WO2024026889A1 - 一种数据类型确定方法/装置/设备及存储介质 - Google Patents
一种数据类型确定方法/装置/设备及存储介质 Download PDFInfo
- Publication number
- WO2024026889A1 WO2024026889A1 PCT/CN2022/110708 CN2022110708W WO2024026889A1 WO 2024026889 A1 WO2024026889 A1 WO 2024026889A1 CN 2022110708 W CN2022110708 W CN 2022110708W WO 2024026889 A1 WO2024026889 A1 WO 2024026889A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- data type
- devices
- model
- model task
- data
- Prior art date
Links
- 238000000034 method Methods 0.000 title claims abstract description 171
- 238000004891 communication Methods 0.000 claims abstract description 66
- 238000004590 computer program Methods 0.000 claims description 29
- 230000015654 memory Effects 0.000 claims description 18
- 238000012549 training Methods 0.000 claims description 12
- 230000002596 correlated effect Effects 0.000 claims description 5
- 238000013473 artificial intelligence Methods 0.000 description 43
- 238000007667 floating Methods 0.000 description 37
- 230000006870 function Effects 0.000 description 18
- 230000000875 corresponding effect Effects 0.000 description 11
- 238000010586 diagram Methods 0.000 description 10
- 230000004044 response Effects 0.000 description 10
- 238000005516 engineering process Methods 0.000 description 9
- 230000008569 process Effects 0.000 description 5
- 239000004065 semiconductor Substances 0.000 description 5
- 238000013461 design Methods 0.000 description 3
- 229910044991 metal oxide Inorganic materials 0.000 description 3
- 150000004706 metal oxides Chemical class 0.000 description 3
- 238000010295 mobile communication Methods 0.000 description 3
- 229910000577 Silicon-germanium Inorganic materials 0.000 description 2
- 230000003190 augmentative effect Effects 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 230000014509 gene expression Effects 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- JBRZTFJDHDCESZ-UHFFFAOYSA-N AsGa Chemical compound [As]#[Ga] JBRZTFJDHDCESZ-UHFFFAOYSA-N 0.000 description 1
- LEVVHYCKPQWKOP-UHFFFAOYSA-N [Si].[Ge] Chemical compound [Si].[Ge] LEVVHYCKPQWKOP-UHFFFAOYSA-N 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 229910052751 metal Inorganic materials 0.000 description 1
- 239000002184 metal Substances 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 230000008054 signal transmission Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000007711 solidification Methods 0.000 description 1
- 230000008023 solidification Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000001356 surgical procedure Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L27/00—Modulated-carrier systems
Definitions
- the present disclosure relates to the field of communication technology, and in particular, to a data type determination method/device/equipment and a storage medium.
- AI Artificial Intelligent, artificial intelligence
- network nodes usually need to perform model tasks (such as AI model training tasks and/or AI model deployment inference tasks) before applying the AI model.
- model tasks such as AI model training tasks and/or AI model deployment inference tasks
- multiple different network nodes may be required to jointly participate in completing model tasks.
- different network nodes may support different data types. Based on this, when multiple network nodes jointly participate in model tasks, different network nodes may use different types of data types, which may reduce the accuracy of multiple network nodes when performing model tasks (for example, due to the need for conversion).
- the data types between different network nodes lead to reduced training accuracy or reduced inference accuracy of the AI model), which in turn leads to lower accuracy of the final AI model and reduces the user experience of the AI model.
- the data type determination method/device/equipment and storage medium proposed in this disclosure are used to solve the technical problems in the methods of related technologies that lead to low accuracy of AI models.
- embodiments of the present disclosure provide a data type determination method in which multiple devices jointly participate in executing model tasks.
- the method is executed by a terminal device among the multiple devices, including:
- the data types to be used when executing the model task determined by multiple devices that jointly participate in executing the model task are all the same, so that the multiple devices can execute the model task based on the same data type. , then there is no need to type-convert the usage data of multiple devices during the execution of the model task, thereby ensuring the execution accuracy of the model task, thereby ensuring the accuracy of the AI model obtained based on the model task, and improving the user experience of the AI model.
- embodiments of the present disclosure provide a data type determination method, which is executed by a network device that jointly participates in executing model tasks with other devices, including:
- an embodiment of the present disclosure provides a communication device, which is configured in a terminal device and includes:
- a determining module configured to determine the data type to be used when performing the model task, wherein the data type determined by the terminal device is the same as the data type determined by other devices in the plurality of devices.
- an embodiment of the present disclosure provides a communication device, which is configured in a network device and includes:
- Determining module configured to determine the data type to be used by each device that performs the model task when performing the model task, wherein the data types used by the network device when performing the model task determined by the network device are the same.
- an embodiment of the present disclosure provides a communication device.
- the communication device includes a processor.
- the processor calls a computer program in a memory, it executes the method described in the first aspect.
- an embodiment of the present disclosure provides a communication device.
- the communication device includes a processor.
- the processor calls a computer program in a memory, it executes the method described in the second aspect.
- an embodiment of the present disclosure provides a communication device.
- the communication device includes a processor and a memory, and a computer program is stored in the memory; the processor executes the computer program stored in the memory, so that the communication device executes The method described in the first aspect above.
- an embodiment of the present disclosure provides a communication device.
- the communication device includes a processor and a memory, and a computer program is stored in the memory; the processor executes the computer program stored in the memory, so that the communication device executes The method described in the second aspect above.
- an embodiment of the present disclosure provides a communication device.
- the device includes a processor and an interface circuit.
- the interface circuit is used to receive code instructions and transmit them to the processor.
- the processor is used to run the code instructions to cause the The device performs the method described in the first aspect.
- an embodiment of the present disclosure provides a communication device.
- the device includes a processor and an interface circuit.
- the interface circuit is used to receive code instructions and transmit them to the processor.
- the processor is used to run the code instructions to cause the The device performs the method described in the second aspect above.
- an embodiment of the present disclosure provides a communication system, which includes the communication device described in the third aspect to the communication device described in the fourth aspect, or the system includes the communication device described in the fifth aspect to The communication device according to the sixth aspect, or the system includes the communication device according to the seventh aspect to the communication device according to the eighth aspect, or the system includes the communication device according to the ninth aspect to the tenth aspect. the above-mentioned communication device.
- embodiments of the present invention provide a computer-readable storage medium for storing instructions used by the above-mentioned network device.
- the terminal device is caused to execute the above-mentioned first to third aspects. The method described in any of the aspects.
- the present disclosure also provides a computer program product including a computer program, which, when run on a computer, causes the computer to execute the method described in any one of the above first to second aspects.
- the present disclosure provides a chip system that includes at least one processor and an interface for supporting a network device to implement the functions involved in the method described in any one of the first to second aspects, For example, at least one of the data and information involved in the above method is determined or processed.
- the chip system further includes a memory, and the memory is used to store necessary computer programs and data for the source secondary node.
- the chip system may be composed of chips, or may include chips and other discrete devices.
- the present disclosure provides a computer program that, when run on a computer, causes the computer to perform the method described in any one of the above first to second aspects.
- Figure 1 is a schematic architectural diagram of a communication system provided by an embodiment of the present disclosure
- Figure 2 is a schematic flowchart of a data type determination method provided by another embodiment of the present disclosure.
- Figure 3 is a schematic flowchart of a data type determination method provided by yet another embodiment of the present disclosure.
- Figure 4 is a schematic flowchart of a data type determination method provided by yet another embodiment of the present disclosure.
- Figure 5 is a schematic flowchart of a data type determination method provided by another embodiment of the present disclosure.
- Figure 6a is a schematic flowchart of a data type determination method provided by yet another embodiment of the present disclosure.
- Figure 6b is a schematic flowchart of a data type determination method provided by yet another embodiment of the present disclosure.
- Figure 7 is a schematic flowchart of a data type determination method provided by yet another embodiment of the present disclosure.
- Figure 8 is a schematic flowchart of a data type determination method provided by an embodiment of the present disclosure.
- Figure 9 is a schematic flowchart of a data type determination method provided by another embodiment of the present disclosure.
- Figure 10 is a schematic flowchart of a data type determination method provided by yet another embodiment of the present disclosure.
- Figure 11a is a schematic flowchart of a data type determination method provided by yet another embodiment of the present disclosure.
- Figure 11b is a schematic flowchart of a data type determination method provided by yet another embodiment of the present disclosure.
- Figure 12 is a schematic structural diagram of a communication device provided by an embodiment of the present disclosure.
- Figure 13 is a schematic structural diagram of a communication device provided by another embodiment of the present disclosure.
- Figure 14 is a block diagram of a user equipment provided by an embodiment of the present disclosure.
- Figure 15 is a block diagram of a network side device provided by an embodiment of the present disclosure.
- first, second, third, etc. may be used to describe various information in the embodiments of the present disclosure, the information should not be limited to these terms. These terms are only used to distinguish information of the same type from each other.
- first information may also be called second information, and similarly, the second information may also be called first information.
- the words "if” and “if” as used herein may be interpreted as “when” or “when” or “in response to determining.”
- AI is a new technical science that studies and develops theories, methods, technologies and application systems for simulating, extending and expanding human intelligence.
- the various network elements/functions involved in the embodiments of the present disclosure can be either an independent hardware device or a function implemented by computer code within the hardware device. This is not the case in the embodiments of the present disclosure. limited.
- FIG. 1 is a schematic architectural diagram of a communication system provided by an embodiment of the present disclosure.
- the communication system may include but is not limited to one network device and one terminal device.
- the number and form of devices shown in Figure 1 are only for examples and do not constitute a limitation on the embodiments of the present disclosure. In actual applications, two or more devices may be included. Network equipment, two or more terminal devices.
- the communication system shown in Figure 1 includes a network device 11 and a terminal device 12 as an example.
- LTE long term evolution
- 5th generation fifth generation
- 5G new radio (NR) system 5th generation new radio
- the network device 11 in the embodiment of the present disclosure is an entity on the network side that is used to transmit or receive signals.
- the network device 11 may be an evolved base station (evolved NodeB, eNB), a transmission reception point (TRP), a next generation base station (next generation NodeB, gNB) in an NR system, or other base stations in future mobile communication systems. Base stations or access nodes in wireless fidelity (WiFi) systems, etc.
- the embodiments of the present disclosure do not limit the specific technologies and specific equipment forms used by network equipment.
- the network equipment provided by the embodiments of the present disclosure may be composed of a centralized unit (CU) and a distributed unit (DU).
- the CU may also be called a control unit (control unit).
- CU-DU is used.
- the structure can separate the protocol layers of network equipment, such as base stations, and place some protocol layer functions under centralized control on the CU. The remaining part or all protocol layer functions are distributed in the DU, and the CU centrally controls the
- the terminal device 12 in the embodiment of the present disclosure is an entity on the user side for receiving or transmitting signals, such as a mobile phone.
- Terminal equipment can also be called terminal equipment (terminal), user equipment (user equipment, UE), mobile station (mobile station, MS), mobile terminal equipment (mobile terminal, MT), etc.
- the terminal device can be a car with communication functions, a smart car, a mobile phone, a wearable device, a tablet computer (Pad), a computer with wireless transceiver functions, a virtual reality (VR) terminal device, an augmented reality (augmented reality (AR) terminal equipment, wireless terminal equipment in industrial control, wireless terminal equipment in self-driving, wireless terminal equipment in remote medical surgery, smart grid ( Wireless terminal equipment in smart grid, wireless terminal equipment in transportation safety, wireless terminal equipment in smart city, wireless terminal equipment in smart home, etc.
- the embodiments of the present disclosure do not limit the specific technology and specific equipment form used by the terminal equipment.
- Figure 2 is a schematic flowchart of a data type determination method provided by an embodiment of the present disclosure.
- multiple devices jointly participate in executing model tasks, and the method is executed by a terminal device among the multiple devices, as shown in Figure 2
- the data type determination method may include the following steps:
- Step 201 Determine the data type to be used when executing the model task.
- the above-mentioned model tasks may include model training tasks and/or model inference tasks.
- the above-mentioned "multiple devices jointly participate in performing model tasks” can be understood as: multiple devices jointly perform a model task (such as a model training task or a model inference task), where different devices are used to be responsible for the model training task. At least part of the model tasks are implemented through interaction between various devices.
- the above-mentioned data type may include at least one of an integer (int), a single floating point number (ie, a 16-bit floating point number (float)), and a double floating point number (ie, a 32-bit floating point number (float)).
- the above-mentioned multiple devices participating in executing the model task may or may not include network devices.
- the terminal device's method of determining the data type to be used when performing model tasks will also be different.
- each terminal device when multiple devices include a network device, each terminal device will uniformly report capability information related to the data types it supports to the network device (i.e., the first mentioned later). data support capability information or second data support capability information), so that the network device determines the data type for performing the model task based on the received capability information, and sends the instruction information of the determined data type to each terminal device respectively, so that each terminal Based on the instructions, the device can determine the type of data to use when performing model tasks.
- capability information related to the data types it supports to the network device (i.e., the first mentioned later).
- data support capability information or second data support capability information so that the network device determines the data type for performing the model task based on the received capability information, and sends the instruction information of the determined data type to each terminal device respectively, so that each terminal Based on the instructions, the device can determine the type of data to use when performing model tasks.
- the various devices when the multiple devices do not include network devices (that is, the multiple devices are all terminal devices), the various devices will exchange capability information related to the data types they support. , so that each device knows the data types supported by other devices, so that each device can independently determine the data type to be used when executing the model task based on the data types supported by multiple devices participating in the execution of the model task.
- the specific implementation method of the above content will be introduced in detail in subsequent embodiments.
