CN118202645A

CN118202645A - AI model transmission method and device

Info

Publication number: CN118202645A
Application number: CN202280002427.8A
Authority: CN
Inventors: 牟勤
Original assignee: Beijing Xiaomi Mobile Software Co Ltd
Current assignee: Beijing Xiaomi Mobile Software Co Ltd
Priority date: 2022-06-30
Filing date: 2022-06-30
Publication date: 2024-06-14
Also published as: WO2024000532A1

Abstract

The method and the device for transmitting the AI model can be applied to the AI model transmission of an AI model interaction party, and the method comprises the following steps: the providing node determines a data type (201) for digitally representing the AI model, converts the AI model into a corresponding AI model bit stream (202) according to the data type, and when the receiving node responds to the AI model bit stream sent by the AI model providing node, the receiving node inverts the AI model bit stream according to a preset conversion rule to obtain the corresponding AI model, and uses the AI model (1101) at the AI receiving node, so that the processing capacity of the AI model receiving node can be met, and the AI model can be flexibly represented according to the data type.

Description

AI model transmission method and device

Technical Field

The present application relates to the field of communications technologies, and in particular, to a transmission method and apparatus for an AI model.

Background

In recent years, artificial intelligence (ARTIFICIAL INTELLIGENCE, AI) technology has been breaking through in a number of fields. The continuous development in the fields of intelligent voice, computer vision and the like not only brings various colorful applications for the intelligent terminal, but also has wide application in a plurality of fields of education, traffic, home, medical treatment, retail, security and the like, brings convenience to life of people, and simultaneously promotes industry upgrading of various industries. AI technology is also accelerating cross-penetration with other discipline fields, and its development incorporates knowledge of different disciplines while also providing new directions and methods for development of different disciplines.

Two important phases are involved in the artificial intelligence (ARTIFICIAL INTELLIGENCE, AI) technology, the first phase is the training phase of the model, i.e., the phase of obtaining the model; the second phase is the deployment phase of the model, i.e., the inference application phase of the model. When training and reasoning of the model are not in the same node, the model is transmitted from the trained node to the reasoning node, and when the model is transmitted from one node to another node, the AI model needs to be digitalized, namely the structure of the model, parameters of the model and the like are expressed by numbers.

In the AI model node transmission process, since different terminal equipment types support different data types, for example, the terminal equipment only supports 8-bit integer type, and the processor supports 16-bit floating point type. Different data types can lead to different inference complexity.

Disclosure of Invention

The embodiment of the disclosure provides a transmission method and a device for an AI model, which can be applied to the fields of artificial intelligence (ARTIFICIAL INTELLIGENCE, AI) and the like, and provides a flexible AI model representation method, wherein the AI model is digitally represented according to the data type, so that the processing capacity of an AI model receiving node can be met, and the AI model can be flexibly represented according to the data type.

In a first aspect, an embodiment of the present disclosure provides a transmission method of an AI model, the method being performed by a providing node, the method including:

Determining a data type for digitally representing the AI model;

And converting the AI model into a corresponding AI model bit stream according to the data type.

In the technical scheme, after the data type for digitally representing the AI model is determined, the AI model is converted into the corresponding AI model bit stream according to the data type by the providing node, so that the data type used when the AI model interaction party synchronizes the AI model transmission is realized, the processing capacity of the AI model receiving node can be met, and the AI model can be flexibly represented according to the data type.

In one implementation, the determining the data type that digitally represents the AI model includes:

and determining the data type digitally representing the AI model according to the supporting capacity of the AI model receiving node for the data type.

In one implementation, the method further comprises:

And receiving the supporting capability of the AI model receiving node to the data type, which is sent by the AI model receiving node.

In one implementation, the determining the data type digitally representing the AI model according to the AI model receiving node's supporting capability for the data type includes:

And according to the supporting capacity of the data types of the AI model receiving nodes, determining the data type corresponding to the data type with the maximum supporting capacity of the AI model receiving nodes as the data type for digitally representing the AI model.

According to the supporting capacity of the AI model receiving node to the data types, selecting one data type supported by the AI model providing node as the data type for digitally representing the AI model.

In one implementation, the determining the data type of the AI model-receiving node includes:

And determining the data type for digitally representing the AI model according to the power consumption and/or the storage capacity reported by the AI model receiving node.

According to the service requirement, the data type of the digital representation of the AI model is determined.

And determining the data type represented by the AI model according to the resource overhead of the AI model.

In one implementation, the method further comprises:

And sending indication information of the data type for digitally representing the AI model to the AI model receiving node, wherein the indication information comprises the data type for digitally representing the AI model.

In a second aspect, there is also provided a transmission method of an AI model, the method being performed by a receiving node, the method including:

And in response to receiving the AI model bit stream sent by the AI model providing node, carrying out inverse transformation on the AI model bit stream according to a preset transformation rule to obtain a corresponding AI model, and using the AI model at the AI receiving node.

In the technical scheme, when the receiving node responds to the received AI model bit stream sent by the AI model providing node, the AI model bit stream is reversely converted according to a preset conversion rule to obtain a corresponding AI model, and the AI model is used by the AI receiving node to realize the data type used when the AI model interaction party synchronizes the AI model transmission, so that the processing capacity of the AI model receiving node can be met, and the AI model can be flexibly represented according to the data type.

In one implementation, the method further comprises:

And receiving indication information of the data type which is sent by the AI model providing node and digitally represents the AI model, wherein the indication information comprises the data type which digitally represents the AI model.

In one implementation, the performing inverse transformation on the AI model bit stream according to the predetermined transformation rule to obtain a corresponding AI model includes:

And responding to the indication information, and according to the data type which is included in the indication information and used for digitally representing the AI model, carrying out inverse conversion on the AI model bit stream according to a preset conversion rule to obtain a corresponding AI model.

In one implementation, the method further comprises:

And providing the support capability of reporting the data type to the AI model by the node.

In one implementation, the method further comprises:

Providing node reporting to the AI model is based on power consumption and/or storage capabilities.

In a third aspect, there is also provided a transmission apparatus of an AI model, the apparatus being provided at a providing node, the apparatus comprising:

A determining unit for determining a data type for digitally representing the AI model;

And the conversion unit is used for converting the AI model into a corresponding AI model bit stream according to the data type.

In one implementation, the determining unit is further configured to determine the data type digitally representing the AI model according to the supporting capability of the AI model receiving node for the data type.

In one implementation, the method further comprises:

And the receiving unit is used for receiving the supporting capability of the AI model receiving node to the data type, which is sent by the AI model receiving node.

In one implementation manner, the determining unit is further configured to determine, according to the supporting capability of the data type of the AI model receiving node, a data type corresponding to a data type with a maximum supporting capability of the AI model receiving node as a data type that digitally represents the AI model.

