Communication method, terminal device and network device

Abstract

The invention provides a communication method, terminal equipment and network equipment. The method comprises the following steps: the terminal equipment sends first capability information; wherein the first capability information is associated with the first model, and the first capability information is used for indicating one of the following information: information of resources capable of being used by the first model; and information of a running process of the first model. In the running process of the terminal equipment, the resources capable of being used by the first model or the running condition of the first model are dynamically changed. Therefore, the first capability information reported by the terminal equipment can be dynamically changed. That is, the terminal device may dynamically indicate capability information of the terminal device based on the first capability information. Based on the dynamic capability information, the network equipment can configure or adjust a proper model or strategy for the terminal equipment, so that the model can continuously and normally operate.

Description

Communication method, terminal device and network device Technical Field

[0001] The present application relates to the field of communication technology, and more specifically, to a method, terminal equipment, and network equipment for communication. Background Art

[0002] The capability information of terminal devices reported by relevant technologies is based on fixed reference indicators. Specifically, capability information is generally reported once, and the specific information reported is the inherent attribute of the terminal device level. The inherent attribute may include, for example, the AI ​​capability level supported by the chip of the terminal device. However, during the actual operation of the terminal device, task models at different levels can work together in parallel. In other words, the resources that can be allocated to models at different times and under different tasks are dynamically changing. And the dynamically changing resources will directly affect whether the model can operate normally and the effect of the operation. In other words, the model determined based on the capability information of the inherent attribute category is difficult to continue to work normally.

[0003] Summary of the Invention

[0004] The present application provides a method, terminal device, and network device for communication. The following introduces various aspects of the present application.

[0005] In a first aspect, a method for communication is provided, comprising: a terminal device sending first capability information; wherein the first capability information is associated with a first model, and the first capability information is used to indicate one of the following information: information about resources that can be used by the first model; information about the operation process of the first model.

[0006] In a second aspect, a method for communication is provided, including: a network device receives first capability information sent by a terminal device; wherein the first capability information is associated with a first model, and the first capability information is used to indicate one of the following information: information about resources that can be used by the first model; information about the operation process of the first model.

[0007] In a third aspect, a terminal device is provided, including: a first sending unit, used to send first capability information; wherein the first capability information is associated with a first model, and the first capability information is used to indicate one of the following information: information about resources that the first model can use; information about the operation process of the first model.

[0008] In a fourth aspect, a network device is provided, including: a second receiving unit, used to receive first capability information sent by a terminal device; wherein the first capability information is associated with a first model, and the first capability information is used to indicate one of the following information: information about resources that the first model can use; information about the operation process of the first model.

[0009] In a fifth aspect, a terminal device is provided, comprising a processor and a memory, wherein the memory is used to store one or more computer programs, and the processor is used to call the computer program in the memory so that the terminal device executes part or all of the steps in the method of the first aspect.

[0010] In a sixth aspect, a network device is provided, comprising a processor, a memory, and a transceiver, wherein the memory is used to store one or more computer programs, and the processor is used to call the computer program in the memory so that the network device executes part or all of the steps in the method of the second aspect.

[0011] In a seventh aspect, an embodiment of the present application provides a communication system, which includes the above-mentioned terminal device and / or network device. In another possible design, the system may also include other devices that interact with the terminal device or network device in the solution provided in the embodiment of the present application.

[0012] In an eighth aspect, an embodiment of the present application provides a computer-readable storage medium, which stores a computer program, and the computer program enables a terminal device and / or a network device to execute part or all of the steps in the methods of the above aspects.

[0013] In a ninth aspect, embodiments of the present application provide a computer program product, wherein the computer program product includes a non-transitory computer-readable storage medium storing a computer program, wherein the computer program is operable to cause a terminal device and / or a network device to perform some or all of the steps of the methods described in each of the above aspects. In some implementations, the computer program product may be a software installation package.

[0014] In the tenth aspect, an embodiment of the present application provides a chip, which includes a memory and a processor. The processor can call and run a computer program from the memory to implement some or all of the steps described in the methods of the above aspects.

[0015] During the operation of the terminal device, the resources available to the first model or the operation status of the first model changes dynamically. Therefore, the first capability information reported by the terminal device can change dynamically. In other words, based on the first capability information, the terminal device can dynamically indicate the capability information of the terminal device. Based on this dynamic capability information, the network device can configure or adjust the appropriate model or policy for the terminal device, thereby ensuring the continuous normal operation of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] FIG1 is a schematic diagram of a wireless communication system used in an embodiment of the present application.

[0017] FIG2 is an example diagram of a neural network model.

[0018] FIG3 is an example diagram of a channel state information feedback system.

[0019] 4A and 4B are exemplary diagrams of the beam scanning process, respectively.

[0020] FIG5 is a flowchart showing an example of a workflow of an online learning solution.

[0021] FIG6 is a schematic flowchart of a method for communication provided in an embodiment of the present application.

[0022] FIG7 is a schematic flowchart of another method for communication provided in an embodiment of the present application.

[0023] FIG8 is a schematic flowchart of another method for communication provided in an embodiment of the present application.

[0024] FIG9 is a schematic flowchart of another method for communication provided in an embodiment of the present application.

[0025] FIG10 is a schematic flowchart of another method for communication provided in an embodiment of the present application.

[0026] FIG11 is a schematic structural diagram of a terminal device provided in an embodiment of the present application.

[0027] FIG12 is a schematic structural diagram of a network device provided in an embodiment of the present application.

[0028] FIG13 is a schematic structural diagram of a device for communication provided in an embodiment of the present application. DETAILED DESCRIPTION

[0029] The technical solution in this application will be described below with reference to the accompanying drawings.

[0030] Communication System

[0031] Figure 1 illustrates a wireless communication system 100 used in an embodiment of the present application. The wireless communication system 100 may include a network device 110 and a terminal device 120. The network device 110 may be a device that communicates with the terminal device 120. The network device 110 may provide communication coverage for a specific geographic area and may communicate with the terminal device 120 within the coverage area.

[0032] FIG1 exemplarily shows a network device and two terminals. Optionally, the wireless communication system 100 may include multiple network devices and each network device may include other numbers of terminal devices within its coverage area, which is not limited in the embodiments of the present application.

[0033] Optionally, the wireless communication system 100 may further include other network entities such as a network controller and a mobility management entity, which is not limited in the embodiment of the present application.

[0034] It should be understood that the technical solutions of the embodiments of the present application can be applied to various communication systems, such as: fifth generation (5G) system or new radio (NR), long term evolution (LTE) system, LTE frequency division duplex (FDD) system, LTE time division duplex (TDD), etc. The technical solutions provided in this application can also be applied to future communication systems, such as the sixth generation mobile communication system, satellite communication system, etc.