- the data type determined by the terminal device by executing the above step 201 and the data type determined by other devices among the above multiple devices should be the same. Therefore, This ensures that the terminal device can use the same data type as other devices to perform model tasks, thereby ensuring the unification of data types when multiple devices perform model tasks and ensuring the accuracy of executing model tasks.
- the terminal device determines the data type to be used when executing the model task, where the data type determined by the terminal device is consistent with the data type among the multiple devices except the terminal device.
- the other devices determine the same data type. That is to say, in the present disclosure, the data types to be used when executing the model task determined by multiple devices that jointly participate in executing the model task are the same. Therefore, the multiple devices can execute the model task based on the same data type, then execute During the model task, there is no need to type-convert the usage data of multiple devices, thereby ensuring the execution accuracy of the model task, thereby ensuring the accuracy of the AI model obtained based on the model task, and improving the user experience of the AI model.
- Figure 3 is a schematic flowchart of a data type determination method provided by an embodiment of the present disclosure.
- the method is executed by a terminal device, and multiple devices jointly participate in executing model tasks.
- the method is executed by a terminal device among the multiple devices, such as As shown in Figure 3, the data type determination method may include the following steps:
- Step 301 Send the first data support capability information of the terminal device to other devices respectively.
- the above-mentioned first data support capability information is used to indicate the data type supported by the terminal device.
- the data types supported by the terminal device may include at least one of integers, single floating point numbers, and double floating point numbers.
- the precision of different data types will be different. However, it should be recognized that when a terminal device supports a data type with higher precision, it is also implicitly instructed that the terminal device supports a data type with lower precision than the data type. For example, it is assumed that the data type supported by the terminal device is double floating point number. Since the precision of integer type is smaller than the precision of single floating point number and smaller than the precision of double floating point number, when the data type supported by the terminal device is double floating point number, it implicitly indicates that the terminal device also supports integers and single floating point numbers.
- the above-mentioned first data support capability information may be sent by the terminal device to other devices through the D2D (Device to Device, device to device) method.
- D2D Device to Device, device to device
- Step 302 Receive second data support capability information sent by other devices.
- the second data support capability information is used to indicate data types supported by other devices.
- the data type supported by other devices may be at least one of integer, single floating point number, and double floating point number.
- the second data support capability information may be sent by other devices to the terminal device in a D2D manner.
- steps 301-302 when the multiple devices participating in executing the model task do not include a network device, the multiple devices will exchange their data support capability information with each other, thereby Make each device participating in the model task aware of the data types supported by other devices except itself, so that each device can subsequently uniformly determine the data types to be used when executing the model task based on the data types supported by other devices and the data types supported by itself. Data type (i.e. subsequent steps 303-304).
- Step 303 Determine at least one data type jointly supported by the terminal device and other devices based on the first data support capability information and the second data support capability information.
- the data type jointly indicated by the first data support capability information and the second data support capability information may be determined as a data type commonly supported by the terminal device and other devices.
- the first data support capability information of the terminal device indicates that the data types supported by the terminal device are: integer and single floating point number
- the second data support capability information of other device #1 Indicates that the data types supported by the other device #1 are integers, single floating point numbers, and double floating point numbers. Then it is determined that the data types supported by the terminal device and other device #1 are integers and single floating point numbers.
- Step 304 Determine the data type to be used when executing the model task from at least one data type jointly supported by the terminal device and other devices.
- the method of determining the data type to be used when executing the model task from at least one data type jointly supported by the terminal device and other devices may include at least one of the following:
- Method 1 Determine the first data type with the highest accuracy among at least one commonly supported data type as the data type to be used when executing the model task.
- the precision of the data type is proportional to the number of bits corresponding to the data type. That is, data types with more bits have greater precision. For example: the number of bits of an integer is 8 bits, the number of bits of a single floating point number is 32 bits, and the number of bits of a double floating point number is 64 bits, then the precision relationships corresponding to the three data types are: double float Point number > single floating point number > integer.
- the above-mentioned method of determining the first data type with the highest accuracy among at least one commonly supported data type as the data type to be used when executing the model task is:
- the first data type with the most bits among the data types is determined as the data type to be used when performing model tasks.
- the commonly supported data types are integers and single floating-point numbers, where the number of bits of a single floating-point number is greater than the number of bits of an integer, and the single floating-point number is a commonly supported data type.
- the single floating point number can be directly determined as the data type to be used when executing the model task.
- Method 2 Determine the second data type of at least one commonly supported data type that is less precise than the first data type as the data type to be used when executing the model task.
- the data type with the highest precision among the commonly supported data types may not be selected as the data type to be used when executing the model task. Instead, the data type with the precision smaller than the maximum precision may be selected as the data type for execution.
- the data type to be used in model tasks so as to ensure the execution accuracy of model tasks and reduce execution costs.
- the commonly supported data types are integers, single floating point numbers, and double floating point numbers
- the data type with the highest accuracy ie, the aforementioned first data type
- Double floating point number you can determine the data type (such as integer or single floating point number) from the commonly supported data types whose precision is smaller than the double floating point number as the data type to be used when performing model tasks.
- Method 3 Determine the data type to be used when executing the model task from at least one commonly supported data type based on the expected resource overhead of the model task.
- the expected resource cost of the model task is small, select a data type with fewer bits and lower precision; if there is no limit on the expected resource cost of the model task, then select the data type with fewer bits. A data type with high precision.
- the commonly supported data types are integers, single floating point numbers, and double floating point numbers.
- the integer type can be selected as the execution The data type to be used when executing the model task; if there is no limit on the expected resource overhead of the model task, you can select double floating point number as the data type to be used when executing the model task.
- Method 4 Determine the data type to be used when executing the model task from at least one commonly supported data type based on the business requirements of the model task.
- the business requirement may be a delay requirement of the model task.
- the above-mentioned method of determining the data type to be used when executing the model task from at least one commonly supported data type based on the business requirements of the model task may include: If the model task has higher latency requirements, select bit A data type with a small number of bits and a low precision; if the model task delay requirement is low, choose a data type with a large number of bits and high precision.
- the commonly supported data types are integers and single floating point numbers.
- the integer type is selected; If the business needs have low latency requirements, choose a single floating point number.
- each terminal device among the multiple devices independently determines the data type to be used from the commonly supported data types by executing the above steps 303-304
- each terminal device Terminal devices should all use the same method to determine the data type to use from the commonly supported data types.
- each terminal device can use the above method three to determine the data type to be used when executing the model task, or each terminal device can use the above method one to determine the data type to be used when executing the model task, thereby ensuring that each The data types determined by the terminal equipment are the same, ensuring the accuracy of executing model tasks.
- step 301-step 304 may be applied to "the multiple devices participating in executing the model task do not include network devices (ie, the multiple devices participating in executing the model task are all terminal devices). )" scene.
- the terminal device determines the data type to be used when executing the model task, where the data type determined by the terminal device is consistent with the data type among the multiple devices except the terminal device.
- the other devices determine the same data type. That is to say, in the present disclosure, the data types to be used when executing the model task determined by multiple devices that jointly participate in executing the model task are the same. Therefore, the multiple devices can execute the model task based on the same data type, then execute During the model task, there is no need to type-convert the usage data of multiple devices, thereby ensuring the execution accuracy of the model task, thereby ensuring the accuracy of the AI model obtained based on the model task, and improving the user experience of the AI model.
- Figure 4 is a schematic flowchart of a method for determining a data type provided by an embodiment of the present disclosure. Multiple devices jointly participate in executing model tasks. The method is executed by a terminal device among the multiple devices. In response to a response from a network among the multiple devices, Device, as shown in Figure 4, the data type determination method may include the following steps:
- Step 401 Send first data support capability information to the network device.
- the above-mentioned first data support capability information is used to indicate the data type supported by the terminal device.
- each terminal device in the multiple devices needs to send a unified message to the network device.
- the first data supports capability information.
- Step 402 Receive instruction information sent by the network device.
- the indication information is used to indicate the data type to be used when executing the model task.
- the data type indicated by the indication information is specifically determined by the network device based on the first data support capability information sent by each terminal device.
- the instruction information received by each terminal device is the same, thereby ensuring that the data types to be used when executing model tasks determined by multiple devices are the same and unified, thus ensuring subsequent execution of the model. accuracy during the task.
- Step 403 Determine the data type to be used when executing the model task based on the instruction information.
- the method of the embodiment corresponding to Figure 4 i.e., steps 401 to 403 can be applied to "multiple devices participating in executing the model task include network devices (i.e., multiple devices participating in executing the model task include terminals at the same time). devices and network equipment)" scenario.
- the terminal device determines the data type to be used when executing the model task, where the data type determined by the terminal device is consistent with the data type among the multiple devices except the terminal device.
- the other devices determine the same data type. That is to say, in the present disclosure, the data types to be used when executing the model task determined by multiple devices that jointly participate in executing the model task are the same. Therefore, the multiple devices can execute the model task based on the same data type, then execute During the model task, there is no need to type-convert the usage data of multiple devices, thereby ensuring the execution accuracy of the model task, thereby ensuring the accuracy of the AI model obtained based on the model task, and improving the user experience of the AI model.
- Figure 5 is a schematic flowchart of a data type determination method provided by an embodiment of the present disclosure. Multiple devices jointly participate in executing model tasks. The method is executed by a terminal device among the multiple devices. In response to a response that no one of the multiple devices includes For network equipment, as shown in Figure 5, the data type determination method may include the following steps:
- Step 501 Send the first data support capability information of the terminal device to other devices respectively.
- step 501 please refer to the description of the above embodiment.
- Step 502 Send the first capability information of the terminal device to other devices respectively.
- the first capability information may be used to indicate at least one of the following:
- the storage capacity of the terminal device is the storage capacity of the terminal device
- the power consumption capability of the terminal device is the power consumption capability of the terminal device.
- the first capability information may be sent by the terminal device to other devices in a D2D manner.
- the storage capability of the terminal device may be the current storage capability of the terminal device
- the power consumption capability of the terminal device may be the current power consumption capability of the terminal device
- Step 503 Receive second data support capability information and second capability information sent by other devices.
- the second data support capability information is used to indicate data types supported by other devices, and the second capability information may be used to indicate at least one of the following:
- the storage capacity of the other device may be the current storage capacity of the other device, and the power consumption capacity of the other device may be the current power consumption capacity of the other device.
- steps 501-503 when the multiple devices participating in executing the model task do not include network devices, the multiple devices will interact with each other in terms of their data support capabilities and storage capabilities. , at least one of power consumption capabilities, so that each device participating in the model task knows the data types supported by other devices except itself and the storage capabilities and/or power consumption capabilities of other devices, so that each device can subsequently be based on other devices.
- the data types supported by the device and the storage capabilities and/or power consumption capabilities of other devices are combined with the data types supported by the device and the storage capabilities and/or power consumption capabilities of the device to uniformly determine the data type to be used when ultimately executing the model task. (i.e. subsequent steps 504-505).
- Step 504 Determine at least one data type jointly supported by the terminal device and other devices based on the first data support capability information and the second data support capability information.
- step 504 For a detailed introduction to step 504, please refer to the description of step 303 above, and the embodiment of the present disclosure will not be repeated here.
- Step 505 Determine the data type to be used when executing the model task from at least one commonly supported data type based on the first capability information and the second capability information.
- the method of determining the data type to be used when executing the model task from at least one commonly supported data type based on the first capability information and the second capability information may include: Select a third data type from at least one data type that accurately matches the capability indicated by the first capability information and/or the second capability information, and determine the third data type as the data type to be used when executing the model task.
- the terminal device determines the third data type based on comprehensive consideration of the storage capabilities and/or power consumption capabilities of multiple devices participating in performing the model task.
- the accuracy of the third data type is matched with the comprehensive storage capabilities and/or power consumption capabilities of multiple devices. For example: when the comprehensive storage capacity and/or power consumption capacity of multiple devices is stronger, it means that multiple devices participating in executing model tasks can support data types with higher precision and more bits. At this time, the selected The accuracy of the third data type can be higher and the number of bits can be higher; when the combined storage capacity and/or power consumption capabilities of multiple devices are weaker, it means that multiple devices participating in executing model tasks can support lower accuracy bits. For a data type with fewer bits, the precision of the selected third data type may be lower and the number of bits may be smaller.
- the commonly supported data types are integers and single floating point numbers.
- the storage capabilities of the first capability information and/or the second capability information are relatively strong, you can Select double floating point number as the data type to be used when performing model tasks; if the storage capacity of the first capability information and/or the second capability information is weak, you can select integer as the data type to be used when executing model tasks.
- each terminal device among the multiple devices independently determines the data type to be used from the commonly supported data types by executing the above steps 504-505
- each terminal device All terminal devices should use the same method (or the same rules) to determine the data type to be used from the commonly supported data types.
- each terminal device should "use storage capabilities and/or power consumption capabilities that are consistent with the determined data types to be used.”
- the method (or rule) of "the storage capacity and/or power consumption capacity is negatively correlated with the accuracy of the determined data type used” is derived from the commonly supported method (or rule).
- the data type used is determined in the data type to ensure that the data type determined by each terminal device is the same, ensuring the accuracy of executing the model task.
- the method of the embodiment corresponding to Figure 5 i.e., steps 501 to 505 can be applied to "the multiple devices participating in executing the model task do not include network devices (that is, the multiple devices participating in executing the model task are all terminal devices). )" scene.
- the terminal device determines the data type to be used when executing the model task, where the data type determined by the terminal device is consistent with the data type among the multiple devices except the terminal device.