In an implementation manner, the determining unit is further configured to select, according to the capability of the AI model receiving node to support data types, a data type supported by the AI model providing node as a data type digitally representing the AI model.

In one implementation manner, the determining unit is further configured to determine a data type digitally representing the AI model according to power consumption and/or storage capability reported by the AI model receiving node.

In an implementation manner, the determining unit is further configured to determine a data type digitally representing the AI model according to the service requirement.

In an implementation manner, the determining unit is further configured to determine a data type represented by the AI model according to resource overhead of the AI model.

In one implementation, the method further comprises:

and the sending unit is used for sending indication information of the data type for digitally representing the AI model to the AI model receiving node, wherein the indication information comprises the data type for digitally representing the AI model.

In a fourth aspect, there is also provided a transmission apparatus of an AI model, the apparatus being disposed at a receiving node, the apparatus including:

The conversion unit is used for reversely converting the AI model bit stream according to a preset conversion rule to obtain a corresponding AI model when the AI model bit stream sent by the AI model providing node is received;

A use unit for using the AI model at the AI receiving node.

In one implementation, the method further comprises:

And the receiving unit is used for receiving the indication information of the data type which is sent by the AI model providing node and digitally represents the AI model, wherein the indication information comprises the data type which digitally represents the AI model.

In one implementation manner, the conversion unit is further configured to, in response to the indication information, reverse the bit stream of the AI model according to a predetermined conversion rule according to the indication information including a data type digitally representing the AI model, so as to obtain a corresponding AI model.

In one implementation, the apparatus further comprises:

and the first reporting unit is used for providing the supporting capability of reporting the data type for the AI model.

In one implementation, the apparatus further comprises:

and the second reporting unit is used for providing the node reporting based on the power consumption and/or the storage capacity for the AI model.

In a fifth aspect, an embodiment of the present disclosure provides an AI model transmission apparatus, including a processor and a memory, where the memory stores a computer program, and the processor executes the computer program stored in the memory to cause the apparatus to perform the method according to the first or second aspect.

In a sixth aspect, an embodiment of the present disclosure provides a transmission apparatus for an AI model, including: a processor and interface circuit;

the interface circuit is used for receiving code instructions and transmitting the code instructions to the processor;

the processor is configured to execute the code instructions to perform the method according to the first or second aspect.

In a seventh aspect, embodiments of the present disclosure provide a computer readable storage medium storing instructions that, when executed, cause a method according to the first or second aspect to be implemented.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments or the background of the present disclosure, the following description will explain the drawings that are required to be used in the embodiments or the background of the present disclosure.

Fig. 1 is a schematic flow chart of a transmission method of an AI model according to an embodiment of the disclosure;

FIG. 2 is a flow chart of another method for transmission of AI models provided in an embodiment of the disclosure;

Fig. 3 is a flowchart of another AI-model transmission method according to an embodiment of the disclosure;

fig. 4 is a flowchart of another AI-model transmission method according to an embodiment of the disclosure;

fig. 5 is a flowchart of another AI-model transmission method according to an embodiment of the disclosure;

fig. 6 is a flowchart of another AI-model transmission method according to an embodiment of the disclosure;

fig. 7 is a flowchart of another AI-model transmission method according to an embodiment of the disclosure;

fig. 8 is a flowchart of another AI-model transmission method according to an embodiment of the disclosure;

Fig. 9 is a flowchart of another AI-model transmission method according to an embodiment of the disclosure;

Fig. 10 is a flowchart of another AI-model transmission method according to an embodiment of the disclosure;

Fig. 11 is a flowchart of another AI-model transmission method according to an embodiment of the disclosure;

Fig. 12 is a flowchart of another AI-model transmission method according to an embodiment of the disclosure;

fig. 13 is a flowchart of another AI-model transmission method according to an embodiment of the disclosure;

fig. 14 is a schematic structural diagram of a transmission device of an AI model according to an embodiment of the disclosure;

Fig. 15 is a schematic structural diagram of a transmission device of an AI model according to an embodiment of the disclosure;

fig. 16 is a schematic structural diagram of a transmission device of an AI model according to an embodiment of the disclosure;

fig. 17 is a schematic structural diagram of a transmission device of an AI model according to an embodiment of the disclosure.

Detailed Description

Reference will now be made in detail to exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, the same numbers in different drawings refer to the same or similar elements, unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the embodiments of the present disclosure. Rather, they are merely examples of apparatus and methods consistent with aspects of embodiments of the present disclosure as detailed in the accompanying claims.

The terminology used in the embodiments of the disclosure is for the purpose of describing particular embodiments only and is not intended to be limiting of the embodiments of the disclosure. As used in this disclosure of embodiments and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It should also be understood that the term "and/or" as used herein refers to and encompasses any or all possible combinations of one or more of the associated listed items.

It should be understood that although the terms first, second, third, etc. may be used in embodiments of the present disclosure to describe various information, these information should not be limited to these terms. These terms are only used to distinguish one type of information from another. For example, the first information may also be referred to as second information, and similarly, the second information may also be referred to as first information, without departing from the scope of embodiments of the present disclosure. The words "if" and "if" as used herein may be interpreted as "at … …" or "at … …" or "in response to a determination", depending on the context.

Reference will now be made in detail to embodiments of the present disclosure, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the exemplary embodiments throughout, and wherein the examples are illustrated in the drawings. When the following description refers to the accompanying drawings, the same numbers in different drawings refer to the same or similar elements, unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the embodiments of the present disclosure. Rather, they are merely examples of apparatus and methods consistent with aspects of embodiments of the present disclosure as detailed in the accompanying claims.

Embodiments of the present disclosure are described in detail below, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the like or similar elements throughout. The embodiments described below by referring to the drawings are exemplary and intended for the purpose of explaining the present disclosure and are not to be construed as limiting the present disclosure.

In order to better understand a control method of direct-connection sidelink discontinuous reception DRX disclosed in an embodiment of the present disclosure, a description is first given below of a communication system to which the embodiment of the present disclosure is applicable. Showing the same or similar elements. The embodiments described below by referring to the drawings are exemplary and intended for the purpose of explaining the present disclosure and are not to be construed as limiting the present disclosure.

In order to better understand a transmission method of an AI model disclosed in an embodiment of the present disclosure, a description is first given below of a communication system to which the embodiment of the present disclosure is applicable.

Referring to fig. 1, fig. 1 is a schematic diagram of a communication system according to an embodiment of the application. The communication system may include, but is not limited to, a network device and a terminal device, where the terminal device communicates with the network device, and the number and the form of the devices shown in fig. 1 are only used as examples and not to limit the embodiments of the present application, and in practical application, two or more network devices and two or more terminal devices may be included. The communication system shown in fig. 1 may comprise a network device 101, a terminal device 102.

As an example, one terminal device may serve as both a providing node of the AI model and as an AI model receiving node. As another example, a network device may act as both a providing node and an AI model receiving node for the AI model.