[0035] The terminal device in the embodiments of the present application may also be referred to as user equipment (UE), access terminal, user unit, user station, mobile station, mobile station (MS), mobile terminal (MT), remote station, remote terminal, mobile device, user terminal, terminal, wireless communication device, user agent or user device. The terminal device in the embodiments of the present application may refer to a device that provides voice and / or data connectivity to a user and can be used to connect people, objects and machines, such as a handheld device with wireless connection function, a vehicle-mounted device, etc. The terminal device in the embodiments of the present application can be a mobile phone, a tablet computer, a laptop computer, a PDA, a mobile internet device (MID), a wearable device, a virtual reality (VR) device, an augmented reality (AR) device, a wireless terminal in industrial control, a wireless terminal in self-driving, a wireless terminal in remote medical surgery, a wireless terminal in a smart grid, a wireless terminal in transportation safety, a wireless terminal in a smart city, a wireless terminal in a smart home, etc. Optionally, the UE can be used to act as a base station. For example, the UE can act as a scheduling entity that provides sidelink signals between UEs in vehicle-to-everything (V2X) or device-to-device (D2D). For example, a cellular phone and a car communicate with each other using sidelink signals. The cellular phone and smart home devices communicate without relaying the communication signal through a base station.

[0036] The network device in the embodiments of the present application may be a device for communicating with a terminal device, and may also be referred to as an access network device or a radio access network device. For example, the network device may be a base station. The network device in the embodiments of the present application may refer to a radio access network (RAN) node (or device) that connects a terminal device to a wireless network. A base station may broadly cover various names as follows, or be replaced with the following names, such as: NodeB, evolved NodeB (eNB), next generation NodeB (gNB), relay station, access point, transmitting and receiving point (TRP), transmitting point (TP), master eNB (MeNB), secondary eNB (SeNB), 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. A base station may be a macro base station, a micro base station, a relay node, a donor node, or the like, or a combination thereof. A base station may also refer to a communication module, modem, or chip used to be set in the aforementioned device or apparatus. A base station may also be a mobile switching center and a device that performs base station functions in D2D, V2X, and machine-to-machine (M2M) communications, a network-side device in a 6G network, or a device that performs base station functions in future communication systems. A base station may support networks with the same or different access technologies. The embodiments of this application do not limit the specific technology and specific device form adopted by the network equipment.

[0037] Base stations can be fixed or mobile. For example, a helicopter or drone can be configured to act as a mobile base station, and one or more cells can move based on 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.

[0038] In some deployments, the network device in the embodiments of the present application may refer to a CU or a DU, or the network device may include a CU and a DU. The gNB may also include an AAU.

[0039] The network equipment and terminal devices can be deployed on land, including indoors or outdoors, handheld or vehicle-mounted; they can also be deployed on water; they can also be deployed in the air on aircraft, balloons, and satellites. The embodiments of this application do not limit the scenarios in which the network equipment and terminal devices are located.

[0040] It should be understood that all or part of the functions of the communication device in this application can also be implemented through software functions running on hardware, or through virtualization functions instantiated on a platform (such as a cloud platform).

[0041] AI

[0042] In recent years, artificial intelligence (AI), exemplified by neural networks (NN), has achieved remarkable success in numerous fields and will continue to play a vital role in our lives and production for a long time to come. In particular, machine learning (ML), a key research area in AI, leverages the nonlinear processing capabilities of neural networks to successfully solve a range of previously intractable problems. AI technology has even demonstrated superior performance to humans in areas such as image recognition, speech processing, natural language processing, and gaming, and is therefore attracting increasing attention.

[0043] A common model in AI technology is the neural network model. Neural networks are nonlinear and data-driven. They can utilize multiple layers. Figure 2 shows an example of a neural network model. As shown in Figure 2, layer-by-layer training of multi-layer neural networks for feature learning significantly enhances the learning and processing capabilities of neural networks. Therefore, neural network models are widely used in pattern recognition, signal processing, optimization and combination, anomaly detection, and other fields.

[0044] Given the tremendous success of AI technology, particularly deep learning, in areas such as computer vision and natural language processing, the communications field is beginning to explore the use of deep learning to address technical challenges that are difficult to address with traditional communications methods. For example, AI can be applied to modeling or learning complex and unknown environments, channel prediction, intelligent signal generation and processing, network status tracking and intelligent scheduling, and network optimization and deployment. AI technology is expected to promote the evolution of future communications paradigms and transform network architectures, and is of great significance and value to 6G technology research.

[0045] The following describes the integration of AI and communications through the application of AI models in channel state feedback and beam management in the communications field.

[0046] Channel state information (CSI) feedback based on AI models

[0047] Using AI models, terminal devices can extract features from actual channel matrix data, allowing network equipment to restore the channel matrix information compressed and fed back by the terminal device as closely as possible. This allows the AI ​​model to restore channel information while also providing the possibility of reducing CSI feedback overhead for terminal devices.

[0048] This article introduces CSI feedback based on AI models, using a deep learning autoencoder as an example. Deep learning-based CSI feedback treats channel information as an image to be compressed, compresses it using a deep learning autoencoder, and reconstructs the compressed channel image at the transmitter. This approach preserves channel information to a greater extent.

[0049] Figure 3 is an example diagram of a channel state information feedback system. The feedback system shown in Figure 3 is implemented based on an autoencoder structure. The autoencoder is divided into an encoder and a decoder. The encoder and decoder are deployed at the transmitter and receiver, respectively. After the transmitter obtains the original CSI (original CSI) through channel estimation, the channel information matrix is ​​compressed and encoded through the neural network of the encoder, and the compressed bit stream is fed back to the receiver through the air interface feedback link. The receiver recovers the channel information based on the feedback bit stream through the decoder to obtain complete feedback channel information or recovered CSI (reconstructed CSI). It should be noted that the network model structure inside the encoder and decoder shown in Figure 3 can be flexibly designed.

[0050] AI-based beam management

[0051] Some communication protocols (such as the first version of the NR system, Release 15) have introduced millimeter wave frequency band communications and corresponding beam management mechanisms. Simply put, beam management can be divided into uplink and downlink beam management. The following mainly uses the downlink beam management mechanism as an example. The downlink beam management mechanism includes downlink beam scanning, beam reporting, and downlink beam indication by network devices.

[0052] The downlink beam scanning process may refer to the network device scanning different transmit beam directions through the downlink reference signal synchronization block (synchronization signal / PBCH block, SSB) and / or the channel state information measurement reference signal (channel state information reference signal, CSI-RS). The terminal device can use different receiving beams for measurement, so that it can traverse all beam pair combinations. During the measurement process, the terminal device can calculate the layer 1 reference signal received power (L1 reference signal received power, L1-RSRP) value of the beam pair. It should be noted that L1-RSRP here can also be replaced by other beam link indicators. For example, other indicators may include: layer 1 signal to interference plus noise ratio (L1-SINR), layer 1 reference signal received quality (L1 reference signal received quality, L1-RSRQ), etc. Among them, L1-SINR is already supported in some communication standards, and L1-RSRQ is not supported in some communication standards.

[0053] Figures 4A and 4B are diagrams illustrating the beam scanning process. Figure 4A illustrates the process of traversing transmit beams and receive beams, while Figure 4B illustrates the process of traversing receive beams for a specific transmit beam.