- the other devices determine the same data type. That is to say, in the present disclosure, the data types to be used when executing the model task determined by multiple devices that jointly participate in executing the model task are the same. Therefore, the multiple devices can execute the model task based on the same data type, then execute During the model task, there is no need to type-convert the usage data of multiple devices, thereby ensuring the execution accuracy of the model task, thereby ensuring the accuracy of the AI model obtained based on the model task, and improving the user experience of the AI model.
- Figure 6a is a schematic flowchart of a method for determining a data type provided by an embodiment of the present disclosure. Multiple devices jointly participate in executing model tasks. The method is executed by a terminal device among the multiple devices. In response to a response from a network among the multiple devices, Device, as shown in Figure 6a, the data type determination method may include the following steps:
- Step 601a Send the first data support capability information to the network device.
- step 601a please refer to the relevant introduction in the above embodiments, and the embodiments of this disclosure will not be repeated here.
- Step 602a Send the first capability information to the network device.
- the first capability information may be used to indicate at least one of the following:
- the storage capacity of the terminal device is the storage capacity of the terminal device
- the power consumption capability of the terminal device is the power consumption capability of the terminal device.
- each terminal device in the multiple devices needs to uniformly send the first capability to the network device. information.
- Step 603a Receive the instruction information sent by the network device.
- the indication information is used to indicate the data type to be used when executing the model task.
- the data type indicated by the indication information is specifically determined by the network device based on the first data support capability information and the first capability information sent by each terminal device.
- the instruction information received by each terminal device is the same, thereby ensuring that the data types to be used when executing model tasks determined by multiple devices are the same and unified, thus ensuring subsequent execution of the model. accuracy during the task.
- Step 604a Determine the data type to be used when executing the model task based on the instruction information.
- step 601a-step 604a can be applied to "multiple devices participating in executing the model task include network devices (ie, multiple devices participating in executing the model task include terminals at the same time). devices and network equipment)" scenario.
- the terminal device determines the data type to be used when executing the model task, where the data type determined by the terminal device is consistent with the data type among the multiple devices except the terminal device.
- the other devices determine the same data type. That is to say, in the present disclosure, the data types to be used when executing the model task determined by multiple devices that jointly participate in executing the model task are the same. Therefore, the multiple devices can execute the model task based on the same data type, then execute During the model task, there is no need to type-convert the usage data of multiple devices, thereby ensuring the execution accuracy of the model task, thereby ensuring the accuracy of the AI model obtained based on the model task, and improving the user experience of the AI model.
- Figure 6b is a schematic flowchart of a method for determining a data type provided by an embodiment of the present disclosure. Multiple devices jointly participate in executing model tasks. The method is executed by a terminal device among the multiple devices. In response to a response from a network among the multiple devices, Device, as shown in Figure 6b, the data type determination method may include the following steps:
- Step 601b Send the first capability information to the network device.
- Step 602b Receive instruction information sent by the network device.
- the indication information is used to indicate the data type to be used when executing the model task.
- the data type indicated by the indication information is specifically determined by the network device based on the first capability information sent by each terminal device and the data type supported by each terminal device. Specifically, the data types supported by each terminal device may be obtained by the network device from the core network.
- the instruction information received by each terminal device should be the same, thereby ensuring that the data types to be used when executing model tasks determined by multiple devices are the same and unified, thus ensuring subsequent execution.
- the accuracy of the model task should be the same.
- Step 603b Determine the data type to be used when executing the model task based on the instruction information.
- step 601b-step 603b can be applied to "multiple devices participating in executing the model task include network devices (ie, multiple devices participating in executing the model task include terminals at the same time). devices and network equipment)" scenario.
- the terminal device determines the data type to be used when executing the model task, where the data type determined by the terminal device is consistent with the data type among the multiple devices except the terminal device.
- the other devices determine the same data type. That is to say, in the present disclosure, the data types to be used when executing the model task determined by multiple devices that jointly participate in executing the model task are the same. Therefore, the multiple devices can execute the model task based on the same data type, then execute During the model task, there is no need to type-convert the usage data of multiple devices, thereby ensuring the execution accuracy of the model task, thereby ensuring the accuracy of the AI model obtained based on the model task, and improving the user experience of the AI model.
- Figure 7 is a schematic flowchart of a data type determination method provided by an embodiment of the present disclosure. Multiple devices jointly participate in executing model tasks. The method is executed by a terminal device among the multiple devices. As shown in Figure 7, the data Type determination methods may include the following steps:
- Step 701 Determine the data type to be used when executing the model task.
- Step 702 Execute the model task based on the determined data type.
- the model since the data type determined by the terminal device is the same as the data type determined by other devices, when multiple devices jointly participate in executing the model task, the model will be executed based on the same unified data type. tasks, ensuring execution accuracy.
- the terminal device determines the data type to be used when executing the model task, where the data type determined by the terminal device is consistent with the data type among the multiple devices except the terminal device.
- the other devices determine the same data type. That is to say, in the present disclosure, the data types to be used when executing the model task determined by multiple devices that jointly participate in executing the model task are the same. Therefore, the multiple devices can execute the model task based on the same data type, then execute During the model task, there is no need to type-convert the usage data of multiple devices, thereby ensuring the execution accuracy of the model task, thereby ensuring the accuracy of the AI model obtained based on the model task, and improving the user experience of the AI model.
- FIG 8 is a schematic flowchart of a data type determination method provided by an embodiment of the present disclosure.
- the method is executed by a network device.
- the network device and other devices jointly participate in executing model tasks.
- the data type determination method can Includes the following steps:
- Step 801 Determine the data type to be used by each device that performs the model task when performing the model task.
- the above-mentioned model tasks may include model training tasks and/or model inference tasks.
- the above-mentioned "multiple devices jointly participate in performing model tasks” can be understood as: multiple devices jointly perform a model task (such as a model training task or a model inference task), where different devices are used to be responsible for the model training task. At least part of the model tasks are implemented through interaction between various devices.
- the above-mentioned data type may include at least one of an integer (int), a single floating point number (ie, a 16-bit floating point number (float)), and a double floating point number (ie, a 32-bit floating point number (float)).
- the data type used by the network device to perform the model task determined by each device is the same, thereby ensuring that multiple devices use the same data type to perform the model task, and thus It ensures the unification of data types when multiple devices perform model tasks and ensures the accuracy of performing model tasks.
- step 801 The method of the embodiment corresponding to Figure 8 (i.e., step 801) is applied to "the multiple devices participating in executing the model task include network devices (that is, the multiple devices participating in executing the model task include both terminal devices and network devices)" A scene.
- step 801 the specific execution method of step 801 will be introduced in detail in subsequent embodiments.
- the terminal device determines the data type to be used when executing the model task, where the data type determined by the terminal device is consistent with the data type among the multiple devices except the terminal device.
- the other devices determine the same data type. That is to say, in the present disclosure, the data types to be used when executing the model task determined by multiple devices that jointly participate in executing the model task are the same. Therefore, the multiple devices can execute the model task based on the same data type, then execute During the model task, there is no need to type-convert the usage data of multiple devices, thereby ensuring the execution accuracy of the model task, thereby ensuring the accuracy of the AI model obtained based on the model task, and improving the user experience of the AI model.
- Figure 9 is a schematic flow chart of a data type determination method provided by an embodiment of the present disclosure.
- the method is executed by a network device.
- the network device and other devices jointly participate in executing model tasks.
- the data type determination method can Includes the following steps:
- Step 901 Determine the data types supported by other devices.
- the method of determining the data types supported by other devices may include any of the following:
- the data support capability information is used to indicate the data types supported by other devices;
- the core network device can store the data types supported by each device reported in advance.
- the data types supported by each device are specifically the data types that the device supports when accessing the core network device. reported to the core network equipment.
- Step 902 Determine the data types supported by the network device.
- Step 903 Determine at least one data type jointly supported by the network device and other devices based on the data types supported by the network device and the data types supported by other devices.
- step 903 please refer to the relevant introduction in the above embodiments, and the embodiments of this disclosure will not be described again here.
- Step 904 Determine the data type to be used when executing the model task from at least one data type commonly supported by the network device and other devices.
- the method for determining the data type to be used when executing the model task among at least one data type jointly supported by the network device and other devices may include at least one of the following:
- step 904 For a detailed introduction to step 904, please refer to the above-mentioned introduction to step 304, and the embodiment of the present disclosure will not be repeated here.
- the terminal device determines the data type to be used when executing the model task, where the data type determined by the terminal device is consistent with the data type among the multiple devices except the terminal device.
- the other devices determine the same data type. That is to say, in the present disclosure, the data types to be used when executing the model task determined by multiple devices that jointly participate in executing the model task are the same. Therefore, the multiple devices can execute the model task based on the same data type, then execute During the model task, there is no need to type-convert the usage data of multiple devices, thereby ensuring the execution accuracy of the model task, thereby ensuring the accuracy of the AI model obtained based on the model task, and improving the user experience of the AI model.
- Figure 10 is a schematic flowchart of a data type determination method provided by an embodiment of the present disclosure.
- the method is executed by a network device.
- the network device and other devices jointly participate in executing model tasks.
- the data type determination method can Includes the following steps:
- Step 1001 Determine the data types supported by other devices.
- step 1001 For a detailed introduction to step 1001, reference may be made to the description of the above embodiments, and the embodiments of the present disclosure will not be described again here.
- Step 1002 Receive capability information sent by other devices.
- the capability information may include at least one of the following:
- Step 1003 Determine the capabilities of the network device.
- the capability may include storage capability and/or power consumption capability.
- Step 1004 Determine at least one data type jointly supported by the network device and other devices based on the data types supported by the network device and the data types supported by other devices.
- step 1004 please refer to the relevant introduction in the above embodiments, and the embodiments of this disclosure will not be described again here.
- Step 1005 Determine the data type to be used when executing the model task from at least one commonly supported data type based on the first capability information and the capability of the network device.
- the method of determining the data type to be used when executing the model task from at least one commonly supported data type based on the first capability information and the capability of the network device may be: from the commonly supported data type Select a third data type that accurately matches the capability indicated by the first capability information and/or the capability of the network device from at least one data type, and determine the third data type as the data type to be used when performing the model task.
- step 1005 please refer to the relevant introduction in the above embodiments, and the embodiments of the present disclosure will not be repeated here.
- the terminal device determines the data type to be used when executing the model task, where the data type determined by the terminal device is consistent with the data type among the multiple devices except the terminal device.
- the other devices determine the same data type. That is to say, in the present disclosure, the data types to be used when executing the model task determined by multiple devices that jointly participate in executing the model task are the same. Therefore, the multiple devices can execute the model task based on the same data type, then execute During the model task, there is no need to type-convert the usage data of multiple devices, thereby ensuring the execution accuracy of the model task, thereby ensuring the accuracy of the AI model obtained based on the model task, and improving the user experience of the AI model.
- Figure 11a is a schematic flow chart of a data type determination method provided by an embodiment of the present disclosure.
- the method is executed by a network device.
- the network device and other devices jointly participate in executing model tasks.
- the data type determination method can Includes the following steps:
- Step 1101a Determine the data type to be used by each device that performs the model task when performing the model task.
- step 1101a For a detailed introduction to step 1101a, reference may be made to the description of the above embodiments, and the embodiments of the present disclosure will not be described again here.
- Step 1102a Send instruction information to other devices.
- the indication information is used to indicate the data type to be used when executing the model task.
- the network device will send the same instruction information to each other device, so that each other device can determine the same data type based on the same instruction information to perform the model task, ensuring that It unifies the data types of other devices when performing model tasks, ensuring the accuracy of performing model tasks.
- the network device should allocate specific model tasks to each other device. And, in one embodiment of the present disclosure, the network device may determine the data type to be used when executing the model task before allocating the model task to each other device, and then send the instruction information to each other device. Model tasks can also be assigned to various other devices simultaneously to indicate the type of data to be used when performing model tasks.
- the terminal device determines the data type to be used when executing the model task, where the data type determined by the terminal device is consistent with the data type among the multiple devices except the terminal device.
- the other devices determine the same data type. That is to say, in the present disclosure, the data types to be used when executing the model task determined by multiple devices that jointly participate in executing the model task are the same. Therefore, the multiple devices can execute the model task based on the same data type, then execute During the model task, there is no need to type-convert the usage data of multiple devices, thereby ensuring the execution accuracy of the model task, thereby ensuring the accuracy of the AI model obtained based on the model task, and improving the user experience of the AI model.
- Figure 11b is a schematic flow chart of a data type determination method provided by an embodiment of the present disclosure.
- the method is executed by a network device.
- the network device and other devices jointly participate in executing model tasks.
- the data type determination method can Includes the following steps:
- Step 1101b Determine the data type to be used by each device that performs the model task when performing the model task.
- Step 1102b Execute the model task based on the determined data type.
- steps 1101b to 1102b please refer to the above embodiment description, and the embodiments of the present disclosure will not be described again here.
- the terminal device determines the data type to be used when executing the model task, where the data type determined by the terminal device is consistent with the data type among the multiple devices except the terminal device.
- the other devices determine the same data type. That is to say, in the present disclosure, the data types to be used when executing the model task determined by multiple devices that jointly participate in executing the model task are the same. Therefore, the multiple devices can execute the model task based on the same data type, then execute During the model task, there is no need to type-convert the usage data of multiple devices, thereby ensuring the execution accuracy of the model task, thereby ensuring the accuracy of the AI model obtained based on the model task, and improving the user experience of the AI model.
- Figure 12 is a schematic structural diagram of a communication device provided by an embodiment of the present disclosure. As shown in Figure 12, the device may include:
- the determination module 1201 is used to determine the data type to be used when performing the model task, wherein the data type determined by the terminal device is the same as the data type determined by other devices in the plurality of devices.
