transceiver (for example, the transmitting unit in the transceiver unit can be replaced by a transmitter, and the receiving unit in the transceiver unit can be replaced by a receiver), and other units, such as the processing unit, can be replaced by a processor, which respectively performs the transceiving operations and related processing operations in each of the method embodiments.

[0264] In addition, the transceiver unit 1010 can also be a transceiver circuit (for example, which can include a receiving circuit and a transmitting circuit), and the processing unit can be a processing circuit.

[0265] It should be noted that the apparatus in FIG. 10 can be a communication device (for example, the first communication device, or the second communication device) in the above-mentioned embodiments, or a chip or a chip system, for example, a system on chip (SoC). The transceiver unit can be an input / output circuit, a communication interface; and the processing unit can be a processor or a microprocessor integrated on the chip or an integrated circuit. Here, no limitation is made.

[0266] Referring to FIG. 11, as an example, FIG. 11 is a schematic diagram of another communication apparatus 1100 provided by the embodiments of the present application. The apparatus 1100 includes a processor 1110, and the processor 1110 is coupled with a memory 1120, the memory 1120 is used to store computer programs or instructions and / or data, and the processor 1110 is used to execute the computer programs or instructions stored in the memory 1120, or read the data stored in the memory 1120, to execute the methods in the above-mentioned method embodiments.

[0267] Optionally, the processor 1110 is one or more.

[0268] Optionally, the memory 1120 is one or more.

[0269] Optionally, the memory 1120 is integrated with the processor 1110, or is separately arranged.

[0270] Optionally, as shown in FIG. 11, the processor 1110 can include a program 1130 (sometimes also referred to as code or instructions) that can be run on the processor 1110, so that the communication apparatus 1100 performs the methods described in the above embodiments. In yet another possible design, the communication apparatus 1100 includes a circuit (not shown in FIG. 11).

[0271] Optionally, as shown in FIG. 11, the memory 1120 can include a program 1140 (sometimes also referred to as code or instructions) that can be run on the processor 1110, so that the communication apparatus 1100 performs the methods described in the above method embodiments.

[0272] Optionally, the processor 1110 and / or the memory 1120 can include AI modules 1170 and 1180, which are used to implement AI-related functions. The AI modules can be implemented in software, hardware, or a combination of software and hardware. For example, the AI modules can include a RIC module. For example, the AI modules can be a near-real-time radio intelligent controller (RIC) or a non-real-time RIC.

[0273] Optionally, as shown in FIG. 11, the apparatus 1100 can further include a transceiver 1150 and / or an antenna 1160. The processor 1110 can sometimes also be referred to as a processing unit, which controls the communication apparatus (such as a RAN node or a terminal). The transceiver 1150 can sometimes also be referred to as a transceiving unit, a transceiver, a transceiving circuit, or a transceiver, etc., which is used to implement the transceiving function of the communication apparatus through the antenna 1160.

[0274] As an example, the processor 1110 can have the functions of the processing unit 1020 shown in FIG. 10, the memory 1120 can have the functions of a storage unit, and the transceiver 1150 can have the functions of the transceiving unit 1010 shown in FIG. 10.

[0275] As an example, the apparatus 1100 is used to implement the operations performed by the communication apparatus (such as the first communication apparatus, and such as the second communication apparatus) in the above various method embodiments.

[0276] For example, the processor 1110 is configured to execute the computer program or instructions stored in the memory 1120, to implement the related operations of the communication apparatus in the above various method embodiments.

[0277] It should be understood that the processor mentioned in the embodiments of the present application can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor.

[0278] It should also be understood that the memory mentioned in the embodiments of the present application can be a volatile memory and / or a non-volatile memory. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically EPROM (EEPROM) or a flash memory. The volatile memory can be a random access memory (RAM). For example, the RAM can be used as an external cache. As an example but not limitation, the RAM includes the following various forms: static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchlink DRAM (SLDRAM) and direct rambus RAM (DR RAM).

[0279] It should be noted that when the processor is a general-purpose processor, a DSP, an ASIC, an FPGA or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component, the memory (storage module) can be integrated in the processor.

[0280] It is also important to note that the storage described herein is intended to comprise, without being limited to such, these and any other suitable types of storage.