[0054] Beam reporting can also be called optimal beam reporting. The terminal device can compare the L1-RSRP values ​​of all measured beam pairs, select the K transmit beams with the highest L1-RSRP, and report these K transmit beams as uplink control information to the network device. K can be a positive integer. After decoding the beam reporting of the terminal device, the network device can complete the beam indication of the terminal device through the transmission configuration indicator (TCI) status (including SSB or CSI-RS as a reference transmit beam) carried by the medium access control control element (MAC CE) or downlink control information (DCI) signaling. The terminal device can use the receive beam corresponding to the transmit beam for reception.

[0055] In discussions of some communication standards (such as R18), AI-based beam management is one of the main use cases for the AI ​​project of the communication standard, and multiple rounds of use case selection and simulation hypothesis discussions have been carried out. Although there is currently no consensus on the details of how to achieve better beam management based on AI, AI-based spatial beam prediction and time-domain prediction are both considered typical use cases. Currently, there is a preliminary consensus on the implementation framework of AI-based beam management as follows: beam prediction is implemented on beam set A (set A) through the measurement results of beam set B (set B); set B can be a subset of set A, or set B and set A can be different beam sets (for example, set A uses narrow beams and set B uses wide beams); the AI ​​model can be deployed on network equipment or terminal devices; the measurement results on set B can be L1-RSRP, or other auxiliary information, such as beam (pair) ID.

[0056] Online and offline learning

[0057] AI networks can create or train AI models based on training data. Model training is typically done to generate more accurate predictions. Online learning and offline learning are two common methods for training models in deep learning.

[0058] Offline learning, also known as offline training, involves obtaining all available training data, randomly shuffling it, and training the model offline using batches of data. For offline learning methods, the model cannot be used for predictions until offline training is complete.

[0059] Online learning, also known as online training, allows model updates to be performed online using streaming data. For example, online learning methods can adjust or update models based on a single or batch of data samples obtained in real time. Online learning methods can capture data changes promptly, effectively increasing the frequency of model updates.

[0060] Currently, most simulation results are evaluated using simulated data, with few evaluations conducted in real systems. Real-world environments are more complex, posing significant challenges to model generalization. Wireless environments are unstable, and data distribution is inevitably affected by factors such as time, environment, and system policies. Therefore, the distribution of real-world data will not be strictly consistent with that of offline data. AI model performance is strongly correlated with data distribution. If there are significant discrepancies between real-world data and offline data, AI models pre-trained using offline data will perform poorly. With the increasing capabilities of terminal devices and network equipment, and driven by more real-world data, online learning solutions will be increasingly discussed to adapt AI models to real-world environments. However, current discussions on online learning solutions primarily focus on frameworks and overall processes. For example, some communication protocols (such as Release 18) discuss the deployment of offline pre-trained models and AI frameworks for online inference.

[0061] FIG5 is a flowchart showing an example of a workflow of an online learning solution. FIG5 is described below.

[0062] During the offline training phase, the offline device uses the collected offline training data to pre-train the task model. After pre-training is completed, the task model can be deployed.

[0063] During the online training phase, online devices collect data from real-world systems as online training data. When a sufficient amount of online training data has accumulated, the online test device performs online training based on the deployed task model to update it. Training continues until the model converges or other default training termination conditions are triggered. The updated task model can be deployed and applied online. Based on the input inference data, the deployed task model outputs the corresponding inference results, which are then delivered to business applications.

[0064] AI capability reporting of terminal devices

[0065] Some companies have proposed a method for determining an AI model. This method can be executed by a terminal device. The terminal device sends its AI capability information. Based on the reported AI capability information, the network device can determine the AI ​​model used by the terminal device. The AI ​​capability information may include at least one of AI capability indication information, AI level indication information, AI model identification information, AI platform identification information, AI reasoning indication information, and AI training indication information.

[0066] In some embodiments, the AI ​​capability information may include the reference computing power of the terminal device (referred to as reference computing power). The relevant technology can judge whether the current terminal device is capable of supporting the operation of the AI ​​model based on the inference delay (or inference time) required for a given task under the reference computing power. The inference delay can also be determined according to the complexity of the AI ​​model. The determined inference delay can be compared with the delay requirement of the AI ​​model. If the inference delay meets the delay requirement, the terminal device can meet the requirements of the AI ​​model, that is, it can support the operation of the AI ​​model. If the inference delay does not meet the delay requirement, the terminal device cannot meet the requirement, that is, it cannot support the operation of the AI ​​model.

[0067] The mathematical expression of inference delay can satisfy: Among them, C can represent the computational complexity of the AI ​​model, P can represent the reference computing power of the terminal device, and T can represent the inference latency of the terminal device for the AI ​​model. The unit of C can be floating point operations per second (FLOPS). The unit of P can be FLOPS or tera operations per second (TOPS). Since C and P are obtained under ideal conditions, T calculated according to the formula is generally also a theoretical reference value. Taking into account constraints such as scheduling, storage, input / output (I / O), the actual inference latency will differ from the theoretical reference value. Therefore, the inference latency obtained through ideal assumptions is difficult to ensure that the model under a certain use case can continue to work normally.

[0068] The capability information reported by terminal devices in related technologies is based on fixed reference indicators. Specifically, capability information is generally reported once, and the specific information reported is inherent attributes at the terminal device level. These inherent attributes may include, for example, the AI ​​capability level supported by the terminal device's chip. However, during actual terminal device operation, task models at different levels can operate concurrently. In other words, the resources allocated to models at different times and for different tasks change dynamically. This dynamic change in resources directly impacts the model's ability to function properly and effectively. In other words, models determined based on capability information of inherent attribute categories are unlikely to function consistently and properly. For example, a terminal device may report its inherent reference computing power as A. The network device can deploy model B for the terminal device based on A. While model B is running, the terminal device can also execute task C in parallel. Therefore, task C requires some computing power A to execute, which will inevitably cause delays or even prevent model B from functioning properly.

[0069] FIG6 is a schematic flow chart of a method provided by an embodiment of the present application to solve the above problem. The method shown in FIG6 can be executed by a network device and a terminal device. The method shown in FIG6 can include step S610.

[0070] In step S610, the terminal device sends first capability information. Correspondingly, the network device may receive the first capability information.

[0071] The first capability information may be associated with the first model. In other words, the first capability information may be used to indicate information about the capabilities of the terminal device associated with the first model. For example, the first capability information may be used to indicate information about resources that the first model can use. Alternatively, the first capability information may be used to indicate information about the operation process of the first model. The first model may be used to perform a first task, that is, the first model may be associated with the first task. In this case, the first capability information may be used to indicate information about resources that the first task corresponding to the first model can use.

[0072] During the operation of the terminal device, the resources available to the first model or the operation status of the first model changes dynamically. Therefore, the first capability information reported by the terminal device can change dynamically. In other words, based on the first capability information, the terminal device can dynamically indicate the capability information of the terminal device. Based on this dynamic capability information, the network device can configure or adjust the appropriate model or policy for the terminal device, thereby ensuring the continuous normal operation of the model.