- the terminal device determines the data type to be used when executing the model task, where the data type determined by the terminal device is consistent with other data types among the multiple devices except the terminal device.
- the data type determined by the device is the same. That is to say, in the present disclosure, the data types to be used when executing the model task determined by multiple devices that jointly participate in executing the model task are the same. Therefore, the multiple devices can execute the model task based on the same data type, then execute During the model task, there is no need to type-convert the usage data of multiple devices, thereby ensuring the execution accuracy of the model task, thereby ensuring the accuracy of the AI model obtained based on the model task, and improving the user experience of the AI model.
- the above device is also used for:
- First data support capability information is sent to the other devices respectively, where the first data support capability information is used to indicate the data type supported by the terminal device.
- the above device is also used for:
- the above-mentioned determination module 1201 is also used to:
- a data type determines the data type to be used when performing the model task.
- the above-mentioned determination module 1201 is also used to:
- the above device is also used for:
- first capability information Send first capability information to other devices respectively, where the first capability information is used to indicate at least one of the following:
- the storage capacity of the terminal device is the storage capacity of the terminal device
- the power consumption capability of the terminal device is the power consumption capability of the terminal device.
- the above device is also used for:
- Receive second data support capability information and second capability information sent by other devices the second data support capability information is used to indicate the data types supported by other devices, and the second capability information is used to indicate at least one of the following:
- the above device is also used for:
- the data type to be used when executing the model task is determined from the at least one commonly supported data type based on the first capability information and the second capability information.
- the above-mentioned determination module 1201 is also used to:
- the above device is also used for:
- the above-mentioned determination module 1202 is also used to:
- the instruction information is used to indicate the data type to be used when performing model tasks
- the above device is further configured to: perform a model task based on the determined data type.
- the model task includes: a model training task and/or a model inference task.
- Figure 13 is a schematic structural diagram of a communication device provided by an embodiment of the present disclosure. As shown in Figure 13, the device may include:
- the determination module 1301 is used to determine the data type to be used by each device that performs the model task when performing the model task, wherein the data types used by the network device for each device when performing the model task are the same.
- the terminal device determines the data type to be used when executing the model task, where the data type determined by the terminal device is consistent with other data types among the multiple devices except the terminal device.
- the data type determined by the device is the same. That is to say, in the present disclosure, the data types to be used when executing the model task determined by multiple devices that jointly participate in executing the model task are the same. Therefore, the multiple devices can execute the model task based on the same data type, then execute During the model task, there is no need to type-convert the usage data of multiple devices, thereby ensuring the execution accuracy of the model task, thereby ensuring the accuracy of the AI model obtained based on the model task, and improving the user experience of the AI model.
- the above-mentioned determination module 1301 is also used to:
- the above-mentioned determination module 1301 is also used to:
- the above-mentioned determination module 1301 is also used to:
- the above-mentioned determination module 1301 is also used to:
- the data type to be used when executing the model task is determined from the at least one commonly supported data type based on the business requirements of the model task.
- the device is also used for:
- the capability information includes at least one of the following:
- the above-mentioned determination module 1301 is also used to:
- the capabilities including storage capabilities and/or power consumption capabilities
- the data type to be used when performing the model task is determined from the at least one commonly supported data type based on the first capability information and the capability of the network device.
- the above-mentioned determination module 1301 is also used to:
- the third data type is determined as the data type to be used when performing the model task.
- the device is also used for:
- Instruction information is sent to other devices among the plurality of devices except the network device, where the instruction information is used to instruct the other devices to use data types when performing model tasks.
- the device is also used for:
- the model task includes: a model training task and/or a model inference task.
- FIG 14 is a schematic structural diagram of a communication device 1400 provided by an embodiment of the present application.
- the communication device 1400 may be a network device, a terminal device, a chip, a chip system, or a processor that supports a network device to implement the above method, or a chip, a chip system, or a processor that supports a terminal device to implement the above method. Processor etc.
- the device can be used to implement the method described in the above method embodiment. For details, please refer to the description in the above method embodiment.
- Communication device 1400 may include one or more processors 1401.
- the processor 1401 may be a general-purpose processor or a special-purpose processor, or the like.
- it can be a baseband processor or a central processing unit.
- the baseband processor can be used to process communication protocols and communication data.
- the central processor can be used to control communication devices (such as base stations, baseband chips, terminal equipment, terminal equipment chips, DU or CU, etc.) and execute computer programs. , processing data for computer programs.
- the communication device 1400 may also include one or more memories 1402, on which a computer program 1404 may be stored.
- the processor 1401 executes the computer program 1404, so that the communication device 1400 performs the steps described in the above method embodiments. method.
- the memory 1402 may also store data.
- the communication device 1400 and the memory 1402 can be provided separately or integrated together.
- the communication device 1400 may also include a transceiver 1405 and an antenna 1406.
- the transceiver 1405 may be called a transceiver unit, a transceiver, a transceiver circuit, etc., and is used to implement transceiver functions.
- the transceiver 1405 may include a receiver and a transmitter.
- the receiver may be called a receiver or a receiving circuit, etc., used to implement the receiving function;
- the transmitter may be called a transmitter, a transmitting circuit, etc., used to implement the transmitting function.
- the communication device 1400 may also include one or more interface circuits 1407.
- the interface circuit 1407 is used to receive code instructions and transmit them to the processor 1401 .
- the processor 1401 executes the code instructions to cause the communication device 1400 to perform the method described in the above method embodiment.
- the communication device 1400 is a terminal device: the transceiver 1405 is used to perform steps 301 to 302 in Figure 3; steps 401 to 402 in Figure 4; steps 501 to 503 in Figure 5; steps 601a to 601 in Figure 6a Step 603a; step 601b to step 602b in Figure 6b.
- the processor 1401 is used to execute step 201 in Figure 2; step 303-step 304 in Figure 3; step 403 in Figure 4; step 504 and step 505 in Figure 5; step 604a in Figure 6a; step 604 in Figure 6b Step 603b; Step 701-Step 702 in Figure 7.
- the communication device 1400 is a network device: the transceiver 1405 is used to perform step 1102a in Figure 11a.
- the processor 1401 is used to execute step 801 in Figure 8; steps 901 to 904 in Figure 9; steps 1001 to 1005 in Figure 10; step 1101a in Figure 11a; and steps 1101b and 1102b in Figure 11b.
- the processor 1401 may include a transceiver for implementing receiving and transmitting functions.
- the transceiver may be a transceiver circuit, an interface, or an interface circuit.
- the transceiver circuits, interfaces or interface circuits used to implement the receiving and transmitting functions can be separate or integrated together.
- the above-mentioned transceiver circuit, interface or interface circuit can be used for reading and writing codes/data, or the above-mentioned transceiver circuit, interface or interface circuit can be used for signal transmission or transfer.
- the processor 1401 may store a computer program 1403, and the computer program 1403 runs on the processor 1401, causing the communication device 1400 to perform the method described in the above method embodiment.
- the computer program 1403 may be solidified in the processor 1401, in which case the processor 1401 may be implemented by hardware.
- the communication device 1400 may include a circuit, which may implement the functions of sending or receiving or communicating in the foregoing method embodiments.
- the processor and transceiver described in this application can be implemented in integrated circuits (ICs), analog ICs, radio frequency integrated circuits RFICs, mixed signal ICs, application specific integrated circuits (ASICs), printed circuit boards ( printed circuit board (PCB), electronic equipment, etc.
- the processor and transceiver can also be manufactured using various IC process technologies, such as complementary metal oxide semiconductor (CMOS), n-type metal oxide-semiconductor (NMOS), P-type Metal oxide semiconductor (positive channel metal oxide semiconductor, PMOS), bipolar junction transistor (BJT), bipolar CMOS (BiCMOS), silicon germanium (SiGe), gallium arsenide (GaAs), etc.
- CMOS complementary metal oxide semiconductor
- NMOS n-type metal oxide-semiconductor
- PMOS P-type Metal oxide semiconductor
- BJT bipolar junction transistor
- BiCMOS bipolar CMOS
- SiGe silicon germanium
- GaAs gallium arsenide
- the communication device described in the above embodiments may be a network device or a terminal device, but the scope of the communication device described in this application is not limited thereto, and the structure of the communication device may not be limited by FIG. 14 .
- the communication device may be a stand-alone device or may be part of a larger device.
- the communication device may be:
- the IC collection may also include storage components for storing data and computer programs;
- the communication device may be a chip or a chip system
- the schematic structural diagram of the chip shown in FIG. 15 refer to the schematic structural diagram of the chip shown in FIG. 15 .
- the chip shown in Figure 15 includes a processor 1501 and an interface 1502.
- the number of processors 1501 may be one or more, and the number of interfaces 1502 may be multiple.
- the chip also includes a memory 1503, which is used to store necessary computer programs and data.
- This application also provides a readable storage medium on which instructions are stored. When the instructions are executed by a computer, the functions of any of the above method embodiments are implemented.
- This application also provides a computer program product, which, when executed by a computer, implements the functions of any of the above method embodiments.
- the above embodiments it may be implemented in whole or in part by software, hardware, firmware, or any combination thereof.
- software it may be implemented in whole or in part in the form of a computer program product.
- the computer program product includes one or more computer programs.
- the computer program When the computer program is loaded and executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part.
- the computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device.
- the computer program may be stored in or transferred from one computer-readable storage medium to another, for example, the computer program may be transferred from a website, computer, server, or data center Transmission to another website, computer, server or data center through wired (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.) means.
- 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 one or more available media integrated.
- the usable media may be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., high-density digital video discs (DVD)), or semiconductor media (e.g., solid state disks, SSD)) etc.
- magnetic media e.g., floppy disks, hard disks, magnetic tapes
- optical media e.g., high-density digital video discs (DVD)
- DVD digital video discs
- semiconductor media e.g., solid state disks, SSD
- At least one in this application can also be described as one or more, and the plurality can be two, three, four or more, which is not limited by this application.
- the technical feature is distinguished by “first”, “second”, “third”, “A”, “B”, “C” and “D”, etc.
- the technical features described in “first”, “second”, “third”, “A”, “B”, “C” and “D” are in no particular order or order.
- the corresponding relationships shown in each table in this application can be configured or predefined.
- the values of the information in each table are only examples and can be configured as other values, which are not limited by this application.
- the corresponding relationships shown in some rows may not be configured.
- appropriate deformation adjustments can be made based on the above table, such as splitting, merging, etc.
- the names of the parameters shown in the titles of the above tables may also be other names understandable by the communication device, and the values or expressions of the parameters may also be other values or expressions understandable by the communication device.
- other data structures can also be used, such as arrays, queues, containers, stacks, linear lists, pointers, linked lists, trees, graphs, structures, classes, heaps, hash tables or hash tables. wait.
- Predefinition in this application can be understood as definition, pre-definition, storage, pre-storage, pre-negotiation, pre-configuration, solidification, or pre-burning.