It should be noted that the technical solution of the embodiment of the present disclosure may be applied to various communication systems. For example: long term evolution (long term evolution, LTE) system, fifth generation (5th generation,5G) mobile communication system, 5G New Radio (NR) system, or other future new mobile communication system, etc.

The network device in the embodiments of the present disclosure is an entity for transmitting or receiving signals at the network side. For example, the network device 101 may be an evolved NodeB (eNB), a transmission and reception point (transmission reception point or TRANSMIT RECEIVE point, TRP), a next generation NodeB (gNB) in an NR system, a base station in other future mobile communication systems, or an access node in a wireless fidelity (WIRELESS FIDELITY, wiFi) system, etc. The embodiments of the present disclosure do not limit the specific technology and specific device configuration employed by the network device. The network device provided by the embodiments of the present disclosure may be composed of a Central Unit (CU) and a Distributed Unit (DU), where the CU may also be referred to as a control unit (control unit), the structure of the CU-DU may be used to split the protocol layers of the network device, such as a base station, and the functions of part of the protocol layers are placed in the CU for centralized control, and the functions of part or all of the protocol layers are distributed in the DU, so that the CU centrally controls the DU.

The terminal device in the embodiments of the present disclosure is an entity on the user side for receiving or transmitting signals, such as a mobile phone. The terminal device may also be referred to as a terminal device (terminal), a Mobile Station (MS), a mobile terminal device (MT), etc. The terminal device may be an automobile with communication function, a smart car, a mobile phone (mobile phone), a wearable device, a tablet computer (Pad), a computer with wireless transceiving function, a Virtual Reality (VR) terminal device, an augmented reality (augmented reality, AR) terminal device, a wireless terminal device in industrial control (industrial control), a wireless terminal device in unmanned-driving (self-driving), a wireless terminal device in teleoperation (remote medical surgery), a wireless terminal device in smart grid (SMART GRID), a wireless terminal device in transportation security (transportation safety), a wireless terminal device in smart city (SMART CITY), a wireless terminal device in smart home (smart home), or the like. The embodiment of the present disclosure does not limit the specific technology and the specific device configuration adopted by the terminal device.

In view of the above problems, an embodiment of the present disclosure provides a transmission method and apparatus for an AI model.

Referring to fig. 2, fig. 2 is a flowchart of a data type transmission method according to an embodiment of the disclosure. As shown in fig. 2, the method is applied to providing nodes, and may include, but is not limited to, the following steps:

step S201: the data type that digitally represents the AI model is determined.

Wherein, the artificial intelligence (ARTIFICIAL INTELLIGENCE, AI) model involves two phases: stage one: training of AI model, i.e. obtaining AI model, stage two: the AI model deployment stage, i.e. the AI model reasoning application stage, when the AI model training and reasoning are not in the same node, the AI model will be transmitted from the trained node (providing node) to the reasoning node (receiving node), and when the AI model is transmitted from one node to another, the AI model needs to be digitally processed, i.e. the AI model structure, the model parameters and the like are digitally represented.

In the embodiment of the disclosure, the method and the device are applied to a scene that an AI model is transmitted from one node to another node, different data types supported by the two sides of the AI model interaction are confirmed, and the AI model is digitally represented according to the data types, namely the AI model is converted into a corresponding AI model bit stream.

As one possible implementation, the different data types supported by the AI model interaction parties include, but are not limited to, the user terminal supporting 8-bit integer, the AI model interaction parties supporting 16-bit floating point, and so on. Different data types may result in different resource overhead to transmit the AI model. For example, the overhead of using a 16-bit floating point type float is higher than that of an 8-bit integer Int.

In addition, if the two sides of the AI model interaction support 8-bit integer, and when the structure of the AI model, parameters of the model, and the like are transmitted, and the 16-bit floating point type of the data type is used for digital representation, problems such as increasing inference complexity, increasing delay, increasing unnecessary functions, and the like may occur, even the two sides of the AI model interaction cannot process the 16-bit floating point type of the data type, that is, cannot meet the processing capability of the receiving node of the AI model, and the two sides of the AI model interaction cannot complete the transmission of the AI model.

Therefore, when the transmission of the AI model is performed, the data types supported by both sides need to be interacted with by referring to the AI model, that is, the data types digitally representing the AI model are determined.

Step S202: and converting the AI model into a corresponding AI model bit stream according to the data type.

And transmitting the converted AI model bit stream to an AI model receiving node, carrying out inverse conversion on the AI model bit stream by the receiving node according to a preset conversion rule to obtain a corresponding AI model, and using the AI model at the AI receiving node.

In summary, different data types also cause different reasoning complexity, and after determining the data type for digitally representing the AI model, the providing node converts the AI model into a corresponding AI model bit stream according to the data type, so that the data type used when the AI model is interacted with the AI model for synchronous transmission is realized, the processing capability of the AI model receiving node can be met, and the AI model can be flexibly represented according to the data type.

An embodiment of the present disclosure provides another AI model transmission method, and fig. 2 is a schematic flow chart of another AI model transmission method provided by an embodiment of the present disclosure, which may be applied to a process of transmitting AI models by two interacting AI models, where the AI model transmission method may be performed alone, may be performed together with any one embodiment of the present disclosure or a possible implementation manner in an embodiment, and may also be performed together with any one technical scheme in the related art.

As shown in fig. 3, the transmission method of the AI model may include the steps of:

Step S301: and determining the data type digitally representing the AI model according to the supporting capacity of the AI model receiving node for the data type.

The AI model involves two phases: stage one: training of AI model, i.e. obtaining AI model, stage two: and the AI model deployment stage, namely the AI model reasoning application stage, corresponds to the AI model receiving node in the reasoning application stage and is used for obtaining a reasoning result through reasoning calculation. In the AI model interaction process, the data type digitally representing the AI model is determined according to the supporting capacity of the AI model receiving node for the data type.

Illustratively, the support capability of the AI model-receiving node is integer, then it is determined that the data type represented by the AI model is also integer.

Step S302: and converting the AI model into a corresponding AI model bit stream according to the data type.

The digitalized processing of the AI model is realized according to the data types determined in step S301, i.e. the structure of the AI model, the parameters of the AI model, etc. are represented by numbers and converted into corresponding AI model bit streams.

In summary, different data types also lead to different reasoning complexity, in the technical scheme, according to the supporting capability of an AI model receiving node on the data types, the data types for digitally representing the AI model are determined, the AI model is converted into corresponding AI model bit streams according to the data types, the data types used when the AI model interaction party synchronizes the AI model transmission are realized, the processing capability of the AI model receiving node can be met, and the AI model can be flexibly represented according to the data types.