[0073] In addition, when the first capability information is used to indicate the actual operation process of the first model, whether the terminal device is capable of supporting the operation of the first model can be determined based on the actual operation process of the first model. In other words, the terminal device can report the actual operation status of the first model on the terminal device, thereby avoiding deviations caused by theoretical calculations based on ideal assumptions.

[0074] The resources of the terminal device may be any resources that support the operation of the first model. For example, the resources of the terminal device may include one or more of the following: memory, video memory, computing power, number of threads, etc. The resources available to the first model may be the resources actually available to the first model at the current time or for a period of time close to the current time.

[0075] As an implementation, the information about resources that can be used by the first model may include information about resources that can be used by the terminal device (i.e., available resources). As an implementation, the resources that can be used by the terminal device may include one or more of the following resources: available memory, available video memory, available computing power, available number of threads, etc.

[0076] In some embodiments, the terminal device may have autonomous resource allocation capabilities. That is, the terminal device can allocate resources on its own. In this case, the terminal device may allocate resources for the first model or the task corresponding to the first model. The resources that the first model can use may include the resources (e.g., resource upper limit) allocated by the terminal device to the first model or the first task corresponding to the first model. As an implementation method, the resources allocated by the terminal device to the first model or the first task may include, for example: one or more of the maximum memory allocated for the first model, the maximum video memory allocated, the upper limit of the computing power allocated, and the upper limit of the number of threads allocated.

[0077] The first model can be deployed and run on a terminal device. The running process of the first model can include an inference process and / or a training process. The training process can be, for example, an online training process. For example, the terminal device can perform online training on the first model. Alternatively, the terminal device can implement inference or prediction based on the first model.

[0078] In some embodiments, the information of the running process may include information related to the most recent run. In this case, the information of the running process may be used to indicate the situation of the current first model running.

[0079] The information of the running process may include one or more of the following information: the duration of the running process, and information about the resources occupied by the running process.

[0080] The running process can be a process in which the first model is run one or more times. For example, during the running process, the first model can perform one or more online trainings. Alternatively, during the running process, one or more inferences or predictions can be performed based on the first model. Correspondingly, the duration of the running process can be used to indicate the duration of the first model running one or more times. For example, the duration of the running process can be used to indicate one or more of the following durations: the duration of a single run of the first model, the average duration of a single run of the first model, the duration of multiple consecutive runs of the first model, or the total duration of the first model running. For example, if the running process is an inference process, the duration of the running process can include the average inference duration. For example, if the running process is an online training process, the duration of the running process can include the average duration of a single training step. The duration of multiple consecutive runs of the first model can include: the duration of N consecutive inferences of the first model and / or the duration of K consecutive online trainings of the first model. Both N and K can be positive integers. Both N and K can be preconfigured, preset, or configured by the network device.

[0081] The information about resources occupied by the running process can be used to indicate information related to the resources occupied by the first model during the running process. The information about resources occupied by the running process can include: one or more of the peak value (maximum value), average value, and valley value (minimum value) of the resources occupied by the running process. For example, the information about resources occupied by the running process can include: one or more of the peak value of memory and the peak value of video memory.

[0082] In some embodiments, the first capability information may also include other capabilities related to the first model. For example, the first capability information may include one or more of the following information: reference indicator information of other chip capabilities of the terminal device, reference indicator information of currently available other chip capabilities, and other statistical information of the first model execution process.

[0083] This application does not limit the reporting method of the information in the first capability information. For example, the first capability information can directly report the specific numerical value of the corresponding information. Alternatively, the first capability information can indicate the range or level corresponding to the specific numerical value of the information to be reported. Taking the case where the first capability information includes the video memory of the terminal device as an example, if the current video memory of the terminal device is 1G, the first capability information can directly indicate 1G, or the first capability information can indicate the level of 500M~1.5G corresponding to 1G. The following examples illustrate the level or range of the information included in the first capability information.

[0084] In some embodiments, the level of the terminal device's video memory indicator may include one or more of: less than 500M, 500M to 1G, 1.5G to 2.5G, and greater than 2.5G. The level of the video memory indicator can be represented by two bits. For example, {00} may indicate that the video memory is less than 500M; {01} may indicate that the video memory is between 500M and 1.5G; {10} may indicate that the video memory is between 1.5G and 2.5G; and {11} may indicate that the video memory is greater than 3G.

[0085] In some embodiments, the level of the computing power indicator of the terminal device may include one or more of: less than 5 TOPS, 5 TOPS to 10 TOPS, 10 TOPS to 30 TOPS, and greater than 30 TOPS. The level of the computing power indicator can be represented by two bits. {00} can indicate that the computing power is less than 5 TOPS; {01} can indicate that the computing power is between 5 TOPS and 10 TOPS; {10} can indicate that the computing power is between 10 TOPS and 30 TOPS; and {11} can indicate that the computing power is greater than 30 TOPS.

[0086] In some embodiments, the maximum video memory allocated by the terminal device to the first model may include one or more of: less than 50M, 50M to 150M, 150M to 300M, and greater than 300M. The maximum video memory allocated to the first model may be represented by two bits. For example, {00} may indicate that the maximum video memory allocated to the first model is less than 50M; {01} may indicate that the maximum video memory allocated to the first model is between 50M and 150M; {10} may indicate that the maximum video memory allocated to the first model is between 150M and 300M; and {11} may indicate that the maximum video memory allocated to the first model is greater than 300M.

[0087] In some embodiments, the level of the computing power upper limit assigned to the first model by the terminal device may include: less than 1TOPS, 1TOPS to 8TOPS, 8TOPS to 16TOPS, and greater than 16TOPS. The level of the computing power upper limit assigned to the first task by the terminal device can be represented by 2 bits. For example, {00} can indicate that the computing power upper limit assigned to the first task is less than 1TOPS; {01} can indicate that the computing power upper limit assigned to the first task is between 1TOPS and 8TOPS; {10} can indicate that the computing power upper limit assigned to the first task is between 8TOPS and 16TOPS; and {11} can indicate that the computing power upper limit assigned to the first task is greater than 16TOPS.

[0088] In some embodiments, the level of the video memory peak value of the first model reasoning process may include: less than 500M, 500M to 1G, 1.5G to 2.5G, and greater than 2.5G. The level of the video memory peak value of the first model reasoning process can be represented by 2 bits. For example, {00} can indicate that the video memory peak value of the first model reasoning process is less than 500M; {01} can indicate that the video memory peak value of the first model reasoning process is within the range of 500M to 1.5G; {10} can indicate that the video memory peak value of the first model reasoning process is within the range of 1.5G to 2.5G; and {11} can indicate that the video memory peak value of the first model reasoning process is greater than 3G.