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
本公开提出一种数据类型确定方法/装置/设备/存储介质,属于通信技术领域。终端设备会确定执行模型任务时所要使用的数据类型,其中,终端设备确定的数据类型与多个设备中的其他设备确定的数据类型相同。由此可知,本公开提供了一种数据类型确定方法,确保了模型任务的执行精度,进而确保了基于模型任务所得到的AI模型的精度,提高了AI模型的用户体验度。
Description
本公开涉及通信技术领域,尤其涉及一种数据类型确定方法/装置/设备及存储介质。
随着AI(ArtificialIntelligent,人工智能)技术的不断发展,AI模型在多个领域得到广泛应用。其中,网络节点在应用AI模型前,通常需要先执行模型任务(如AI模型的训练任务和/或AI模型的部署推理任务)。相关技术中,可能需要多个不同的网络节点来共同参与完成模型任务。
但是,相关技术中,不同的网络节点可能支持不同的数据类型。基于此,当多个网络节点共同参与模型任务时,可能会出现不同网络节点使用不同类型的数据类型的情形,则可能会降低多个网络节点执行模型任务时的准确度(如可能由于需要转换不同网络节点之间的数据类型而导致AI模型的训练准确度降低或推理准确度降低),进而导致最终得到的AI模型准确度较低,降低了AI模型的用户体验度。
发明内容
本公开提出的数据类型确定方法/装置/设备及存储介质,用于解决相关技术的方法中导致AI模型准确度较低的技术问题。
第一方面,本公开实施例提供一种数据类型确定方法,多个设备共同参与执行模型任务,所述方法由所述多个设备中的终端设备执行,包括:
确定执行所述模型任务时所要使用的数据类型,其中,所述终端设备确定的数据类型与所述多个设备中的其他设备确定的数据类型相同。
本公开提供的数据类型确定方法中,共同参与执行模型任务的多个设备所确定的执行模型任务时所要使用的数据类型均相同,由此该多个设备可以基于相同的数据类型来执行模型任务,则执行模型任务期间无需对多个设备的使用数据进行类型转换,从而确保了模型任务的执行精度,进而确保了基于模型任务所得到的AI模型的精度,提高了AI模型的用户体验度。
第二方面,本公开实施例提供一种数据类型确定方法,该方法由网络设备执行,所述网络设备与其他设备共同参与执行模型任务,包括:
确定执行所述模型任务的各个设备执行所述模型任务时所要使用的数据类型,其中,所述网络设备针对所述各个设备确定的执行所述模型任务时使用的数据类型相同。
第三方面,本公开实施例提供一种通信装置,该装置被配置在终端设备中,包括:
确定模块,用于确定执行所述模型任务时所要使用的数据类型,其中,所述终端设备确定的数据类型与所述多个设备中的其他设备确定的数据类型相同。
第四方面,本公开实施例提供一种通信装置,该装置被配置在网络设备中,包括:
确定模块,用于确定执行所述模型任务的各个设备执行所述模型任务时所要使用的数据类型,其中,所述网络设备针对所述各个设备确定的执行所述模型任务时使用的数据类型相同。
第五方面,本公开实施例提供一种通信装置,该通信装置包括处理器,当该处理器调用存储器中的计算机程序时,执行上述第一方面所述的方法。
第六方面,本公开实施例提供一种通信装置,该通信装置包括处理器,当该处理器调用存储器中的计算机程序时,执行上述第二方面所述的方法。
第七方面,本公开实施例提供一种通信装置,该通信装置包括处理器和存储器,该存储器中存储有计算机程序;所述处理器执行该存储器所存储的计算机程序,以使该通信装置执行上述第一方面所述的方法。
第八方面,本公开实施例提供一种通信装置,该通信装置包括处理器和存储器,该存储器中存储有计算机程序;所述处理器执行该存储器所存储的计算机程序,以使该通信装置执行上述第二方面所述的方法。
第九方面,本公开实施例提供一种通信装置,该装置包括处理器和接口电路,该接口电路用于接收代码指令并传输至该处理器,该处理器用于运行所述代码指令以使该装置执行上述第一方面所述的方法。
第十方面,本公开实施例提供一种通信装置,该装置包括处理器和接口电路,该接口电路用于接收代码指令并传输至该处理器,该处理器用于运行所述代码指令以使该装置执行上述第二方面所述的方法。
第十一方面,本公开实施例提供一种通信系统,该系统包括第三方面所述的通信装置至第四方面所述的通信装置,或者,该系统包括第五方面所述的通信装置至第六方面所述的通信装置,或者,该系统包括第七方面所述的通信装置至第八方面所述的通信装置,或者,该系统包括第九方面所述的通信装置至第十方面所述的通信装置。
第十二方面,本发明实施例提供一种计算机可读存储介质,用于储存为上述网络设备所用的指令,当所述指令被执行时,使所述终端设备执行上述第一方面至第三方面的任一方面所述的方法。
第十三方面,本公开还提供一种包括计算机程序的计算机程序产品,当其在计算机上运行时,使得计算机执行上述第一方面至第二方面的任一方面所述的方法。
第十四方面,本公开提供一种芯片系统,该芯片系统包括至少一个处理器和接口,用于支持网络设备实现第一方面至第二方面的任一方面所述的方法所涉及的功能,例如,确定或处理上述方法中所涉及的数据和信息中的至少一种。在一种可能的设计中,所述芯片系统还包括存储器,所述存储器,用于保存源辅节点必要的计算机程序和数据。该芯片系统,可以由芯片构成,也可以包括芯片和其他分立器件。
第十五方面,本公开提供一种计算机程序,当其在计算机上运行时,使得计算机执行上述第一方面至第二方面的任一方面所述的方法。
本公开上述的和/或附加的方面和优点从下面结合附图对实施例的描述中将变得明显和容易理解,其中:
图1为本公开实施例提供的一种通信系统的架构示意图;
图2为本公开另一个实施例所提供的数据类型确定方法的流程示意图;
图3为本公开再一个实施例所提供的数据类型确定方法的流程示意图;
图4为本公开又一个实施例所提供的数据类型确定方法的流程示意图;
图5为本公开另一个实施例所提供的数据类型确定方法的流程示意图;
图6a为本公开再一个实施例所提供的数据类型确定方法的流程示意图;
图6b为本公开再一个实施例所提供的数据类型确定方法的流程示意图;
图7为本公开又一个实施例所提供的数据类型确定方法的流程示意图;
图8为本公开一个实施例所提供的数据类型确定方法的流程示意图;
图9为本公开另一个实施例所提供的数据类型确定方法的流程示意图;
图10为本公开再一个实施例所提供的数据类型确定方法的流程示意图;
图11a为本公开又一个实施例所提供的数据类型确定方法的流程示意图;
图11b为本公开又一个实施例所提供的数据类型确定方法的流程示意图;
图12为本公开一个实施例所提供的通信装置的结构示意图;
图13为本公开另一个实施例所提供的通信装置的结构示意图;
图14是本公开一个实施例所提供的一种用户设备的框图;
图15为本公开一个实施例所提供的一种网络侧设备的框图。
这里将详细地对示例性实施例进行说明,其示例表示在附图中。下面的描述涉及附图时,除非另有表示,不同附图中的相同数字表示相同或相似的要素。以下示例性实施例中所描述的实施方式并不代表与本公开实施例相一致的所有实施方式。相反,它们仅是与如所附权利要求书中所详述的、本公开实施例的一些方面相一致的装置和方法的例子。
在本公开实施例使用的术语是仅仅出于描述特定实施例的目的,而非旨在限制本公开实施例。在本公开实施例和所附权利要求书中所使用的单数形式的“一种”和“该”也旨在包括多数形式,除非上下文清楚地表示其他含义。还应当理解,本文中使用的术语“和/或”是指并包含一个或多个相关联的列出项目的任何或所有可能组合。
应当理解,尽管在本公开实施例可能采用术语第一、第二、第三等来描述各种信息,但这些信息不应限于这些术语。这些术语仅用来将同一类型的信息彼此区分开。例如,在不脱离本公开实施例范围的情况下,第一信息也可以被称为第二信息,类似地,第二信息也可以被称为第一信息。取决于语境,如在此所使用的词语“如果”及“若”可以被解释成为“在……时”或“当……时”或“响应于确定”。
下面详细描述本公开的实施例,所述实施例的示例在附图中示出,其中自始至终相同或类似的标号表示相同或类似的要素。下面通过参考附图描述的实施例是示例性的,旨在用于解释本公开,而不能理解为对本公开的限制。
为了便于理解,首先介绍本申请涉及的术语。
1、人工智能(Artificial Intelligence,AI)
AI是研究、开发用于模拟、延伸和扩展人的智能的理论、方法、技术及应用系统的一门新的技术科学。
本公开实施例中涉及到的各种网元/功能,其既可以是一个独立的硬件设备,也可以是在硬件设备内的通过计算机代码实现的功能,本公开实施例中并不对此做出限定。
请参见图1,图1为本公开实施例提供的一种通信系统的架构示意图。该通信系统可包括但不限于一个网络设备和一个终端设备,图1所示的设备数量和形态仅用于举例并不构成对本公开实施例的限定,实际应用中可以包括两个或两个以上的网络设备,两个或两个以上的终端设备。图1所示的通信系统以包括一个网络设备11、一个终端设备12为例。
需要说明的是,本公开实施例的技术方案可以应用于各种通信系统。例如:长期演进(long term evolution,LTE)系统、第五代(5th generation,5G)移动通信系统、5G新空口(new radio,NR)系统,或者其他未来的新型移动通信系统等。
本公开实施例中的网络设备11是网络侧的一种用于发射或接收信号的实体。例如,网络设备11可以为演进型基站(evolved NodeB,eNB)、发送接收点(transmission reception point,TRP)、NR系统中的下一代基站(next generation NodeB,gNB)、其他未来移动通信系统中的基站或无线保真(wireless fidelity,WiFi)系统中的接入节点等。本公开的实施例对网络设备所采用的具体技术和具体设备形态不做限定。本公开实施例提供的网络设备可以是由集中单元(central unit,CU)与分布式单元(distributed unit,DU)组成的,其中,CU也可以称为控制单元(control unit),采用CU-DU的结构可以将网络设备,例如基站的协议层拆分开,部分协议层的功能放在CU集中控制,剩下部分或全部协议层的功能分布在DU中,由CU集中控制DU。
本公开实施例中的终端设备12是用户侧的一种用于接收或发射信号的实体,如手机。终端设备也可以称为终端设备(terminal)、用户设备(user equipment,UE)、移动台(mobile station,MS)、移动终端设备(mobile terminal,MT)等。终端设备可以是具备通信功能的汽车、智能汽车、手机(mobile phone)、穿戴式设备、平板电脑(Pad)、带无线收发功能的电脑、虚拟现实(virtual reality,VR)终端设备、增强现实(augmented reality,AR)终端设备、工业控制(industrial control)中的无线终端设备、无人驾驶(self-driving)中的无线终端设备、远程手术(remote medical surgery)中的无线终端设备、智能电网(smart grid)中的无线终端设备、运输安全(transportation safety)中的无线终端设备、智慧城市(smart city)中的无线终端设备、智慧家庭(smart home)中的无线终端设备等等。本公开的实施例对终端设备所采用的具体技术和具体设备形态不做限定。
可以理解的是,本公开实施例描述的通信系统是为了更加清楚的说明本公开实施例的技术方案,并不构成对于本公开实施例提供的技术方案的限定,本领域普通技术人员可知,随着系统架构的演变和新业务场景的出现,本公开实施例提供的技术方案对于类似的技术问题,同样适用。
下面参考附图对本公开实施例所提供的数据类型确定方法/装置/设备及存储介质进行详细描述。
图2为本公开实施例所提供的一种数据类型确定方法的流程示意图,其中,本公开中多个设备共同参与执行模型任务,所述方法由多个设备中的终端设备执行,如图2所示,该数据类型确定方法可以包括以下步骤:
步骤201、确定执行模型任务时所要使用的数据类型。
其中,在本公开的一个实施例之中,上述模型任务可以包括模型训练任务,和/或,模型推理任务。以及,上述的“多个设备共同参与执行模型任务”可以理解为:由多个设备来共同执行一模型任务(如模型训练任务或模型推理任务),其中,不同设备用于负责模型训练任务中的至少一部分,各个设备之间通过交互来实现该模型任务。
以及,上述的数据类型可以包括整型(int)、单浮点数(即16位浮点数(float))、双浮点数(即32位浮点数(float))中的至少一种。
以及,在本公开的一个实施例之中,上述的参与执行模型任务的多个设备中可以包括有网络设备也可以不包括有网络设备。其中,当包括或不包括网络设备时,终端设备对于执行模型任务时所要使用的数据类型的确定方法也会有所不同。
具体的,在本公开的一个实施例之中,当多个设备中包括网络设备时,各个终端设备统一会向该网络设备上报其支持的数据类型相关的能力信息(即后续提到的第一数据支持能力信息或第二数据支持能力信息),以便网络设备基于接收到的能力信息确定执行模型任务的数据类型,并将确定的数据类型的指示信息分别下发至各个终端设备,使得各个终端设备可以基于指示信息确定执行模型任务时所要使用的数据类型。
以及,在本公开的另一个实施例之中,当多个设备不包括网络设备(即该多个设备均为终端设备)时,该各个设备之间会交互其支持的数据类型相关的能力信息,以便各个设备均知晓其他设备支持的数据类型,从而各个设备即可基于参与执行模型任务的多个设备所支持的数据类型自主确定执行模型任务时所要使用的数据类型。其中,关于上述内容的具体执行方法会在后续实施例中进行详细介绍。
此外,需要说明的是,在本公开的一个实施例之中,终端设备通过执行上述步骤201所确定的数据类型与上述的多个设备中的其他设备确定的数据类型应当是相同的,由此以确保终端设备可以与其他设备采用相同的数据类型来执行模型任务,进而保证了多个设备在执行模型任务时的数据类型的统一,确保了执行模型任务的精度。