An embodiment of the present disclosure provides another AI model transmission method, and fig. 4 is a schematic flow chart of another AI model transmission method provided by an embodiment of the present disclosure, which may be applied to a process of transmitting AI models by two interacting AI models, where the AI model transmission method may be performed alone, may be performed together with any one embodiment of the present disclosure or a possible implementation manner in an embodiment, and may also be performed together with any one technical scheme in the related art.

As shown in fig. 4, the transmission method of the AI model may include the steps of:

Step S401: and receiving the supporting capability of the AI model receiving node to the data type, which is sent by the AI model receiving node.

In the embodiment of the disclosure, an AI model interaction party (AI model providing node and AI model receiving) determines the data type represented by the AI model according to the supporting capability of the AI model receiving node to the AI model providing node.

Step S402: and determining the data type digitally representing the AI model according to the supporting capacity of the AI model receiving node for the data type.

As an achievable implementation manner of the embodiment of the disclosure, according to the data type with the maximum accuracy supported by the AI model receiving node, determining the data type corresponding to the data type with the maximum supporting capacity of the AI model receiving node as the data type for digitally representing the AI model. For example, the data type of maximum accuracy supported by the AI model receiving node is a floating-point type float, the integer Int, the AI model providing node determines that the data type digitally representing the AI model is a floating-point type float, and converts the AI model into a corresponding AI model bit stream according to the floating-point type float.

As another implementation manner of the embodiments of the present disclosure, according to the supporting capability of the AI model receiving node for the data types, a data type that is supported by the AI model providing node is selected as the data type that digitally represents the AI model, and the data type that is supported by the AI model receiving node is illustratively floating point type float and integer Int. The AI model providing node selects any data type to determine as the data type for digitizing the AI model based on the floating point type float supported by the AI model receiving node, the integer type Int. For example, the AI model receiving node may support floating point type float, integer Int. At this point the AI model receiving node selects a floating point float as the data type for which the AI model is digitally represented.

Step S403: and converting the AI model into a corresponding AI model bit stream according to the data type.

In summary, different data types also cause different reasoning complexity, time delay, power consumption and the like, and in the technical scheme, the supporting capability of the AI model receiving node to the data types sent by the AI model receiving node is received, the data types digitally representing the AI model are determined according to the supporting capability of the AI model receiving node to the data types, the AI model is converted into corresponding AI model bit streams according to the data types, so that the processing capability of the AI model receiving node can be met, and the AI model can be flexibly represented according to the data types.

An embodiment of the present disclosure provides another AI model transmission method, and fig. 5 is a schematic flow chart of another AI model transmission method provided by an embodiment of the present disclosure, which may be applied to a process of transmitting AI models by two interacting AI models, where the AI model transmission method may be performed alone, may be performed together with any one embodiment of the present disclosure or a possible implementation manner in an embodiment, and may also be performed together with any one technical scheme in the related art.

As shown in fig. 5, the transmission method of the AI model may include the steps of:

Step S501: and according to the supporting capacity of the data types of the AI model receiving nodes, determining the data type corresponding to the data type with the maximum supporting capacity of the AI model receiving nodes as the data type for digitally representing the AI model.

In order to facilitate understanding, according to the data type of the maximum accuracy supported by the AI model receiving node, determining the data type corresponding to the data type of the maximum supporting capacity of the AI model receiving node as the data type for digitally representing the AI model. For example, the data type of maximum accuracy supported by the AI model receiving node is a floating-point type float, the integer Int, the AI model providing node determines that the data type digitally representing the AI model is a floating-point type float, and converts the AI model into a corresponding AI model bit stream according to the floating-point type float.

Step S502: and converting the AI model into a corresponding AI model bit stream according to the data type.

The digitizing process of the AI model is implemented according to the data type determined in step S501, i.e. the structure of the AI model, the parameters of the AI model, etc. are converted into a corresponding AI model bit stream.

And transmitting the converted AI model bit stream to an AI model receiving node, carrying out inverse conversion on the AI model bit stream by the receiving node according to a preset conversion rule to obtain a corresponding AI model, and using the AI model at the AI receiving node to finish the transmission of the AI model.

In summary, different data types also cause different reasoning complexity, time delay, power consumption and the like, in the technical scheme, according to the supporting capability of the data type of the AI model receiving node, the data type corresponding to the data type with the maximum supporting capability of the AI model receiving node is determined to be the data type for digitally representing the AI model, and the AI model is converted into a corresponding AI model bit stream according to the data type, so that the processing capability of the AI model receiving node can be met, and the AI model can be flexibly represented according to the data type.

An embodiment of the present disclosure provides another AI model transmission method, and fig. 6 is a schematic flow chart of another AI model transmission method provided by an embodiment of the present disclosure, which may be applied to a process of transmitting AI models by two interacting AI models, where the AI model transmission method may be performed alone, may be performed together with any one embodiment of the present disclosure or a possible implementation manner in an embodiment, and may also be performed together with any one technical scheme in the related art.

As shown in fig. 6, the transmission method of the AI model may include the steps of:

step S601: according to the supporting capacity of the AI model receiving node to the data types, selecting one data type supported by the AI model providing node as the data type for digitally representing the AI model.

Illustratively, the data type that the AI model receiving node reports to its support is floating point type float, integer type Int. The AI model providing node selects any data type to determine as the data type for digitizing the AI model based on the floating point type float supported by the AI model receiving node, the integer type Int. For example, the AI model receiving node may support floating point type float, integer Int. At this time, the AI model receiving node selects a floating-point float as the data type for digitizing the AI model, or the AI model receiving node selects an integer Int as the data type for digitizing the AI model.

Step S602: and converting the AI model into a corresponding AI model bit stream according to the data type.

The digitalized processing of the AI model is realized according to the data type determined in step S601, i.e. the structure of the AI model, the parameters of the AI model, etc. are converted into a corresponding AI model bit stream.

In summary, different data types also cause different reasoning complexity, time delay, power consumption and the like, in the technical scheme, according to the supporting capability of an AI model receiving node on the data types, a data type supported by an AI model providing node is selected as a data type for digitally representing the AI model, and the AI model is converted into a corresponding AI model bit stream according to the data type, so that the processing capability of the AI model receiving node can be met, and the AI model can be flexibly represented according to the data type.

Fig. 7 is a schematic flow chart of another AI model transmission method provided by the embodiment of the present disclosure, which may be applied to a process of AI model interaction and AI model transmission, where the AI model transmission method may be performed alone, may be performed in combination with any one of the embodiments of the present disclosure or a possible implementation manner in the embodiment, or may be performed in combination with any one of the technical schemes in the related art.

As shown in fig. 7, the transmission method of the AI model may include the steps of:

step S701: and determining the data type for digitally representing the AI model according to the power consumption and/or the storage capacity reported by the AI model receiving node.