[0089] In some embodiments, the level to which the average single inference time of the first model belongs may include: one or more of: higher than 1e-4s, 1e-4s to 1e-5s, 1e-5s to 1e-6s, and lower than 1e-6s. The level to which the average single inference time of the first model belongs may be represented by 2 bits. For example, {00} may indicate that the average single inference time of the first model is higher than 1e-4s; {01} may indicate that the average single inference time of the first model is between 1e-4s and 1e-5s; {10} may indicate that the average single inference time of the first model is between 1e-5s and 1e-6s; and {11} may indicate that the average single inference time of the first model is lower than 1e-6s.

[0090] In some embodiments, the level of the video memory peak value during the online training process of the first model may include: less than 500M, 500M to 1G, 1.5G to 2.5G, and greater than 2.5G. The level of the video memory peak value during the online training process of the first model can be represented by 2 bits. For example, {00} can indicate that the video memory peak value during the online training process of the first model is less than 500M; {01} can indicate that the video memory peak value during the online training process of the first model is within the range of 500M to 1.5G; {10} can indicate that the video memory peak value during the online training process of the first model is within the range of 1.5G to 2.5G; and {11} can indicate that the video memory peak value during the online training process of the first model is greater than 3G.

[0091] In some embodiments, the level to which the average single-step training duration of the first model belongs may include: one or more of: higher than 1e-4s, 1e-4s to 1e-5s, 1e-5s to 1e-6s, and lower than 1e-6s. The level to which the average single-step training duration of the first model belongs may be represented by 2 bits. For example, {00} may indicate that the average single-step training duration of the first model is higher than 1e-4s; {01} may indicate that the average single-step training duration of the first model is between 1e-4s and 1e-5s; {10} may indicate that the average single-step training duration of the first model is between 1e-5s and 1e-6s; and {11} may indicate that the average single-step training duration of the first model is lower than 1e-6s.

[0092] In some embodiments, the first capability information may also include the type of the first capability information. This type can be represented by two bits. For example, {00} may represent type 1, which indicates that the terminal device has autonomous resource allocation capabilities; {01} may represent type 2, which indicates that the terminal device does not have autonomous resource allocation capabilities; {10} may represent type 3, which indicates that the first capability information is used to provide feedback on the inference operation status; and {11} may represent type 4, which indicates that the first capability information is used to provide feedback on the online training operation status. It is understood that type 4 first capability information may only be activated and reported when there is an online training task.

[0093] As an implementation method, the first capability information can be indicated by 6 bits. The first 2 bits can be used to indicate the type of the first capability information. The last 4 bits can be used to indicate the specific content. For example, in the case where the first capability information is type 1, the last 4 bits can be used to indicate the level of the terminal device's video memory indicator and the level of the computing power indicator. In the case where the first capability information is type 2, the last 4 bits can be used to indicate the level of the maximum video memory allocated to the first model by the terminal device and the level of the computing power upper limit. In the case where the first capability information is type 3, the last 4 bits can be used to indicate the level of the video memory peak value and the level of the average single inference duration of the first model inference process. In the case where the first capability information is type 4, the last 4 bits can be used to indicate the level of the video memory peak value and the level of the average single-step training duration of the first model online training process.

[0094] It should be noted that the present application does not limit the tasks performed by the first model. For example, the first model can be used for CSI feedback or beam prediction. Taking the first model being used for CSI feedback as an example, when the terminal device performs the CSI prediction task through the first model, the first model may include an encoder; when the terminal device performs online training on the model corresponding to the CSI feedback task, the first model may include an encoder and a decoder. During the prediction process, the encoder and decoder are deployed on the terminal device and the network device side respectively, and the online training process requires joint training of the encoder and decoder. Therefore, the first model deployed on the device side may include the same decoder as that on the network device side. During the online training process, only the parameters of the encoder may be updated. Taking the first model being used for beam management as an example, the first model may be a beam prediction model.

[0095] It should be noted that the first model may be an AI model. For example, the first model may be a neural network model or a machine learning model. In the case where the first model is an AI model, the first capability information may be used to indicate information about the AI-related capabilities of the terminal device.

[0096] It should be noted that this application does not limit the type of message carrying the first capability information. The first capability information can be carried in UCI or other uplink signaling.

[0097] The first capability information can be reported periodically or triggered by an event. For example, when the first capability information changes, the reporting of the first capability information can be triggered. The reporting period of the first capability information can meet one of the following: a preset value, a network configuration, or a preset value. This application does not limit the reporting period of the first capability information. For example, the period can be 30ms, 50ms, or 70ms.

[0098] In some embodiments, the first capability information may be reported when the first model has not yet been deployed on the terminal device. For example, in response to the terminal device accessing the network, the terminal device may report the first capability information. In some embodiments, the first capability information may be reported when the first model has already been deployed on the terminal device.

[0099] The first capability information can provide dynamic change information of the resources of the terminal device and / or the actual operation status of the first model. Based on the received first capability information, the network device can determine the scale, parameters or training model strategy of the model for the terminal device. Based on the first capability information, the network device can further reasonably select a model and / or an online training strategy. For example, the network device can determine the inference delay based on the inference time indicated by the first capability information, thereby selecting the largest model and / or the best online training strategy that meets the inference delay requirement. Larger models tend to have higher performance (such as accuracy), and the best training strategy can achieve higher online update efficiency.

[0100] From this, it can be seen that based on the first capability information, the network device can configure a more suitable model and / or online training strategy for the terminal device, so as to maximize the capabilities of the terminal device as much as possible, thereby improving the operating performance of the first model while avoiding problems such as operation failure of the first model, inference timeout, and failure of online training to converge within the specified time.

[0101] When the network device determines the model and / or online training strategy, the network device may send first indication information to the terminal device.

[0102] In some embodiments, the first indication information can be used to deploy or update a model on a terminal device. For example, if the first model has not yet been deployed on the terminal device, the first indication information can be used to instruct the terminal device to deploy the first model. If the first model has already been deployed on the terminal device, the first indication information can be used to instruct the terminal device to update the first model to a second model.

[0103] When the terminal device performs online training on the first model, the first indication information may be used to indicate an online training strategy for the first model. The online training strategy may be used to instruct the terminal device on how to perform online training on the first model. The online training strategy may include, for example, one or more of the following information: the frequency of online training, the number of samples used in online training, starting parameters for online training, whether online training is periodic, whether online training is performed sample by sample, etc.

[0104] The method provided by the present application is described in detail below with reference to Figure 7. The method shown in Figure 7 may include steps S710 to S750.

[0105] Step S710: The terminal device sends first capability information to the network device.

[0106] Step S710 may occur in the initial access phase. In the initial access phase, the model corresponding to the first task has not yet been configured on the terminal device.

[0107] In step S710 , the first capability information may be used to indicate information about resources that can be used by the first model or the first task.

[0108] If the terminal device does not have autonomous resource allocation capabilities, the information about resources available to the first model may include: the class of the terminal device's current video memory indicator and / or the class of the terminal device's current computing power indicator. In this case, the type of the first capability information may be marked as Type 1.