综上所述,在本公开实施例提供的数据类型确定方法之中,终端设备会确定执行模型任务时所要使用的数据类型,其中,终端设备确定的数据类型与多个设备中除终端设备外的其他设备确定的数据类型相同。也即是,本公开中共同参与执行模型任务的多个设备所确定的执行模型任务时所要使用的数据类型均相同,由此该多个设备可以基于相同的数据类型来执行模型任务,则执行模型任务期间无需对多个设备的使用数据进行类型转换,从而确保了模型任务的执行精度,进而确保了基于模型任务所得到的AI模型的精度,提高了AI模型的用户体验度。
图3为本公开实施例所提供的一种数据类型确定方法的流程示意图,该方法由终端设备执行,多个设备共同参与执行模型任务,所述方法由多个设备中的终端设备执行,如图3所示,该数据类型确定方法可以包括以下步骤:
步骤301、分别向其他设备发送终端设备的第一数据支持能力信息。
其中,在本公开的一个实施例之中,上述第一数据支持能力信息用于指示终端设备支持的数据类型。该终端设备支持的数据类型可以包括整型、单浮点数、双浮点数中的至少一种。
需要说明的是,不同的数据类型的精度会有所不同。但是,应当认识到,当终端设备支持精度较高的某数据类型时,也隐式指示该终端设备支持精度低于该数据类型的数据类型。示例的,假设终端设备支持的数据类型为双浮点数,其中,由于整型的精度小于单浮点数的精度且小于双浮点数的精度,因此,当终端设备支持的数据类型为双浮点数时,则隐式说明该终端设备也支持整型和单浮点数。
此外,在本公开的一个实施例之中,上述的第一数据支持能力信息可以是终端设备通过D2D(Device to Device,设备到设备)方法向其他设备发送的。
步骤302、接收其他设备发送的第二数据支持能力信息。
其中,在本公开的一个实施例之中,第二数据支持能力信息用于指示其他设备支持的数据类型。以及,其他设备支持的数据类型可以为整型、单浮点数、双浮点数中的至少一种。
以及,在本公开的一个实施例之中,该第二数据支持能力信息可以是其他设备通过D2D方式发送至该终端设备的。
则由步骤301-302可知,在本公开的一个实施例之中,当参与执行模型任务的多个设备中不包括网络设备时,该多个设备之间会互相交互其数据支持能力信息,从而使得参与模型任务的各个设备均知晓除自身外的其他设备支持的数据类型,以便各个设备后续可以基于其他设备支持的数据类型和自身支持的数据类型来统一确定出最终执行模型任务时所要使用的数据类型(即后续步骤303-304)。
步骤303、基于第一数据支持能力信息和第二数据支持能力信息确定终端设备和其他设备共同支持的至少一种数据类型。
其中,在本公开的一个实施例之中,可以将第一数据支持能力信息和第二数据支持能力信息共同指示的数据类型确定为终端设备和其它设备共同支持的数据类型。
示例的,在本公开的一个实施例之中,假设终端设备的第一数据支持能力信息指示终端设备支持的数据类型为:整型和单浮点数,其他设备#1的第二数据支持能力信息指示该其他设备#1支持的数据类型为整型、单浮点数和双浮点数,则确定终端设备和其他设备#1共同支持的数据类型为整型和单浮点数。
步骤304、从终端设备和其他设备共同支持的至少一种数据类型中确定执行模型任务时所要使用的数据类型。
其中,在本公开的一个实施例之中,从终端设备和其他设备共同支持的至少一种数据类型中确定执行模型任务时所要使用的数据类型的方法可以包括以下至少一种:
方法一、将共同支持的至少一种数据类型中精度最大的第一数据类型确定为执行模型任务时所要使用的数据类型。
其中,在本公开的一个实施例之中,数据类型的精度与该数据类型对应的比特位数呈正比。即,比特位数越多的数据类型的精度越大。例如:整型的比特位数为8位、单浮点数的比特位数为32位、双浮点数的比特位数为64位,则该三种数据类型对应的精度大小关系依次为:双浮点数>单浮点数>整型。
基于此,在本公开的一个实施例之中,上述将共同支持的至少一种数据类型中精度最大的第一数据类型确定为执行模型任务时所要使用的数据类型的方法即为:将共同支持的数据类型中比特位数最多的第一数据类型确定为执行模型任务时所要使用的数据类型。
示例的,在本公开的一个实施例之中,假设共同支持的数据类型为整型和单浮点数,其中,单浮点数的比特位数大于整型的比特位数,单浮点数为共同支持的数据类型中精度最大的数据,此时可以直接将单浮点数确定为执行模型任务时所要使用的数据类型。
方法二、将共同支持的至少一种数据类型中精度小于第一数据类型的第二数据类型确定为执行模型任务时所要使用的数据类型。
其中,在本公开的一个实施例之中,也可以不选取共同支持的数据类型中的最大精度的数据类型作为执行模型任务时所要使用的数据类型,可以选取精度小于最大精度的数据类型作为执行模型任务时所要使用的数据类型,以此在确保模型任务的执行精度的同时,还可以降低执行成本。
示例的,在本公开的一个实施例之中,假设共同支持的数据类型为整型、单浮点数、双浮点数,则可以确定出精度最高的数据类型(即前述的第一数据类型)为双浮点数,此时,可以将共同支持的数据类型中从精度小于双浮点数的数据类型(如整型或单浮点数)确定为执行模型任务时所要使用的数据类型。
方法三、基于模型任务的期望资源开销从共同支持的至少一种数据类型中确定执行模型任务时所要使用的数据类型。
其中,在本公开的一个实施例之中,若模型任务的期望资源开销较小,则选择比特位数少精度较低的数据类型;若对模型任务的期望资源开销没有限制,则选择比特位数大精度较高的数据类型。
示例的,在本公开的一个实施例之中,假设共同支持的数据类型为整型、单浮点数、双浮点数,此时若模型任务的期望资源开销较小,则可以选择整型作为执行模型任务时所要使用的数据类型;若对模型任务的期望资源开销没有限制,则可以选择双浮点数作为执行模型任务时所要使用的数据类型。
方法四、基于模型任务的业务需求从共同支持的至少一种数据类型中确定执行模型任务时所要使用的数据类型。
其中,在本公开的一个实施例之中,该业务需求可以为模型任务的时延要求。基于此,上述的基于模型任务的业务需求从共同支持的至少一种数据类型中确定执行模型任务时所要使用的数据类型的方法可以包括:若模型任务对时延要求较高,则选择比特位数少精度小的数据类型;若模型任务时延要求较低,则选择比特位数多精度高的数据类型。
示例的,在本公开的一个实施例之中,假设共同支持的数据类型为整型、单浮点数,此时若模型任务的业务需求对时延要求高,则选择整型;若模型任务的业务需求对时延要求低,则选择单浮点数。
此外,需要说明的是,在本公开的一个实施例之中,多个设备中的各个终端设备在通过执行上述步骤303-304以自主从共同支持的数据类型中确定使用的数据类型时,各个终端设备均应当是采用相同的方法来从共同支持的数据类型中确定使用的数据类型。如,各个终端设备可以均采用上述方法三来确定执行模型任务时所要使用的数据类型,或者,各个终端设备可以均采用上述方法一来确定为执行模型任务时所要使用的数据类型,从而确保各个终端设备确定的数据类型相同,确保了执行模型任务的精度。
其中,图3对应的实施例的方法(即步骤301-步骤304)可以是应用于“参与执行模型任务的多个设备中不包括网络设备(即参与执行模型任务的多个设备均为终端设备)”这一场景。
综上所述,在本公开实施例提供的数据类型确定方法之中,终端设备会确定执行模型任务时所要使用的数据类型,其中,终端设备确定的数据类型与多个设备中除终端设备外的其他设备确定的数据类型相同。也即是,本公开中共同参与执行模型任务的多个设备所确定的执行模型任务时所要使用的数据类型均相同,由此该多个设备可以基于相同的数据类型来执行模型任务,则执行模型任务期间无需对多个设备的使用数据进行类型转换,从而确保了模型任务的执行精度,进而确保了基于模型任务所得到的AI模型的精度,提高了AI模型的用户体验度。
图4为本公开实施例所提供的一种数据类型确定方法的流程示意图,多个设备共同参与执行模型任务,所述方法由多个设备中的终端设备执行,响应于多个设备中包括网络设备,如图4所示,该数据类型确定方法可以包括以下步骤:
步骤401、向网络设备发送第一数据支持能力信息。
其中,在本公开的一个实施例之中,上述第一数据支持能力信息用于指示终端设备支持的数据类型。
需要说明的是,在本公开的一个实施例之中,当参与执行模型任务的多个设备中包括有网络设备和终端设备时,该多个设备中的各个终端设备均需要统一向网络设备发送第一数据支持能力信息。
步骤402、接收网络设备发送的指示信息。
其中,在本公开的一个实施例之中,指示信息用于指示执行模型任务时所要使用的数据类型。该指示信息指示的数据类型具体是网络设备基于各个终端设备发送的第一数据支持能力信息确定的。
以及,在本公开的一个实施例之中,各个终端设备接收到的指示信息是相同,从而确保多个设备确定的执行模型任务时所要使用的数据类型是相同统一的,则确保了后续执行模型任务时的精度。
步骤403、基于指示信息确定执行模型任务时所要使用的数据类型。
其中,图4对应的实施例的方法(即步骤401-步骤403)可以是应用于“参与执行模型任务的多个设备中包括网络设备(即参与执行模型任务的多个设备中同时包括有终端设备和网络设备)”这一场景。
综上所述,在本公开实施例提供的数据类型确定方法之中,终端设备会确定执行模型任务时所要使用的数据类型,其中,终端设备确定的数据类型与多个设备中除终端设备外的其他设备确定的数据类型相同。也即是,本公开中共同参与执行模型任务的多个设备所确定的执行模型任务时所要使用的数据类型均相同,由此该多个设备可以基于相同的数据类型来执行模型任务,则执行模型任务期间无需对多个设备的使用数据进行类型转换,从而确保了模型任务的执行精度,进而确保了基于模型任务所得到的AI模型的精度,提高了AI模型的用户体验度。
图5为本公开实施例所提供的一种数据类型确定方法的流程示意图,多个设备共同参与执行模型任务,所述方法由多个设备中的终端设备执行,响应于多个设备中不包括网络设备,如图5所示,该数据类型确定方法可以包括以下步骤:
步骤501、分别向其他设备发送终端设备的第一数据支持能力信息。
其中,关于步骤501的其他相关介绍可以参考上述实施例描述。
步骤502、分别向其他设备发送终端设备的第一能力信息。
其中,在本公开的一个实施例之中,第一能力信息可以用于指示以下至少一种:
终端设备的存储能力;
终端设备的功耗能力。
以及,在本公开的一个实施例之中,该第一能力信息可以是终端设备通过D2D方式发送至其他设备的。
以及,在本公开的一个实施例之中,上述的终端设备的存储能力可以是终端设备当前的存储能力,上述的终端设备的功耗能力可以是终端设备当前的功耗能力。
步骤503、接收其他设备发送的第二数据支持能力信息和第二能力信息。
其中,在本公开的一个实施例之中,第二数据支持能力信息用于指示其他设备支持的数据类型,第二能力信息可以用于指示以下至少一种:
其他设备的存储能力;
其他设备的功耗能力。
其中,上述的其他设备的存储能力可以是其他设备当前的存储能力,上述的其他设备的功耗能力可以是其他设备当前的功耗能力。
则由步骤501-503可知,在本公开的一个实施例之中,当参与执行模型任务的多个设备中不包括网络设备时,该多个设备之间会互相交互其数据支持能力、存储能力、功耗能力中的至少一种,从而使得参与模型任务的各个设备均知晓除自身外的其他设备支持的数据类型和其他设备的存储能力和/或功耗能力,以便各个设备后续可以基于其他设备支持的数据类型和其他设备的存储能力和/或功耗能力并结合自身支持的数据类型和自身设备的存储能力和/或功耗能力来统一确定出最终执行模型任务时所要使用的数据类型(即后续步骤504-505)。
步骤504、基于第一数据支持能力信息和第二数据支持能力信息确定终端设备和其他设备共同支持的至少一种数据类型。
其中,关于步骤504的详细介绍可以参考上述步骤303的描述,本公开实施例在此不做赘述。
步骤505、基于第一能力信息和第二能力信息从共同支持的至少一种数据类型中确定执行模型任务时所要使用的数据类型。
其中,在本公开的一个实施例之中,基于第一能力信息和第二能力信息从共同支持的至少一种数据类型中确定执行模型任务时所要使用的数据类型的方法可以包括:从共同支持的至少一种数据类型中选择精度匹配于第一能力信息和/或第二能力信息所指示的能力的第三数据类型,将第三数据类型确定为执行模型任务时所要使用的数据类型。
具体的,在本公开的一个实施例之中,终端设备是基于参与执行模型任务的多个设备的存储能力和/或功耗能力来综合考虑确定第三数据类型的。其中,第三数据类型的精度是匹配于多个设备的综合的存储能力和/或功耗能力的。如:当多个设备的综合的存储能力和/或功耗能力越强,则说明参与执行模型任务的多个设备可以支持精度较高比特位数较多的数据类型,此时所选择的该第三数据类型的精度可以较高比特位数可以较多;当多个设备的综合的存储能力和/或功耗能力越弱,则说明参与执行模型任务的多个设备可以支持精度较低比特位数较少的数据类型,此时所选择的该第三数据类型的精度可以较低比特位数可以较少。
示例的,在本公开的一个实施例之中,假设共同支持的数据类型为整型、单浮点数,此时若第一能力信息和/或第二能力信息中存储能力较强时,则可以选择双浮点数作为执行模型任务时所要使用的数据类型;若第一能力信息和/或第二能力信息中存储能力较弱时,则可以选择整型作为执行模型任务时所要使用的数据类型。
此外,需要说明的是,在本公开的一个实施例之中,多个设备中的各个终端设备在通过执行上述步骤504-505以自主从共同支持的数据类型中确定使用的数据类型时,各个终端设备均应当是采用相同的方法(或相同的规则)从共同支持的数据类型中确定使用的数据类型,如各个终端设备均“以存储能力和/或功耗能力与所确定的使用的数据类型的精度成正相关”的方法(或规则),或者,“以存储能力和/ 或功耗能力与所确定的使用的数据类型的精度成负相关”的方法(或规则)来从共同支持的数据类型中确定使用的数据类型,以此确保各个终端设备确定的数据类型相同,确保了执行模型任务的精度。
其中,图5对应的实施例的方法(即步骤501-步骤505)可以是应用于“参与执行模型任务的多个设备中不包括网络设备(即参与执行模型任务的多个设备为均终端设备)”这一场景。