[0281] Referring to FIG. 12, as an example, FIG. 12 is a schematic diagram of a chip system 1200 provided by embodiments of the present application. The chip system 1200 (or also can be referred to as a processing system) comprises a logic circuit 1210 and an input / output interface 1220.

[0282] The logic circuit 1210 can be a processing circuit in the chip system 1200. The logic circuit 1210 can be coupled to a storage unit, and invoke instructions in the storage unit, so that the chip system 1200 can implement the methods and functions of embodiments of the present application. The input / output interface 1220 can be an input / output circuit in the chip system 1200, and output information processed by the chip system 1200, or input data or signaling information to be processed by the chip system 1200.

[0283] As an example, the chip system 1200 is configured to implement operations performed by a communication device (such as the first communication device, or the second communication device) in the above various method embodiments.

[0284] For example, the logic circuit 1210 is configured to implement processing-related operations performed by a communication device (such as the first communication device, or the second communication device) in the above method embodiments; and the input / output interface 1220 is configured to implement sending and / or receiving-related operations performed by a communication device (such as the first communication device, or the second communication device) in the above method embodiments.

[0285] Embodiments of the present application also provide a computer readable storage medium having stored thereon a computer program or instructions for implementing the method performed by a communication device (such as the first communication device, or the second communication device) in the above various method embodiments. For example, the computer program or instructions, when run on the communication device, cause the communication device (such as the first communication device, or the second communication device) to perform the above method (such as the communication method 400, the communication method 700, the communication method 800, or the communication method 900).

[0286] Embodiments of the present application also provide a computer program product comprising instructions, which, when executed by a computer, implement the method performed by a communication device (such as the first communication device, or the second communication device) in the above various method embodiments. For example, the computer program or instructions, when run on the communication device, cause the communication device (such as the first communication device, or the second communication device) to perform the above method (such as the communication method 400, the communication method 700, the communication method 800, or the communication method 900).

[0287] The explanations and beneficial effects of the related contents in any of the above provided devices can refer to the corresponding method embodiments provided above, and will not be repeated here.

[0288] In several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other manners. For example, the above described device embodiments are merely schematic. For example, the division of the units is only a logical function division. There can be another division manner for the actual implementation. For example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed mutual couplings or direct couplings or communication connections can be indirect couplings or communication connections through some interfaces, devices or units, and can be electrical, mechanical or in other forms.

[0289] In the above embodiments, all or part of the embodiments can be implemented by software, hardware, firmware or any combination thereof. When implemented by software, all or part of the embodiments can be implemented in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of the present application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. For example, the computer can be a personal computer, a server, a network device, etc. The computer instructions can be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another computer readable storage medium, for example, the computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center through wired (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.) manner. The computer readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. integrated with one or more available media. The available media can be magnetic media (such as floppy disk, hard disk, magnetic tape), optical media (such as DVD), or semiconductor media (such as solid state disk (SSD), etc. For example, the foregoing available media includes but is not limited to: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk, etc. Various media that can store program codes.

[0290] The above merely provides the specific implementation of the present application, but the protection scope of the present application is not limited to this. Any person skilled in the art can easily think of the changes or replacements within the technical range disclosed by the present application, which should be covered in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A communication method characterized by comprising: The method is applied to a first communication device, and comprises: determining a nested model, the nested model being obtained by connecting a plurality of sub-models, each of the plurality of sub-models being capable of implementing a function of the nested model; sending first indication information to a second communication device, the first indication information indicating information of the nested model.

2. The method of claim 1, wherein, The determination of the nested model comprises: receiving first training data; determining the nested model according to the first training data.

3. The method of claim 2, wherein, The determination of the nested model according to the first training data comprises: determining the nested model according to the first training data and performance information; wherein the performance information comprises performance information of the first communication device and / or performance information of the second communication device.

4. The method of claim 3, wherein, The method further comprises: receiving performance information of the second communication device from the second communication device.

5. The method according to any one of claims 1 to 4, characterized in that, The method further comprises: receiving first request information from the second communication device, the first request information being used to request a first sub-model in the nested model, the first sub-model comprising one or more of the sub-models; The sending of the first indication information to the second communication device, the first indication information indicating information of the nested model, comprises: sending first indication information to the second communication device based on the first request information, the first indication information indicating the first sub-model.