[0109] If the terminal device has autonomous resource allocation capabilities, the information about resources available to the first model may include: the maximum level of video memory allocated by the terminal device to the first task, and / or the upper limit of computing power allocated by the terminal device to the first task. In this case, the type of the first capability information may be marked as Type 2.

[0110] According to the content and / or type of the first capability information, the network device may determine the scale of the first model or the online training strategy of the first model.

[0111] Step S720: The network device sends first indication information to the terminal device.

[0112] In step S720, the first indication information may be used to indicate deployment of the first model on the terminal device; and / or, the first indication information may be used to indicate an online training strategy for the first model.

[0113] In step S730, the terminal device may complete the deployment of the first model and perform one or more of the following operations: use the first model for inference; and perform online training on the first model.

[0114] In some embodiments, the terminal device can start online training using real data collected from a real system.

[0115] After the first model is deployed, the terminal device may report the first capability information once or multiple times. That is, the terminal device may execute steps S741 to S749. It should be noted that steps S741 to S749 in Figure 7 are only examples, and this application does not limit the number of times the first capability information may be reported.

[0116] In steps S741 to S749, the type of the first capability information may be type 1 or type 2 as described above. The first capability information may also be used to indicate information about the operation process of the first model, that is, the type of the first capability information may be type 3 or type 4 as described below.

[0117] When the running process is an inference process, the first capability information may include: the level of the peak memory usage during the inference process, and / or the average inference duration. In this case, the type of the first capability information may be type 3.

[0118] In the case where the running process is an online training process, the first capability information may include: the level of the video memory peak of the online training process, and / or the average single-step online training time. In this case, the type of the first capability information may be type 4.

[0119] Based on part or all of the first capability information reported in steps S741 to S749, the network device may perform one or more of the following operations: determining whether to update the first model, determining the scale of the updated second model, and adjusting the online training strategy of the first model.

[0120] Step S750: The network device sends first indication information to the terminal device.

[0121] In step S750, the first indication information may include one or more of the following information: the first model deployed on the terminal device is updated to the second model, and the online training strategy of the first model.

[0122] Among them, steps 710 to S720 may belong to the first stage. Steps S730 to S750 may belong to the second stage. It can be understood that in both the first and second stages, the terminal device can send the first capability information. The type and content of the first capability information may be different in different stages. For example, in the first stage, the terminal device may report first capability information of type 1 and / or type 2. In the second stage, the terminal device may report first capability information of any one or more types of type 1, type 2, type 3, and type 4. Among them, the first capability information of type 3 or type 4 may be reported simultaneously or successively with the first capability information of type 1 or type 2.

[0123] The present application is briefly introduced below with reference to FIG8-FIG10 , in combination with the first model being a CSI feedback model or a beam prediction model.

[0124] Figure 8 takes the first model as an example of a first encoder model for CSI feedback. The method shown in Figure 8 may include steps S810 to S850.

[0125] Step S810: The terminal device sends first capability information to the network device.

[0126] Based on the first capability information, the network device determines a first encoder model.

[0127] Step S820: The network device deploys a first encoder model for the terminal device.

[0128] Step S830: The terminal device performs inference using the first encoder model.

[0129] In steps S841 to S849, the terminal device sends the first capability information to the network device once or multiple times.

[0130] According to the first capability information in steps S841 to S849, the network device determines to update the first encoder model on the terminal device to the second encoder model.

[0131] Step S850: The network device updates the first encoder model deployed by the terminal device to a second encoder model.

[0132] Figure 9 takes the first model as the first encoder model and the first decoder model for CSI feedback as an example. The method shown in Figure 9 may include steps S910 to S950.

[0133] Step S910: The terminal device sends first capability information to the network device.

[0134] Based on the first capability information, the network device determines a first encoder model and a first decoder model.

[0135] Step S920: The network device deploys a first encoder model and a first decoder model for the terminal device.

[0136] In step S930 , the terminal device performs online training on the first encoder model and the first decoder model.

[0137] In steps S941 to S949, the terminal device sends the first capability information to the network device once or multiple times.

[0138] According to the first capability information in steps S941 to S949, the network device determines to update the first encoder model and the first decoder model on the terminal device to the second encoder model and the second decoder model; or, the network device determines the online training strategy of the first encoder model and the first decoder model.

[0139] In step S950, the network device updates the first encoder model and the first decoder model deployed by the terminal device to a second encoder model and a second decoder model; or, the network device sends an online training strategy to the terminal device.

[0140] Figure 10 takes the first model as the first beam prediction model as an example. The method shown in Figure 10 may include steps S1010 to S1050.

[0141] Step S1010: The terminal device sends first capability information to the network device.

[0142] Based on the first capability information, the network device determines a first beam prediction model.

[0143] Step S1020: The network device deploys a first beam prediction model for the terminal device.

[0144] Step S1030: The terminal device performs online training on the first beam prediction model; and / or, the terminal device performs inference based on the first beam prediction model.

[0145] In steps S1041 to S1049, the terminal device sends the first capability information to the network device once or multiple times.

[0146] According to the first capability information in steps S1041 to S1049, the network device determines to update the first beam prediction model on the terminal device to the second beam prediction model; or, the network device determines an online training strategy for the first beam prediction model.

[0147] In step S1050, the network device updates the first beam prediction model deployed by the terminal device to a second beam prediction model; or, the network device sends an online training strategy to the terminal device.

[0148] The method embodiment of the present application is described in detail above. The following device embodiment of the present application is described in detail in conjunction with Figures 11 to 13. It should be understood that the description of the method embodiment corresponds to the description of the device embodiment. Therefore, for parts not described in detail, reference can be made to the above method embodiment.

[0149] FIG11 is a schematic structural diagram of a terminal device 1100 provided in an embodiment of the present application. The terminal device 1100 may include a first sending unit 1110 .

[0150] The first sending unit 1110 is configured to send first capability information; wherein the first capability information is associated with the first model, and the first capability information is configured to indicate one of the following information: information about resources that can be used by the first model;

[0151] Information about the running process of the first model.

[0152] In some embodiments, the resources that can be used by the first model include: information about resources that can be used by the terminal device; and / or information about resources allocated by the terminal device to the first model.

[0153] In some embodiments, the operation process of the first model includes: an inference process of the first model; and / or an online training process of the first model.

[0154] In some embodiments, the information of the running process includes one or more of the following information: the duration of the running process; and information about resources occupied by the running process.

[0155] In some embodiments, the duration of the running process includes one or more of the following durations: the average duration of the first model running once, the duration of the first model running one or more times continuously, and the total duration of the first model running.

[0156] In some embodiments, the information about resources occupied by the running process includes: a peak value of resources occupied by the running process.

[0157] In some embodiments, the resources include one or more of the following: memory, video memory, computing power, and number of threads.

[0158] In some embodiments, it also includes: a first receiving unit for receiving first indication information; wherein, the first indication information is determined based on the first capability information, and the first indication information is used to indicate one or more of the following: deploying a first model on the terminal device; updating the deployed first model to a second model on the terminal device; and an online training strategy for the first model.