When determining the data type of the AI model for digital representation, determining according to the power consumption and/or storage capacity of the AI model receiving node, and not exceeding the data type corresponding to the power consumption and/or storage capacity of the AI model receiving node. For example, when the storage capacity of the terminal is weak, a data type with lower precision can be selected for representation; or when the receiving node does not expect to process the AI task with a larger power consumption, a data type with lower accuracy may be selected for the representation at this time.

Step S702: and converting the AI model into a corresponding AI model bit stream according to the data type.

The digitalized processing of the AI model is realized according to the data type determined in step S701, i.e. the structure of the AI model, the parameters of the AI model, etc. are converted into a corresponding AI model bit stream.

In summary, different data types also cause different reasoning complexity, time delay, power consumption and the like, and in the technical scheme, according to the power consumption and/or storage capacity reported by the AI model receiving node, the data type for digitally representing the AI model is determined, and according to the data type, the AI model is converted into a corresponding AI model bit stream, so that the processing capacity of the AI model receiving node can be met, and the AI model can be flexibly represented according to the data type.

An embodiment of the present disclosure provides another AI model transmission method, and fig. 8 is a flow chart of another AI model transmission method provided by an embodiment of the present disclosure, which may be applied to a process of transmitting AI models by two interacting AI models, where the AI model transmission method may be performed alone, may be performed together with any one embodiment of the present disclosure or a possible implementation manner in an embodiment, and may also be performed together with any one technical scheme in the related art.

As shown in fig. 8, the transmission method of the AI model may include the steps of:

step S801: according to the service requirement, the data type of the digital representation of the AI model is determined.

In the embodiment of the disclosure, the data type digitally representing the AI model is determined according to the requirements of the service on the time delay. Illustratively, int may be selected with high latency requirements and float may be selected with low latency requirements.

Step S802: and converting the AI model into a corresponding AI model bit stream according to the data type.

In summary, different data types also cause different reasoning complexity, time delay, power consumption and the like, in the technical scheme, the data type for digitally representing the AI model is determined according to the service requirement, the AI model is converted into a corresponding AI model bit stream according to the data type, the processing capacity of an AI model receiving node can be met, and the AI model can be flexibly represented according to the data type.

An embodiment of the present disclosure provides another AI model transmission method, and fig. 9 is a schematic flow chart of another AI model transmission method provided by an embodiment of the present disclosure, which may be applied to a process of transmitting AI models by two interacting AI models, where the AI model transmission method may be performed alone, may be performed together with any one embodiment of the present disclosure or a possible implementation manner in an embodiment, and may also be performed together with any one technical scheme in the related art.

As shown in fig. 9, the transmission method of the AI model may include the steps of:

Step S901: and determining the data type represented by the AI model according to the resource overhead of the AI model.

In the embodiment of the disclosure, the data type of the receiving node of the AI model is determined according to the size of the resources occupied by the AI model, for example, if the AI model is expected to occupy less or smaller resource overhead, the data type digitally representing the AI model may be determined to be integer Int, the resource overhead is not limited, and the data type digitally representing the AI model may be determined to be floating-point float.

Step S902: and converting the AI model into a corresponding AI model bit stream according to the data type.

In summary, different data types also cause different reasoning complexity, time delay, power consumption and the like, in the technical scheme, according to the resource expense of the AI model, the data type represented by the AI model is determined, and the AI model is converted into a corresponding AI model bit stream according to the data type, so that the processing capacity of an AI model receiving node can be met, and the AI model can be flexibly represented according to the data type.

An embodiment of the present disclosure provides another AI model transmission method, and fig. 10 is a flow chart of another AI model transmission method provided by an embodiment of the present disclosure, which may be applied to a process of transmitting AI models by two interacting AI models, where the AI model transmission method may be performed alone, may be performed together with any one embodiment of the present disclosure or a possible implementation manner in an embodiment, and may also be performed together with any one technical scheme in the related art.

As shown in fig. 10, the transmission method of the AI model may include the steps of:

Step S1001: and converting the AI model into a corresponding AI model bit stream according to the data type.

Step S1002: and sending indication information of the data type for digitally representing the AI model to the AI model receiving node, wherein the indication information comprises the data type for digitally representing the AI model.

And transmitting indication information of the data type for digitally representing the AI model in response to the transmission of the AI model, wherein the indication information comprises the data type for digitally representing the AI model, so that a receiving node inverts the bit stream of the AI model according to the data type in the indication information to obtain a corresponding AI model, and the AI model is used at the AI receiving node.

In summary, different data types also cause different reasoning complexity, time delay, power consumption and the like, and in the technical scheme, the AI model is converted into a corresponding AI model bit stream according to the data types, indication information of the data types for digitally representing the AI model is sent to the AI model receiving node, and the indication information comprises the data types for digitally representing the AI model, so that the processing capacity of the AI model receiving node can be met, and the AI model can be flexibly represented according to the data types.

An embodiment of the present disclosure provides another AI model transmission method, and fig. 11 is a flowchart of another AI model transmission method provided in an embodiment of the present disclosure, where the method is applied to a receiving node side. As shown in fig. 11, the transmission method of the AI model may include the steps of:

step S1101: and in response to receiving the AI model bit stream sent by the AI model providing node, carrying out inverse transformation on the AI model bit stream according to a preset transformation rule to obtain a corresponding AI model, and using the AI model at the AI receiving node.

Determining a data type for digitally representing the AI model at an AI model providing node, converting the AI model into a corresponding AI model bit stream according to the data type, and transmitting the AI model bit stream to an AI model receiving node.

As an example of an embodiment of the present disclosure, when a providing node transmits an AI model bit stream, indication information of a data type digitally representing an AI model is transmitted to the AI model receiving node, where the indication information includes the data type digitally representing the AI model. And the receiving node carries out inverse transformation on the AI model bit stream according to a preset transformation rule to obtain a corresponding AI model.

As another example of the embodiment of the present disclosure, after the receiving node receives the AI model bit stream, according to the maximum support capability data type of the AI model receiving node, performing inverse transformation on the AI model bit stream according to a predetermined transformation rule to obtain a corresponding AI model.

In the embodiment of the disclosure, the predetermined conversion rule used for performing the inverse conversion on the bit stream of the AI model to obtain the corresponding AI model is not limited.

In summary, different data types also cause different reasoning complexity, time delay, power consumption and the like, in the technical scheme, a providing node determines a data type for digitally representing an AI model, converts the AI model into a corresponding AI model bit stream according to the data type, a receiving node reversely converts the AI model bit stream according to a preset conversion rule to obtain the corresponding AI model when receiving the AI model bit stream sent by the AI model providing node, and the AI model is used at the AI receiving node, so that the processing capacity of the AI model receiving node can be met, and the AI model can be flexibly represented according to the data type.