关于本公开实施例中的其他内容的详细介绍,可以参考上述实施例中的相关介绍。
综上所述,在本公开实施例提供的数据类型确定方法之中,终端设备会确定执行模型任务时所要使用的数据类型,其中,终端设备确定的数据类型与多个设备中除终端设备外的其他设备确定的数据类型相同。也即是,本公开中共同参与执行模型任务的多个设备所确定的执行模型任务时所要使用的数据类型均相同,由此该多个设备可以基于相同的数据类型来执行模型任务,则执行模型任务期间无需对多个设备的使用数据进行类型转换,从而确保了模型任务的执行精度,进而确保了基于模型任务所得到的AI模型的精度,提高了AI模型的用户体验度。
图6a为本公开实施例所提供的一种数据类型确定方法的流程示意图,多个设备共同参与执行模型任务,所述方法由多个设备中的终端设备执行,响应于多个设备中包括网络设备,如图6a所示,该数据类型确定方法可以包括以下步骤:
步骤601a、向网络设备发送第一数据支持能力信息。
关于步骤601a的详细介绍可以参考上述实施例中的相关介绍,本公开实施例在此不做赘述。
步骤602a、向网络设备发送第一能力信息。
其中,在本公开的一个实施例之中,第一能力信息可以用于指示以下至少一种:
终端设备的存储能力;
终端设备的功耗能力。
以及,在本公开的一个实施例之中,当参与执行模型任务的多个设备中包括有网络设备和终端设备时,该多个设备中的各个终端设备均需要统一向网络设备发送第一能力信息。
步骤603a、接收网络设备发送的指示信息。
其中,在本公开的一个实施例之中,指示信息用于指示执行模型任务时所要使用的数据类型。该指示信息指示的数据类型具体是网络设备基于各个终端设备发送的第一数据支持能力信息和第一能力信息确定的。
以及,在本公开的一个实施例之中,各个终端设备接收到的指示信息是相同,从而确保多个设备确定的执行模型任务时所要使用的数据类型是相同统一的,则确保了后续执行模型任务时的精度。
步骤604a、基于指示信息确定执行模型任务时所要使用的数据类型。
其中,图6a对应的实施例的方法(即步骤601a-步骤604a)可以是应用于“参与执行模型任务的多个设备中包括网络设备(即参与执行模型任务的多个设备中同时包括有终端设备和网络设备)”这一场景。
关于实施例其它内容的详细介绍可以参考上述实施例中的相关介绍,本公开实施例在此不做赘述。
综上所述,在本公开实施例提供的数据类型确定方法之中,终端设备会确定执行模型任务时所要使用的数据类型,其中,终端设备确定的数据类型与多个设备中除终端设备外的其他设备确定的数据类型相同。也即是,本公开中共同参与执行模型任务的多个设备所确定的执行模型任务时所要使用的数据类型均相同,由此该多个设备可以基于相同的数据类型来执行模型任务,则执行模型任务期间无需对多个设备的使用数据进行类型转换,从而确保了模型任务的执行精度,进而确保了基于模型任务所得到的AI模型的精度,提高了AI模型的用户体验度。
图6b为本公开实施例所提供的一种数据类型确定方法的流程示意图,多个设备共同参与执行模型任务,所述方法由多个设备中的终端设备执行,响应于多个设备中包括网络设备,如图6b所示,该数据类型确定方法可以包括以下步骤:
步骤601b、向网络设备发送第一能力信息。
关于第一能力信息的详细介绍可以参考上述实施例中的相关介绍,本公开实施例在此不做赘述。
步骤602b、接收网络设备发送的指示信息。
其中,在本公开的一个实施例之中,指示信息用于指示执行模型任务时所要使用的数据类型。该指示信息指示的数据类型具体是网络设备基于各个终端设备发送的第一能力信息以及各个终端设备支持的数据类型确定的。其中,该各个终端设备支持的数据类型具体可以是网络设备从核心网处获取的。
以及,在本公开的一个实施例之中,各个终端设备接收到的指示信息应当是相同,从而确保多个设备确定的执行模型任务时所要使用的数据类型是相同统一的,则确保了后续执行模型任务时的精度。
步骤603b、基于指示信息确定执行模型任务时所要使用的数据类型。
其中,图6b对应的实施例的方法(即步骤601b-步骤603b)可以是应用于“参与执行模型任务的多个设备中包括网络设备(即参与执行模型任务的多个设备中同时包括有终端设备和网络设备)”这一场景。
关于实施例其它内容的详细介绍可以参考上述实施例中的相关介绍,本公开实施例在此不做赘述。
综上所述,在本公开实施例提供的数据类型确定方法之中,终端设备会确定执行模型任务时所要使用的数据类型,其中,终端设备确定的数据类型与多个设备中除终端设备外的其他设备确定的数据类型相同。也即是,本公开中共同参与执行模型任务的多个设备所确定的执行模型任务时所要使用的数据类型均相同,由此该多个设备可以基于相同的数据类型来执行模型任务,则执行模型任务期间无需对多个设备的使用数据进行类型转换,从而确保了模型任务的执行精度,进而确保了基于模型任务所得到的AI模型的精度,提高了AI模型的用户体验度。
图7为本公开实施例所提供的一种数据类型确定方法的流程示意图,多个设备共同参与执行模型任务,所述方法由多个设备中的终端设备执行,如图7所示,该数据类型确定方法可以包括以下步骤:
步骤701、确定执行模型任务时所要使用的数据类型。
步骤702、基于确定的数据类型执行模型任务。
其中,在本公开的一个实施例之中,由于终端设备确定的数据类型与其他设备确定的数据类型相同,由此当多个设备共同参与执行模型任务时会基于相同统一的数据类型来执行模型任务,则确保了执行精度。
关于本实施例中其它内容的详细介绍可以参考上述实施例中的相关介绍,本公开实施例在此不做赘述。
综上所述,在本公开实施例提供的数据类型确定方法之中,终端设备会确定执行模型任务时所要使用的数据类型,其中,终端设备确定的数据类型与多个设备中除终端设备外的其他设备确定的数据类型相同。也即是,本公开中共同参与执行模型任务的多个设备所确定的执行模型任务时所要使用的数据类型均相同,由此该多个设备可以基于相同的数据类型来执行模型任务,则执行模型任务期间无需对多个设备的使用数据进行类型转换,从而确保了模型任务的执行精度,进而确保了基于模型任务所得到的AI模型的精度,提高了AI模型的用户体验度。
图8为本公开实施例所提供的一种数据类型确定方法的流程示意图,该方法由网络设备执行,网络设备与其他设备共同参与执行模型任务,如图8所示,该数据类型确定方法可以包括以下步骤:
步骤801、确定执行模型任务的各个设备执行模型任务时所要使用的数据类型。
其中,在本公开的一个实施例之中,上述模型任务可以包括模型训练任务,和/或,模型推理任务。以及,上述的“多个设备共同参与执行模型任务”可以理解为:由多个设备来共同执行一模型任务(如模型训练任务或模型推理任务),其中,不同设备用于负责模型训练任务中的至少一部分,各个设备之间通过交互来实现该模型任务。
以及,上述的数据类型可以包括整型(int)、单浮点数(即16位浮点数(float))、双浮点数(即32位浮点数(float))中的至少一种。
需要说明的是,在本公开的一个实施例之中,网络设备针对各个设备确定的执行模型任务时使用的数据类型相同,由此可以确保多个设备采用相同的数据类型来执行模型任务,进而保证了多个设备在执行模型任务时的数据类型的统一,确保了执行模型任务的精度。
图8对应的实施例的方法(即步骤801)应用于“参与执行模型任务的多个设备中包括网络设备(即参与执行模型任务的多个设备中同时包括有终端设备和网络设备)”这一场景。
以及,关于步骤801的具体执行方法会在后续实施例中详细介绍。
关于本实施例中其它内容的详细介绍可以参考上述实施例中的相关介绍,本公开的实施例在此不做赘述。
综上所述,在本公开实施例提供的数据类型确定方法之中,终端设备会确定执行模型任务时所要使用的数据类型,其中,终端设备确定的数据类型与多个设备中除终端设备外的其他设备确定的数据类型相同。也即是,本公开中共同参与执行模型任务的多个设备所确定的执行模型任务时所要使用的数据类型均相同,由此该多个设备可以基于相同的数据类型来执行模型任务,则执行模型任务期间无需对多个设备的使用数据进行类型转换,从而确保了模型任务的执行精度,进而确保了基于模型任务所得到的AI模型的精度,提高了AI模型的用户体验度。
图9为本公开实施例所提供的一种数据类型确定方法的流程示意图,该方法由网络设备执行,网络设备与其他设备共同参与执行模型任务,如图9所示,该数据类型确定方法可以包括以下步骤:
步骤901、确定其他设备支持的数据类型。
其中,在本公开的一个实施例之中,确定其他设备支持的数据类型的方法可以包括以下任意一种:
接收其他设备发送的数据支持能力信息,数据支持能力信息用于指示其他设备支持的数据类型;
从核心网设备处确定其他设备支持的数据类型,其中,核心网设备处可以存储有各个设备预先上报的其支持的数据类型,该各个设备支持的数据类型具体是该设备在接入核心网设备时上报至核心网设备的。
步骤902、确定网络设备支持的数据类型。
步骤903、基于网络设备支持的数据类型和其他设备支持的数据类型确定网络设备和其他设备共同支持的至少一种数据类型。
关于步骤903的详细介绍可以参考上述实施例中的相关介绍,本公开实施例在此不做赘述。
步骤904、从网络设备和其他设备共同支持的至少一种数据类型中确定执行模型任务时所要使用的数据类型。
其中,在本公开的一个实施例之中,网络设备和其他设备共同支持的至少一种数据类型中确定执行模型任务时所要使用的数据类型的方法可以包括以下至少一种:
将共同支持的至少一种数据类型中精度最大的第一数据类型确定为执行模型任务时所要使用的数据类型;
将共同支持的至少一种数据类型中精度小于第一数据类型的第二数据类型确定为执行模型任务时所要使用的数据类型;
基于模型任务的期望资源开销从共同支持的至少一种数据类型中确定执行模型任务时所要使用的数据类型;
基于模型任务的业务需求从共同支持的至少一种数据类型中确定执行模型任务时所要使用的数据类型。
关于步骤904的详细介绍可以参考上述步骤304的相关介绍,本公开实施例在此不做赘述。
关于本实施例中其它内容的详细介绍可以参考上述实施例中的相关介绍,本公开的实施例在此不做赘述。
综上所述,在本公开实施例提供的数据类型确定方法之中,终端设备会确定执行模型任务时所要使用的数据类型,其中,终端设备确定的数据类型与多个设备中除终端设备外的其他设备确定的数据类型相同。也即是,本公开中共同参与执行模型任务的多个设备所确定的执行模型任务时所要使用的数据类型均相同,由此该多个设备可以基于相同的数据类型来执行模型任务,则执行模型任务期间无需对多个设备的使用数据进行类型转换,从而确保了模型任务的执行精度,进而确保了基于模型任务所得到的AI模型的精度,提高了AI模型的用户体验度。
图10为本公开实施例所提供的一种数据类型确定方法的流程示意图,该方法由网络设备执行,网络设备与其他设备共同参与执行模型任务,如图10所示,该数据类型确定方法可以包括以下步骤:
步骤1001、确定其他设备支持的数据类型。
其中,关于步骤1001的详细介绍可以参考上述实施例描述,本公开实施例在此不做赘述。
步骤1002、接收其他设备发送的能力信息。
其中,在本公开的一个实施例之中,能力信息可以包括以下至少一种:
其他设备的存储能力;
其他设备的功耗能力。
关于能力信息的详细介绍可以参考上述实施例中的相关介绍,本公开实施例在此不做赘述。
步骤1003、确定网络设备的能力。
其中,在本公开的一个实施例之中,能力可以包括存储能力和/或功耗能力。
关于存储能力和功耗能力的详细介绍可以参考上述实施例中的相关介绍,本公开的实施例在此不做赘述。
步骤1004、基于网络设备支持的数据类型和其他设备支持的数据类型确定网络设备和其他设备共同支持的至少一种数据类型。
关于步骤1004的详细介绍可以参考上述实施例中的相关介绍,本公开实施例在此不做赘述。
步骤1005、基于第一能力信息和网络设备的能力从共同支持的至少一种数据类型中确定执行模型任务时所要使用的数据类型。
其中,在本公开的一个实施例之中,基于第一能力信息和网络设备的能力从共同支持的至少一种数据类型中确定执行模型任务时所要使用的数据类型的方法可以为:从共同支持的至少一种数据类型中选择出精度匹配于第一能力信息所指示的能力和/或网络设备的能力的第三数据类型,将第三数据类型确定为执行模型任务时所要使用的数据类型。
关于步骤1005的详细介绍可以参考上述实施例中的相关介绍,本公开实施例在此不做赘述。
关于本实施例中其它内容的详细介绍可以参考上述实施例中的相关介绍,本公开的实施例在此不做赘述。
综上所述,在本公开实施例提供的数据类型确定方法之中,终端设备会确定执行模型任务时所要使用的数据类型,其中,终端设备确定的数据类型与多个设备中除终端设备外的其他设备确定的数据类型相同。也即是,本公开中共同参与执行模型任务的多个设备所确定的执行模型任务时所要使用的数据类型均相同,由此该多个设备可以基于相同的数据类型来执行模型任务,则执行模型任务期间无需对多个设备的使用数据进行类型转换,从而确保了模型任务的执行精度,进而确保了基于模型任务所得到的AI模型的精度,提高了AI模型的用户体验度。
图11a为本公开实施例所提供的一种数据类型确定方法的流程示意图,该方法由网络设备执行,网络设备与其他设备共同参与执行模型任务,如图11a所示,该数据类型确定方法可以包括以下步骤:
步骤1101a、确定执行模型任务的各个设备执行模型任务时所要使用的数据类型。
其中,关于步骤1101a的详细介绍可以参考上述实施例描述,本公开实施例在此不做赘述。
步骤1102a、向其他设备发送指示信息。
其中,在本公开的一个实施例之中,指示信息用于指示执行模型任务时所要使用的数据类型。
以及,在本公开的一个实施例之中,网络设备会向各个其他设备发送相同的指示信息,由此各个其他设备即可基于相同的指示信息确定出相同的数据类型来执行模型任务,则确保了各个其他设备在执行模型任务时的数据类型的统一,确保了执行模型任务的精度。