[0159] In some embodiments, the first capability information is sent in response to a change in the first capability information; and / or the first capability information is sent periodically.

[0160] In some embodiments, it is characterized in that the first model is an AI model.

[0161] FIG12 is a schematic structural diagram of a network device 1200 provided in an embodiment of the present application. The network device 1200 may include a second receiving unit 1210 .

[0162] The second receiving unit 1210 is used to receive first capability information sent by the terminal device; wherein, the first capability information is associated with the first model, and the first capability information is used to indicate one of the following information: information about resources that can be used by the first model; information about the operation process of the first model.

[0163] In some embodiments, the resources that can be used by the first model include: information about resources that can be used by the terminal device; and / or information about resources allocated by the terminal device to the first model.

[0164] In some embodiments, the operation process of the first model includes: the reasoning process of the first model; and / or,

[0165] The online training process of the first model.

[0166] In some embodiments, the information of the running process includes one or more of the following information: the duration of the running process; and information about resources occupied by the running process.

[0167] In some embodiments, the duration of the running process includes one or more of the following durations: the average duration of the first model running once, the duration of the first model running one or more times continuously, and the total duration of the first model running.

[0168] In some embodiments, the information about resources occupied by the running process includes: a peak value of resources occupied by the running process.

[0169] In some embodiments, the resources include one or more of the following: memory, video memory, computing power, and number of threads.

[0170] In some embodiments, it also includes: a second sending unit, used to send first indication information to the terminal; wherein, the first indication information is determined based on the first capability information, and the first indication information is used to indicate one or more of the following: deploying a first model on the terminal device; updating the deployed first model to a second model on the terminal device; online training strategy of the first model.

[0171] In some embodiments, the first capability information is sent in response to a change in the first capability information; and / or the first capability information is sent periodically.

[0172] In some embodiments, the first model is an AI model.

[0173] In an optional embodiment, the first sending unit 1110 or the second receiving unit 1210 may be a transceiver 1330. The terminal device 1100 or the network device 1200 may further include a memory 1320 and a processor 1310, as specifically shown in FIG13 .

[0174] Figure 13 is a schematic block diagram of a communication device according to an embodiment of the present application. The dashed lines in Figure 13 indicate that the unit or module is optional. Apparatus 1300 may be used to implement the method described in the above method embodiment. Apparatus 1300 may be a chip, a terminal device, or a network device.

[0175] The device 1300 may include one or more processors 1310. The processor 1310 may support the device 1300 to implement the method described in the above method embodiment. The processor 1310 may be a general-purpose processor or a special-purpose processor. For example, the processor may be a central processing unit (CPU). Alternatively, the processor may be another general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic device, discrete hardware component, etc. The general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc.

[0176] The apparatus 1300 may further include one or more memories 1320. The memories 1320 store programs that can be executed by the processor 1310, causing the processor 1310 to perform the methods described in the above method embodiments. The memories 1320 may be independent of the processor 1310 or integrated into the processor 1310.

[0177] The apparatus 1300 may further include a transceiver 1330. The processor 1310 may communicate with other devices or chips via the transceiver 1330. For example, the processor 1310 may transmit and receive data with other devices or chips via the transceiver 1330.

[0178] The present application also provides a computer-readable storage medium for storing a program. The computer-readable storage medium can be applied to a terminal or network device provided in the present application, and the program enables a computer to execute the method performed by the terminal or network device in each embodiment of the present application.

[0179] The present application also provides a computer program product. The computer program product includes a program. The computer program product can be applied to a terminal or network device provided in the present application, and the program causes a computer to execute the method performed by the terminal or network device in each embodiment of the present application.

[0180] The embodiments of the present application also provide a computer program. The computer program can be applied to the terminal or network device provided in the embodiments of the present application, and the computer program enables a computer to execute the method performed by the terminal or network device in each embodiment of the present application.

[0181] It should be understood that the terms "system" and "network" in this application can be used interchangeably. In addition, the terms used in this application are only used to explain the specific embodiments of this application and are not intended to limit this application. The terms "first", "second", "third", and "fourth" in the specification and claims of this application and the accompanying drawings are used to distinguish different objects rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions.

[0182] In the embodiments of this application, the term "indication" may refer to a direct indication, an indirect indication, or an indication of an association. For example, "A indicates B" may refer to a direct indication of B, e.g., B can obtain information through A; it may refer to an indirect indication of B, e.g., A indicates C, e.g., B can obtain information through C; or it may refer to an association between A and B.

[0183] In the embodiment of the present application, "B corresponding to A" means that B is associated with A and B can be determined based on A. However, it should be understood that determining B based on A does not mean determining B based solely on A, but B can also be determined based on A and / or other information.

[0184] In the embodiments of the present application, the term "corresponding" may indicate a direct or indirect correspondence between the two, or an association relationship between the two, or a relationship between indication and indication, configuration and configuration, etc.

[0185] In the embodiments of the present application, "pre-definition" or "pre-configuration" may be implemented by pre-storing corresponding codes, tables, or other methods that can be used to indicate relevant information in a device (e.g., a terminal device and a network device). The present application does not limit the specific implementation method. For example, pre-definition may refer to information defined in a protocol.

[0186] In the embodiments of the present application, the “protocol” may refer to a standard protocol in the communications field, for example, it may include an LTE protocol, an NR protocol, and related protocols used in future communication systems, and the present application does not limit this.

[0187] In the embodiments of this application, the term "and / or" is simply a description of the association relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this document generally indicates that the related objects are in an "or" relationship.

[0188] In various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0189] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0190] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0191] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0192] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part 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, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. 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, computer, server or data center to another website, computer, server or data center via a wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) method. The computer-readable storage medium can be any available medium that can be read by a computer or a data storage device such as a server or data center that includes one or more available media integrated therein. The available medium may be a magnetic medium (eg, a floppy disk, a hard disk, a magnetic tape), an optical medium (eg, a digital versatile disc (DVD)), or a semiconductor medium (eg, a solid state disk (SSD)).

[0193] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. A method for communication, characterized in that include: The terminal device sends first capability information; The first capability information is associated with the first model, and the first capability information is used to indicate one of the following information: information about resources that can be used by the first model; Information about the running process of the first model.

2. The method according to claim 1, characterized in that The resources that can be used in the first model include: Information about resources that can be used by the terminal device; and / or, Information about resources allocated by the terminal device to the first model.

3. The method according to claim 1 or 2, characterized in that: The operation process of the first model includes: the reasoning process of the first model; and / or, The online training process of the first model.

4. The method according to any one of claims 1 to 3, characterized in that The information of the operation process includes one or more of the following information: the duration of the running process; Information about resources occupied by the running process.

5. The method according to claim 4, characterized in that The duration of the running process includes one or more of the following durations: the average duration of the first model running once, the duration of the first model running one or more times continuously, and the total duration of the first model running.

6. The method according to claim 4 or 5, characterized in that: The information about resources occupied by the running process includes: a peak value of resources occupied by the running process.