An embodiment of the present disclosure provides another AI model transmission method, and fig. 12 is a flow chart of another AI model transmission method provided by an embodiment of the present disclosure, which may be applied to a scenario in which a receiving node receives an AI model bit stream sent by a providing node, where the AI model transmission method may be performed alone, may be performed together with any one of the embodiments of the present disclosure or a possible implementation manner of the embodiments, and may also be performed together with any one of the technical schemes of the related art.

As shown in fig. 12, the transmission method of the AI model may include the steps of:

Step S1201: and receiving indication information of the data type which is sent by the AI model providing node and digitally represents the AI model, wherein the indication information comprises the data type which digitally represents the AI model.

As an example of an embodiment of the present disclosure, when the providing node transmits the AI model bit stream, indication information of a data type digitally representing the AI model is transmitted to the AI model receiving node, where the indication information includes the data type digitally representing the AI model.

Step S1202: and responding to the indication information, and according to the data type which is included in the indication information and used for digitally representing the AI model, carrying out inverse conversion on the AI model bit stream according to a preset conversion rule to obtain a corresponding AI model.

In summary, different data types also cause different reasoning complexity, time delay, power consumption and the like, in the technical scheme, a providing node determines a data type for digitally representing an AI model, the AI model is converted into a corresponding AI model bit stream according to the data type, a receiving node receives indication information of the data type for digitally representing the AI model sent by the AI model providing node, the indication information comprises the data type for digitally representing the AI model, and in response to the indication information, the AI model bit stream is reversely converted according to a preset conversion rule to obtain the corresponding AI model according to the indication information, so that the processing capacity of the AI model receiving node can be met, and the AI model can be flexibly represented according to the data type.

An embodiment of the present disclosure provides another AI model transmission method, and fig. 13 is a flow chart of another AI model transmission method provided by an embodiment of the present disclosure, which may be applied to a scenario in which a receiving node receives an AI model bit stream sent by a providing node, where the AI model transmission method may be performed alone, may also be performed together with any one of the embodiments of the present disclosure or a possible implementation manner of the embodiments, and may also be performed together with any one of the technical schemes of the related art.

As shown in fig. 13, the transmission method of the AI model may include the steps of:

Step S1301: and providing the support capability of reporting the data type to the AI model by the node.

Step S1302: providing node reporting to the AI model is based on power consumption and/or storage capabilities.

For the supporting capability of the AI-model receiving node to transmit the data types to the providing node, based on the power consumption and/or the storage capability, refer to the above-mentioned detailed description of related contents, and the embodiments of the disclosure will not be described in detail herein.

Corresponding to the above-mentioned transmission of the AI model provided by the embodiment of fig. 2 to 10, the present disclosure also provides a transmission device of the AI model, and since the transmission device of the AI model provided by the embodiment of the present disclosure corresponds to the transmission method of the AI model provided by the embodiment of fig. 2 to 13, the implementation of the transmission method of the AI model is also applicable to the transmission device of the AI model provided by the embodiment of the present disclosure, which is not described in detail in the embodiment of the present disclosure.

Fig. 14 is a schematic structural diagram of a transmission device of an AI model according to an embodiment of the disclosure. The apparatus is provided at a providing node, the apparatus comprising:

a determining unit 1401 for determining a data type digitally representing the AI model;

A conversion unit 1402, configured to convert the AI model into a corresponding AI model bit stream according to the data type.

As a possible implementation manner of the embodiments of the present disclosure, the determining unit 1401 is further configured to determine, according to the support capability of the AI model receiving node for the data type, the data type digitally representing the AI model.

As one possible implementation of the embodiments of the present disclosure, further includes:

A receiving unit 1403, configured to receive a supporting capability of the AI model receiving node for a data type, where the supporting capability is sent by the AI model receiving node.

As a possible implementation manner of the embodiments of the present disclosure, the determining unit 1401 is further configured to determine, according to the support capability of the data type of the AI model receiving node, a data type corresponding to the data type of the maximum support capability of the AI model receiving node, as a data type that digitally represents the AI model.

As a possible implementation manner of the embodiment of the present disclosure, the determining unit 1401 is further configured to select, according to the capability of the AI model receiving node to support data types, a data type supported by the AI model providing node as a data type digitally representing the AI model.

As a possible implementation manner of the embodiments of the present disclosure, the determining unit 1401 is further configured to determine a data type digitally representing the AI model according to power consumption and/or storage capability reported by the AI model receiving node.

As a possible implementation manner of the embodiments of the present disclosure, the determining unit 1401 is further configured to determine a data type that digitally represents the AI model according to a service requirement.

As a possible implementation manner of the embodiments of the present disclosure, the determining unit 1401 is further configured to determine, according to resource overhead of the AI model, a data type represented by the AI model.

A sending unit 1404, configured to send, to the AI-model receiving node, indication information of a data type that digitally represents the AI model, where the indication information includes the data type that digitally represents the AI model.

Corresponding to the transmission of the AI model provided by the embodiments of fig. 11 to 13, the present disclosure also provides a transmission device of the AI model, and since the transmission device of the AI model provided by the embodiments of the present disclosure corresponds to the transmission method of the AI model provided by the embodiments of fig. 11 to 13, the implementation of the transmission method of the AI model is also applicable to the transmission device of the AI model provided by the embodiments of the present disclosure, which is not described in detail in the embodiments of the present disclosure.

Fig. 15 is a schematic structural diagram of a transmission device of an AI model according to an embodiment of the disclosure. The apparatus is provided at a receiving node, the apparatus comprising:

A conversion unit 1501, configured to reverse-convert the AI model bit stream according to a predetermined conversion rule to obtain a corresponding AI model in response to receiving the AI model bit stream sent by the AI model providing node;

a usage unit 1502 is configured to use the AI model at the AI receiving node.

The receiving unit 1503 is configured to receive indication information of a data type digitally representing the AI model sent by the AI model providing node, where the indication information includes the data type digitally representing the AI model.

As a possible implementation manner of the embodiment of the present disclosure, the transforming unit 1501 is further configured to, in response to the indication information, reverse-transform the bit stream of the AI model according to a predetermined transformation rule to obtain a corresponding AI model according to a data type that includes a digitalized representation of the AI model in the indication information.

As one possible implementation of the embodiments of the present disclosure, the apparatus further includes:

a first reporting unit 1504, configured to provide the AI model with a supporting capability of reporting the data type by a node.

a second reporting unit 1505, configured to provide the AI model with node reporting based on power consumption and/or storage capability.

In order to implement the above-mentioned embodiments, the present disclosure also proposes a transmission apparatus of an AI model, the apparatus including a processor and a memory, the memory storing a computer program, the processor executing the computer program stored in the memory to cause the apparatus to perform the method as described in fig. 2 to 13.

In order to implement the above embodiment, the present disclosure further proposes another AI-model transmission apparatus, including: a processor and interface circuit;

The processor is configured to execute the code instructions to perform the methods described in fig. 2-13.