6. The method of claim 5, wherein, The indication of the first sub-model by the first indication information comprises: the first indication information indicating at least one of an index, an identifier, and a model parameter of the first sub-model.

7. The method according to any one of claims 1 to 6, characterized in that, After the sending of the first indication information to the second communication device, the method further comprises: performing model inference and / or performance monitoring on the nested model.

8. The method of claim 7, wherein, The method further comprises: updating a second sub-model in the nested model, the second sub-model comprising one or more of the sub-models, the second sub-model being deployed at the first communication device, the second sub-model being determined based on at least one of a performance monitoring result of the second sub-model and second indication information from the second communication device, the second indication information indicating an update to the second sub-model.

9. The method of claim 8, wherein, The indication of the update to the second sub-model by the second indication information comprises: the second indication information indicating at least one of an index, an identifier, and a model parameter of a first sub-model in the nested model.

10. The method according to claim 8 or 9, characterized in that, The method further comprises: sending third indication information to the second communication device, the third indication information indicating the second sub-model.

11. The method of claim 10, wherein, The indication of the second sub-model by the third indication information comprises: the third indication information indicating at least one of an index, an identifier, and a model parameter of the second sub-model.

12. A communication method characterized by comprising: The method is applied to a second communication device, and comprises: receiving first indication information from a first communication device, the first indication information indicating information of a nested model; determining the nested model based on the first indication information, the nested model being obtained by connecting a plurality of sub-models, each of the plurality of sub-models being capable of implementing a function of the nested model.

13. The method of claim 12, wherein, After receiving the first indication information from the first communication device, the method further comprises: performing model inference and / or performance monitoring according to performance information; wherein the performance information comprises performance information of the first communication device and / or performance information of the second communication device.

14. The method of claim 13, wherein, The method further comprises: sending performance information of the second communication device to the first communication device; and / or receiving performance information of the first communication device from the first communication device.

15. The method according to any one of claims 12 to 14, characterized in that, After receiving the first indication information from the first communication device, the method further comprises: updating a first sub-model in the nested model, the first sub-model comprising one or more of the sub-models, the first sub-model being deployed at the second communication device, the first sub-model being determined based on at least one of: performance monitoring result of the first sub-model, the first indication information from the first communication device, third indication information from the first communication device, the third indication information indicating to update the first sub-model.

16. The method of claim 15, wherein, The third indication information indicating to update the first sub-model comprises: the third indication information indicating at least one of: index, identification, model parameter of a second sub-model in the nested model, the second sub-model being deployed at the first communication device.

17. The method according to claim 15 or 16, characterized in that The method further comprises: sending first request information to the first communication device, the first request information being used to request the first sub-model in the nested model.

18. The method of any one of claims 15-17, wherein, The method further comprises: sending second indication information to the first communication device, the second indication information indicating the first sub-model.

19. The method of claim 18, wherein, The second indication information indicating the first sub-model comprises: the second indication information indicating at least one of: index, identification, model parameter of the first sub-model.

20. The method of any one of claims 1 to 19, wherein, The information of the nested model comprises model split information, the model split information being used to distinguish one or more of the sub-models in the nested model.

21. The method of claim 20, wherein, The model split information comprises at least one of: number of layers, number of heads, linear layer dimension, hidden layer dimension, parameter quantity of the nested model.

22. The method of any one of claims 3, 4, 13, 14, or 17, wherein, The performance information comprises local computing power and / or performance requirement information.

23. A communications device, characterized by comprise a module or unit for performing the method of any one of claims 1-11; or comprise a module or unit for performing the method of any one of claims 12-22.

24. A communications device, characterized by comprise a processor configured to cause the communication device to perform the method of any one of claims 1-11, or configured to cause the communication device to perform the method of any one of claims 12-22.

25. A computer-readable storage medium, characterized in that, The computer program or instructions stored on the computer readable storage medium cause the communication device to perform the method of any one of claims 1-11, or cause the communication device to perform the method of any one of claims 12-22, when the computer program or instructions are run on the communication device.

26. A computer program product, characterised in that, The computer program product comprises computer programs or instructions which, when run on a communication device, cause the communication device to perform the method of any one of claims 1 to 11, or cause the communication device to perform the method of any one of claims 12 to 22.

Images