7. The method according to any one of claims 1 to 6, characterized in that The resources include one or more of the following: memory, video memory, computing power, and number of threads.

8. The method according to any one of claims 1 to 7, characterized in that Also includes: The terminal receives first indication information; The first indication information is determined based on the first capability information, and the first indication information is used to indicate one or more of the following: Deploy a first model on the terminal device; Updating the deployed first model to the second model on the terminal device; The online training strategy of the first model.

9. The method according to any one of claims 1 to 8, characterized in that The first capability information is sent in response to a change in the first capability information; and / or The first capability information is sent periodically.

10. The method according to any one of claims 1 to 9, characterized in that The first model is an artificial intelligence AI model.

11. A method for communication, characterized in that include: The network device receives the first capability information sent by the terminal device; The first capability information is associated with the first model, and the first capability information is used to indicate one of the following information: information about resources that can be used by the first model; Information about the running process of the first model.

12. The method according to claim 11, characterized in that The resources that can be used in the first model include: Information about resources that can be used by the terminal device; and / or, Information about resources allocated by the terminal device to the first model.

13. The method according to claim 11 or 12, characterized in that: The operation process of the first model includes: the reasoning process of the first model; and / or, The online training process of the first model.

14. The method according to any one of claims 11 to 13, characterized in that The information of the operation process includes one or more of the following information: the duration of the running process; Information about resources occupied by the running process.

15. The method according to claim 14, characterized in that The duration of the running process includes one or more of the following durations: the average duration of the first model running once, the duration of the first model running one or more times continuously, and the total duration of the first model running.

16. The method according to claim 14 or 15, characterized in that The information about resources occupied by the running process includes: a peak value of resources occupied by the running process.

17. The method according to any one of claims 11 to 16, characterized in that The resources include one or more of the following: memory, video memory, computing power, and number of threads.

18. The method according to any one of claims 11 to 17, characterized in that Also includes: The network device sends first indication information to the terminal; The first indication information is determined based on the first capability information, and the first indication information is used to indicate one or more of the following: Deploy a first model on the terminal device; Updating the deployed first model to the second model on the terminal device; The online training strategy of the first model.

19. The method according to any one of claims 11 to 18, characterized in that The first capability information is sent in response to a change in the first capability information; and / or The first capability information is sent periodically.

20. The method according to any one of claims 11 to 19, characterized in that The first model is an artificial intelligence AI model.

21. A terminal device, characterized in that: include: A first sending unit, configured to send first capability information; The first capability information is associated with the first model, and the first capability information is used to indicate one of the following information: information about resources that can be used by the first model; Information about the running process of the first model.

22. The terminal device according to claim 21, characterized in that: The resources that can be used in the first model include: Information about resources that can be used by the terminal device; and / or, Information about resources allocated by the terminal device to the first model.

23. The terminal device according to claim 21 or 22, characterized in that: The operation process of the first model includes: the reasoning process of the first model; and / or, The online training process of the first model.

24. The terminal device according to any one of claims 21 to 23, characterized in that: The information of the operation process includes one or more of the following information: the duration of the running process; Information about resources occupied by the running process.

25. The terminal device according to claim 24, characterized in that: The duration of the running process includes one or more of the following durations: the average duration of the first model running once, the duration of the first model running one or more times continuously, and the total duration of the first model running.

26. The terminal device according to claim 24 or 25, characterized in that: The information about resources occupied by the running process includes: a peak value of resources occupied by the running process.

27. The terminal device according to any one of claims 21 to 26, characterized in that: The resources include one or more of the following: memory, video memory, computing power, and number of threads.

28. The terminal device according to any one of claims 21 to 27, characterized in that: Also includes: A first receiving unit, configured to receive first indication information; The first indication information is determined based on the first capability information, and the first indication information is used to indicate one or more of the following: Deploy a first model on the terminal device; Updating the deployed first model to the second model on the terminal device; The online training strategy of the first model.

29. The terminal device according to any one of claims 21 to 28, characterized in that: The first capability information is sent in response to a change in the first capability information; and / or The first capability information is sent periodically.

30. The terminal device according to any one of claims 21 to 29, characterized in that: The first model is an artificial intelligence AI model.

31. A network device, characterized in that: include: A second receiving unit, configured to receive first capability information sent by a terminal device; The first capability information is associated with the first model, and the first capability information is used to indicate one of the following information: information about resources that can be used by the first model; Information about the running process of the first model.

32. The network device according to claim 31, characterized in that: The resources that can be used in the first model include: Information about resources that can be used by the terminal device; and / or, Information about resources allocated by the terminal device to the first model.

33. The network device according to claim 31 or 32, characterized in that: The operation process of the first model includes: the reasoning process of the first model; and / or, The online training process of the first model.

34. The network device according to any one of claims 31 to 33, characterized in that: The information of the operation process includes one or more of the following information: the duration of the running process; Information about resources occupied by the running process.

35. The network device according to claim 34, characterized in that: The duration of the running process includes one or more of the following durations: the average duration of the first model running once, the duration of the first model running one or more times continuously, and the total duration of the first model running.

36. The network device according to claim 34 or 35, characterized in that: The information about resources occupied by the running process includes: a peak value of resources occupied by the running process.

37. The network device according to any one of claims 31 to 36, characterized in that: The resources include one or more of the following: memory, video memory, computing power, and number of threads.

38. The network device according to any one of claims 31 to 37, characterized in that: Also includes: A second sending unit, configured to send first indication information to the terminal; The first indication information is determined based on the first capability information, and the first indication information is used to indicate one or more of the following: Deploy a first model on the terminal device; Updating the deployed first model to the second model on the terminal device; The online training strategy of the first model.

39. The network device according to any one of claims 31 to 38, characterized in that: The first capability information is sent in response to a change in the first capability information; and / or The first capability information is sent periodically.

40. The network device according to any one of claims 31 to 39, characterized in that: The first model is an artificial intelligence AI model.

41. A terminal device, characterized in that: The method comprises a memory and a processor, wherein the memory is used to store a program, and the processor is used to call the program in the memory so that the terminal device executes the method according to any one of claims 1 to 10.

42. A network device, characterized in that: The network device comprises a memory and a processor, wherein the memory is used to store a program, and the processor is used to call the program in the memory so that the network device executes the method according to any one of claims 11 to 20.

43. A device, characterized in that The device comprises a processor, configured to call a program from a memory so as to enable the device to execute the method according to any one of claims 1 to 20.

44. A chip, characterized in that: The device comprises a processor, which is used to call a program from a memory so that a device equipped with the chip executes a method as claimed in any one of claims 1 to 20.

45. A computer-readable storage medium, characterized in that A program is stored thereon, and the program enables a computer to execute the method according to any one of claims 1 to 20.

46. ​​A computer program product, characterized in that The method comprises a program which causes a computer to execute the method according to any one of claims 1 to 20.

47. A computer program, characterized in that The computer program enables a computer to execute the method according to any one of claims 1 to 20.

Images