To implement the above-described embodiments, the present disclosure proposes a computer-readable storage medium storing instructions that, when executed, cause the method as described in fig. 2 to 13 to be implemented.

In the embodiments provided by the present application, the method provided by the embodiments of the present disclosure is described from the perspective of the AI model interacting party. In order to implement the functions in the method provided by the embodiments of the present disclosure, the AI model interaction party may include a hardware structure, a software module, and implement the functions in the form of a hardware structure, a software module, or a hardware structure plus a software module. Some of the functions described above may be implemented in a hardware structure, a software module, or a combination of a hardware structure and a software module.

Referring to fig. 16, fig. 16 is a schematic structural diagram of a transmission device of an AI model according to an embodiment of the disclosure. Referring to fig. 16, network device 1600 includes a processing component 1622 that further includes at least one processor, and memory resources represented by memory 1632 for storing instructions, such as applications, executable by processing component 1622. The application programs stored in memory 1632 may include one or more modules that each correspond to a set of instructions. Further, the processing component 1622 is configured to execute instructions to perform any of the methods described above as applied to the network device, e.g., as described in the embodiments of fig. 2-12.

The network device 1600 may also include a power component 1626 configured to perform power management of the network device 1600, a wired or wireless network interface 1650 configured to connect the network device 1600 to a network, and an input output (I/O) interface 1658. The network device 1600 may operate based on an operating system stored in memory 1632, such as Windows Server TM, mac OS XTM, unixTM, linuxTM, freeBSDTM, or the like.

Fig. 17 is a block diagram of a transmission device of an AI model according to an embodiment of the disclosure. For example, terminal device 1700 may be a mobile phone, computer, digital broadcast terminal device, messaging device, game console, tablet device, medical device, exercise device, personal digital assistant, or the like.

Referring to fig. 17, a terminal device 1700 may include at least one of the following components: a processing component 1702, a memory 1704, a power source component 1706, a multimedia component 1708, an audio component 1710, an input/output (I/O) interface 1712, a sensor component 1714, and a communications component 1716.

The processing component 1702 generally controls overall operation of the terminal device 1700, such as operations associated with display, telephone calls, data communications, camera operations, and recording operations. The processing component 1702 may include at least one processor 1720 to execute instructions to perform all or part of the steps of the methods described above. Further, the processing component 1702 can include at least one module that facilitates interaction between the processing component 1702 and other components. For example, the processing component 1702 may include a multimedia module to facilitate interaction between the multimedia component 1708 and the processing component 1702.

The memory 1704 is configured to store various types of data to support operation at the terminal device 1700. Examples of such data include instructions for any application or method operating on terminal device 1700, contact data, phonebook data, messages, pictures, video, and the like. The memory 1704 may be implemented by any type of volatile or non-volatile memory device or combination thereof, such as Static Random Access Memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk, or optical disk.

The power supply component 1706 provides power to the various components of the terminal device 1700. The power supply component 1706 can include a power management system, at least one power supply, and other components associated with generating, managing, and distributing power for the terminal device 1700.

The multimedia component 1708 includes a screen between the terminal device 1700 and the user that provides an output interface. In some embodiments, the screen may include a Liquid Crystal Display (LCD) and a Touch Panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from a user. The touch panel includes at least one touch sensor to sense touch, swipe, and gestures on the touch panel. The touch sensor may sense not only the boundary of a touch or slide action, but also a wake-up time and pressure associated with the touch or slide operation. In some embodiments, the multimedia component 1708 includes a front-facing camera and/or a rear-facing camera. When the terminal device 1700 is in an operation mode, such as a photographing mode or a video mode, the front camera and/or the rear camera may receive external multimedia data. Each front camera and rear camera may be a fixed optical lens system or have focal length and optical zoom capabilities.

The audio component 1710 is configured to output and/or input audio signals. For example, audio component 1710 includes a Microphone (MIC) configured to receive external audio signals when terminal device 1700 is in an operational mode, such as a call mode, a recording mode, and a voice recognition mode. The received audio signals may be further stored in the memory 1704 or transmitted via the communication component 1716. In some embodiments, audio component 1710 also includes a speaker for outputting audio signals.

The I/O interface 1712 provides an interface between the processing component 1702 and peripheral interface modules, which may be keyboards, click wheels, buttons, etc. These buttons may include, but are not limited to: homepage button, volume button, start button, and lock button.

The sensor assembly 1714 includes at least one sensor for providing the terminal device 1700 with a status assessment of various aspects. For example, sensor assembly 1714 may detect the on/off state of terminal device 1700, the relative positioning of the components, such as the display and keypad of terminal device 1700, sensor assembly 1714 may also detect the change in position of terminal device 1700 or a component of terminal device 1700, the presence or absence of user contact with terminal device 1700, the orientation or acceleration/deceleration of terminal device 1700, and the change in temperature of terminal device 1700. The sensor assembly 1714 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact. The sensor assembly 1714 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor assembly 1714 may also include an acceleration sensor, a gyroscopic sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

The communication component 1716 is configured to facilitate communication between the terminal device 1700 and other devices, either wired or wireless. Terminal device 1700 may access a wireless network based on a communication standard, such as WiFi,2G, or 3G, or a combination thereof. In one exemplary embodiment, the communication component 1716 receives broadcast signals or broadcast related information from an external broadcast management system via a broadcast channel. In one exemplary embodiment, the communication component 1716 further includes a Near Field Communication (NFC) module to facilitate short range communications. For example, the NFC module may be implemented based on Radio Frequency Identification (RFID) technology, infrared data association (IrDA) technology, ultra Wideband (UWB) technology, bluetooth (BT) technology, and other technologies.

In an exemplary embodiment, the terminal device 1700 may be implemented by at least one Application Specific Integrated Circuit (ASIC), a Digital Signal Processor (DSP), a Digital Signal Processing Device (DSPD), a Programmable Logic Device (PLD), a Field Programmable Gate Array (FPGA), a controller, a microcontroller, a microprocessor, or other electronic components for performing the methods shown in fig. 1-11 described above.

In an exemplary embodiment, a non-transitory computer readable storage medium is also provided, such as a memory 1704, including instructions executable by the processor 1720 of the terminal device 1700 to perform the methods shown in fig. 2-13 described above. For example, the non-transitory computer readable storage medium may be ROM, random Access Memory (RAM), CD-ROM, magnetic tape, floppy disk, optical data storage device, etc.

Those of skill in the art will further appreciate that the various illustrative logical blocks (illustrative logical block) and steps (steps) described in connection with the embodiments of the disclosure may be implemented by electronic hardware, computer software, or combinations of both. Whether such functionality is implemented as hardware or software depends upon the particular application and design requirements of the overall system. Those skilled in the art may implement the described functionality in varying ways for each particular application, but such implementation is not to be understood as beyond the scope of the embodiments of the present disclosure.

In the above embodiments, it may be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product. The computer program product comprises one or more computer programs. When the computer program is loaded and executed on a computer, the processes or functions described in accordance with the embodiments of the present disclosure are produced in whole or in part. The computer may be a general purpose computer, a special purpose computer, a computer network, or other programmable apparatus. The computer program may be stored in or transmitted from one computer readable storage medium to another, for example, by wired (e.g., coaxial cable, optical fiber, digital subscriber line (digital subscriber line, DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means from one website, computer, server, or data center. The computer readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. that contains an integration of one or more available media. The usable medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a magnetic tape), an optical medium (e.g., a high-density digital video disc (digital video disc, DVD)), or a semiconductor medium (e.g., a solid-state disk (solid-state drive STATE DISK, SSD)), or the like.

Those of ordinary skill in the art will appreciate that: the first, second, etc. numbers referred to in the present application are merely for convenience of description and are not intended to limit the scope of the embodiments of the present disclosure, but also to indicate the order of the steps.

At least one of the present application may also be described as one or more, and a plurality may be two, three, four or more, and the present application is not limited thereto. In the embodiment of the disclosure, for a technical feature, the technical features in the technical feature are distinguished by "first", "second", "third", "a", "B", "C", and "D", and the technical features described by "first", "second", "third", "a", "B", "C", and "D" are not in sequence or in order of magnitude.

The correspondence relation shown in each table in the application can be configured or predefined. The values of the information in each table are merely examples, and may be configured as other values, and the present application is not limited thereto. In the case of the correspondence between the configuration information and each parameter, it is not necessarily required to configure all the correspondence shown in each table. For example, in the table of the present application, the correspondence relation shown by some rows may not be configured. For another example, appropriate morphing adjustments, e.g., splitting, merging, etc., may be made based on the tables described above. The names of the parameters indicated in the tables may be other names which are understood by the transmission device of the AI model, and the values or expression of the parameters may be other values or expression which are understood by the transmission device of the AI model. When the tables are implemented, other data structures may be used, for example, an array, a queue, a container, a stack, a linear table, a pointer, a linked list, a tree, a graph, a structure, a class, a heap, a hash table, or a hash table.

Predefined in the present application may be understood as defining, predefining, storing, pre-negotiating, pre-configuring, curing, or pre-sintering.

Those of ordinary skill in the art will appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.

It will be clear to those skilled in the art that, for convenience and brevity of description, specific working procedures of the above-described systems, apparatuses and units may refer to corresponding procedures in the foregoing method embodiments, and are not repeated herein.

The foregoing is merely illustrative of the present application, and the present application is not limited thereto, and any person skilled in the art will readily recognize that variations or substitutions are within the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

A transmission method of an AI model, the method being performed by a providing node, the method comprising:

Determining a data type for digitally representing the AI model;

And converting the AI model into a corresponding AI model bit stream according to the data type.
The transmission method of claim 1, wherein the determining the data type for digitally representing the AI model comprises:

and determining the data type digitally representing the AI model according to the supporting capacity of the AI model receiving node for the data type.
The transmission method according to claim 2, further comprising:

And receiving the supporting capability of the AI model receiving node to the data type, which is sent by the AI model receiving node.
The transmission method according to claim 2, wherein determining the data type digitally representing the AI model based on the AI model receiving node's supporting capability for the data type comprises:

And according to the supporting capacity of the data types of the AI model receiving nodes, determining the data type corresponding to the data type with the maximum supporting capacity of the AI model receiving nodes as the data type for digitally representing the AI model.
The transmission method according to claim 2, wherein determining the data type digitally representing the AI model based on the AI model receiving node's supporting capability for the data type comprises:

According to the supporting capacity of the AI model receiving node to the data types, selecting one data type supported by the AI model providing node as the data type for digitally representing the AI model.
The transmission method according to claim 1, wherein the determining the data type of the AI model-receiving node comprises:

And determining the data type for digitally representing the AI model according to the power consumption and/or the storage capacity reported by the AI model receiving node.
The transmission method according to claim 1, wherein the determining the data type of the AI model-receiving node comprises:

According to the service requirement, the data type of the digital representation of the AI model is determined.
The transmission method according to claim 1, wherein the determining the data type of the AI model-receiving node comprises:

And determining the data type represented by the AI model according to the resource overhead of the AI model.
The transmission method according to any one of claims 1 to 7, characterized by further comprising:

And sending indication information of the data type for digitally representing the AI model to the AI model receiving node, wherein the indication information comprises the data type for digitally representing the AI model.
A transmission method of an AI model, the method being performed by a receiving node, the method comprising:

And in response to receiving the AI model bit stream sent by the AI model providing node, carrying out inverse transformation on the AI model bit stream according to a preset transformation rule to obtain a corresponding AI model, and using the AI model at the AI receiving node.
The transmission method according to claim 10, further comprising:

And receiving indication information of the data type which is sent by the AI model providing node and digitally represents the AI model, wherein the indication information comprises the data type which digitally represents the AI model.
The transmission method according to claim 11, wherein the inverse transforming the AI model bit stream according to the predetermined transformation rule to obtain the corresponding AI model includes:

And responding to the indication information, and according to the data type which is included in the indication information and used for digitally representing the AI model, carrying out inverse conversion on the AI model bit stream according to a preset conversion rule to obtain a corresponding AI model.
The transmission method according to claim 10, characterized in that the method further comprises:

And providing the support capability of reporting the data type to the AI model by the node.
The transmission method according to claim 10, characterized in that the method further comprises:

Providing node reporting to the AI model is based on power consumption and/or storage capabilities.
An AI-model transmission apparatus provided at a providing node, the apparatus comprising:

A determining unit for determining a data type for digitally representing the AI model;

And the conversion unit is used for converting the AI model into a corresponding AI model bit stream according to the data type.
A transmission apparatus of an AI model, the apparatus being provided at a receiving node, the apparatus comprising:

The conversion unit is used for reversely converting the AI model bit stream according to a preset conversion rule to obtain a corresponding AI model when the AI model bit stream sent by the AI model providing node is received;

A use unit for using the AI model at the AI receiving node.
An AI-model transmission apparatus, characterized in that the apparatus comprises a processor and a memory, the memory having stored therein a computer program, the processor executing the computer program stored in the memory to cause the apparatus to perform the method of any one of claims 1-9 or the method of any one of claims 10-14.
A transmission device of an AI model, characterized by comprising: a processor and interface circuit;

the interface circuit is used for receiving code instructions and transmitting the code instructions to the processor;

The processor being operative to execute the code instructions to perform the method of any one of claims 1 to 9 or the method of any one of claims 10 to 14.
A computer readable storage medium storing instructions which, when executed, cause a method as claimed in any one of claims 1 to 9 or a method as claimed in any one of claims 10 to 14 to be implemented.