需要说明的是,当参与执行模型任务的多个设备中包括网络设备时,则应当是由网络设备来向各个其他设备分配具体的模型任务的。以及,在本公开的一个实施例之中,网络设备可以是在向各个其他设备分配模型任务之前就确定好了确定执行模型任务时所要使用的数据类型,之后,在向各个其他设备发送指示信息以指示确定执行模型任务时所要使用的数据类型时,还可以同步向各个其他设备分配模型任务。
综上所述,在本公开实施例提供的数据类型确定方法之中,终端设备会确定执行模型任务时所要使用的数据类型,其中,终端设备确定的数据类型与多个设备中除终端设备外的其他设备确定的数据类型相同。也即是,本公开中共同参与执行模型任务的多个设备所确定的执行模型任务时所要使用的数据类 型均相同,由此该多个设备可以基于相同的数据类型来执行模型任务,则执行模型任务期间无需对多个设备的使用数据进行类型转换,从而确保了模型任务的执行精度,进而确保了基于模型任务所得到的AI模型的精度,提高了AI模型的用户体验度。
图11b为本公开实施例所提供的一种数据类型确定方法的流程示意图,该方法由网络设备执行,网络设备与其他设备共同参与执行模型任务,如图11b所示,该数据类型确定方法可以包括以下步骤:
步骤1101b、确定执行模型任务的各个设备执行模型任务时所要使用的数据类型。
步骤1102b、基于确定的数据类型执行模型任务。
其中,关于步骤1101b-步骤1102b的详细介绍可以参考上述实施例描述,本公开实施例在此不做赘述。
综上所述,在本公开实施例提供的数据类型确定方法之中,终端设备会确定执行模型任务时所要使用的数据类型,其中,终端设备确定的数据类型与多个设备中除终端设备外的其他设备确定的数据类型相同。也即是,本公开中共同参与执行模型任务的多个设备所确定的执行模型任务时所要使用的数据类型均相同,由此该多个设备可以基于相同的数据类型来执行模型任务,则执行模型任务期间无需对多个设备的使用数据进行类型转换,从而确保了模型任务的执行精度,进而确保了基于模型任务所得到的AI模型的精度,提高了AI模型的用户体验度。
图12为本公开实施例所提供的一种通信装置的结构示意图,如图12所示,装置可以包括:
确定模块1201,用于确定执行模型任务时所要使用的数据类型,其中,终端设备确定的数据类型与多个设备中的其他设备确定的数据类型相同。
综上所述,在本公开实施例提供的通信装置之中,终端设备会确定执行模型任务时所要使用的数据类型,其中,终端设备确定的数据类型与多个设备中除终端设备外的其他设备确定的数据类型相同。也即是,本公开中共同参与执行模型任务的多个设备所确定的执行模型任务时所要使用的数据类型均相同,由此该多个设备可以基于相同的数据类型来执行模型任务,则执行模型任务期间无需对多个设备的使用数据进行类型转换,从而确保了模型任务的执行精度,进而确保了基于模型任务所得到的AI模型的精度,提高了AI模型的用户体验度。
可选的,在本公开的一个实施例之中,上述装置还用于:
分别向所述其他设备发送第一数据支持能力信息,所述第一数据支持能力信息用于指示所述终端设备支持的数据类型。
可选的,在本公开的一个实施例之中,上述装置还用于:
接收所述其他设备发送的第二数据支持能力信息,所述第二数据支持能力信息用于指示所述其他设备支持的数据类型。
响应于多个设备中不包括网络设备,上述确定模块1201,还用于:
基于第一数据支持能力信息和第二数据支持能力信息确定终端设备和其他设备共同支持的至少一种数据类型;
从所述终端设备和其他设备共同支持的至少一种数据类型中确定执行所述模型任务时所要使用的数据类型,其中,所述终端设备与所述其他设备采用相同的方法从共同支持的至少一种数据类型中确定执行所述模型任务时所要使用的数据类型。
可选的,在本公开的一个实施例之中,上述确定模块1201,还用于:
将共同支持的至少一种数据类型中精度最大的第一数据类型确定为执行模型任务时所要使用的数据类型;
将共同支持的至少一种数据类型中精度小于第一数据类型的第二数据类型确定为执行模型任务时所要使用的数据类型;
基于模型任务的期望资源开销从共同支持的至少一种数据类型中确定执行模型任务时所要使用的数据类型;
基于模型任务的业务需求从共同支持的至少一种数据类型中确定执行模型任务时所要使用的数据类型。
可选的,在本公开的一个实施例之中,上述装置还用于:
分别向其他设备发送第一能力信息,所述第一能力信息用于指示以下至少一种:
所述终端设备的存储能力;
所述终端设备的功耗能力。
可选的,在本公开的一个实施例之中,上述装置还用于:
接收其他设备发送的第二数据支持能力信息和第二能力信息,第二数据支持能力信息用于指示其他设备支持的数据类型,第二能力信息用于指示以下至少一种:
其他设备的存储能力;
其他设备的的功耗能力;
可选的,在本公开的一个实施例之中,上述装置还用于:
基于所述第一数据支持能力信息和第二数据支持能力信息确定所述终端设备和其他设备共同支持的至少一种数据类型;
基于所述第一能力信息和所述第二能力信息从所述共同支持的至少一种数据类型中确定执行所述模型任务时所要使用的数据类型。
可选的,在本公开的一个实施例之中,上述确定模块1201,还用于:
从共同支持的至少一种数据类型中选择精度匹配于所述第一能力信息和/或所述第二能力信息所指示的能力的第三数据类型;其中,第三数据类型的精度与所述存储能力和所述功耗能力均呈正相关;
将第三数据类型确定为执行模型任务时所要使用的数据类型。
可选的,在本公开的一个实施例之中,上述装置还用于:
向所述网络设备发送第一数据支持能力信息和/或第一能力信息。
可选的,在本公开的一个实施例之中,上述确定模块1202,还用于:
接收网络设备发送的指示信息,指示信息用于指示执行模型任务时所要使用的数据类型;
基于指示信息确定执行模型任务时所要使用的数据类型。
可选的,在本公开的一个实施例之中,上述装置还用于:基于确定的数据类型执行模型任务。
可选的,在本公开的一个实施例之中,模型任务包括:模型训练任务,和/或,模型推理任务。
图13为本公开实施例所提供的一种通信装置的结构示意图,如图13所示,装置可以包括:
确定模块1301,用于确定执行模型任务的各个设备执行模型任务时所要使用的数据类型,其中,网络设备针对各个设备确定的执行模型任务时使用的数据类型相同。
综上所述,在本公开实施例提供的通信装置之中,终端设备会确定执行模型任务时所要使用的数据类型,其中,终端设备确定的数据类型与多个设备中除终端设备外的其他设备确定的数据类型相同。也即是,本公开中共同参与执行模型任务的多个设备所确定的执行模型任务时所要使用的数据类型均相同,由此该多个设备可以基于相同的数据类型来执行模型任务,则执行模型任务期间无需对多个设备的使用数据进行类型转换,从而确保了模型任务的执行精度,进而确保了基于模型任务所得到的AI模型的精度,提高了AI模型的用户体验度。
可选的,在本公开的一个实施例之中,上述确定模块1301,还用于:
确定所述多个设备中除所述网络设备外的所述其他设备支持的数据类型。
可选的,在本公开的一个实施例之中,上述确定模块1301,还用于:
接收其他设备发送的数据支持能力信息,所述数据支持能力信息用于指示所述其他设备支持的数据类型;
从核心网设备处确定所述其他设备支持的数据类型。
可选的,在本公开的一个实施例之中,上述确定模块1301,还用于:
确定所述网络设备支持的数据类型;
基于所述网络设备支持的数据类型和其他设备支持的数据类型确定所述网络设备和所述其他设备共同支持的至少一种数据类型;
从所述网络设备和其他设备共同支持的至少一种数据类型中确定执行所述模型任务时所要使用的 数据类型。
可选的,在本公开的一个实施例之中,上述确定模块1301,还用于:
将所述共同支持的至少一种数据类型中精度最大的第一数据类型确定为执行所述模型任务时所要使用的数据类型;
将所述共同支持的至少一种数据类型中精度小于所述第一数据类型的第二数据类型确定为执行所述模型任务时所要使用的数据类型;
基于所述模型任务的期望资源开销从所述共同支持的至少一种数据类型中确定执行所述模型任务时所要使用的数据类型;
基于所述模型任务的业务需求从所述共同支持的至少一种数据类型中确定执行所述模型任务时所要使用的数据类型。
可选的,在本公开的一个实施例之中,所述装置还用于:
接收其他设备发送的能力信息,能力信息包括以下至少一种:
其他设备的存储能力;
其他设备的功耗能力。
可选的,在本公开的一个实施例之中,上述确定模块1301,还用于:
确定所述网络设备的能力,所述能力包括存储能力和/或功耗能力;
基于所述网络设备支持的数据类型和其他设备支持的数据类型确定所述网络设备和所述其他设备共同支持的至少一种数据类型;
基于所述第一能力信息和所述网络设备的能力从所述共同支持的至少一种数据类型中确定执行所述模型任务时所要使用的数据类型。
可选的,在本公开的一个实施例之中,上述确定模块1301,还用于:
从所述共同支持的至少一种数据类型中选择出精度匹配于所述第一能力信息所指示的能力和/或所述网络设备的能力的第三数据类型;其中,所述第三数据类型的精度与所述存储能力和所述功耗能力均呈正相关;
将所述第三数据类型确定为执行所述模型任务时所要使用的数据类型。
可选的,在本公开的一个实施例之中,所述装置还用于:
向所述多个设备中除所述网络设备外的其他设备发送指示信息,所述指示信息用于指示所述其他设备执行模型任务时所要使用的数据类型。
可选的,在本公开的一个实施例之中,所述装置还用于:
基于确定的数据类型执行模型任务。
可选的,在本公开的一个实施例之中,模型任务包括:模型训练任务,和/或,模型推理任务。
请参见图14,图14是本申请实施例提供的一种通信装置1400的结构示意图。通信装置1400可以是网络设备,也可以是终端设备,也可以是支持网络设备实现上述方法的芯片、芯片系统、或处理器等,还可以是支持终端设备实现上述方法的芯片、芯片系统、或处理器等。该装置可用于实现上述方法实施例中描述的方法,具体可以参见上述方法实施例中的说明。
通信装置1400可以包括一个或多个处理器1401。处理器1401可以是通用处理器或者专用处理器等。例如可以是基带处理器或中央处理器。基带处理器可以用于对通信协议以及通信数据进行处理,中央处理器可以用于对通信装置(如,基站、基带芯片,终端设备、终端设备芯片,DU或CU等)进行控制,执行计算机程序,处理计算机程序的数据。
可选的,通信装置1400中还可以包括一个或多个存储器1402,其上可以存有计算机程序1404,处理器1401执行所述计算机程序1404,以使得通信装置1400执行上述方法实施例中描述的方法。可选的,所述存储器1402中还可以存储有数据。通信装置1400和存储器1402可以单独设置,也可以集成在一起。
可选的,通信装置1400还可以包括收发器1405、天线1406。收发器1405可以称为收发单元、收发机、或收发电路等,用于实现收发功能。收发器1405可以包括接收器和发送器,接收器可以称为接 收机或接收电路等,用于实现接收功能;发送器可以称为发送机或发送电路等,用于实现发送功能。
可选的,通信装置1400中还可以包括一个或多个接口电路1407。接口电路1407用于接收代码指令并传输至处理器1401。处理器1401运行所述代码指令以使通信装置1400执行上述方法实施例中描述的方法。
通信装置1400为终端设备:收发器1405用于执行图3中的步骤301-步骤302;图4中的步骤401至步骤402;图5中的步骤501至步骤503;图6a中的步骤601a至步骤603a;图6b中的步骤601b至步骤602b。处理器1401用于执行图2中的步骤201;图3中的步骤303-步骤304;图4中的步骤403;图5中的步骤504和步骤505;图6a中的步骤604a;图6b中的步骤603b;图7中的步骤701-步骤702。
通信装置1400为网络设备:收发器1405用于执行图11a中的步骤1102a。处理器1401用于执行图8中的步骤801;图9中的步骤901至步骤904;图10中的步骤1001至步骤1005;图11a中的步骤1101a;图11b中的步骤1101b和步骤1102b。
在一种实现方式中,处理器1401中可以包括用于实现接收和发送功能的收发器。例如该收发器可以是收发电路,或者是接口,或者是接口电路。用于实现接收和发送功能的收发电路、接口或接口电路可以是分开的,也可以集成在一起。上述收发电路、接口或接口电路可以用于代码/数据的读写,或者,上述收发电路、接口或接口电路可以用于信号的传输或传递。
在一种实现方式中,处理器1401可以存有计算机程序1403,计算机程序1403在处理器1401上运行,可使得通信装置1400执行上述方法实施例中描述的方法。计算机程序1403可能固化在处理器1401中,该种情况下,处理器1401可能由硬件实现。
在一种实现方式中,通信装置1400可以包括电路,所述电路可以实现前述方法实施例中发送或接收或者通信的功能。本申请中描述的处理器和收发器可实现在集成电路(integrated circuit,IC)、模拟IC、射频集成电路RFIC、混合信号IC、专用集成电路(application specific integrated circuit,ASIC)、印刷电路板(printed circuit board,PCB)、电子设备等上。该处理器和收发器也可以用各种IC工艺技术来制造,例如互补金属氧化物半导体(complementary metal oxide semiconductor,CMOS)、N型金属氧化物半导体(nMetal-oxide-semiconductor,NMOS)、P型金属氧化物半导体(positive channel metal oxide semiconductor,PMOS)、双极结型晶体管(bipolar junction transistor,BJT)、双极CMOS(BiCMOS)、硅锗(SiGe)、砷化镓(GaAs)等。
以上实施例描述中的通信装置可以是网络设备或者终端设备,但本申请中描述的通信装置的范围并不限于此,而且通信装置的结构可以不受图14的限制。通信装置可以是独立的设备或者可以是较大设备的一部分。例如所述通信装置可以是:
(1)独立的集成电路IC,或芯片,或,芯片系统或子系统;
(2)具有一个或多个IC的集合,可选的,该IC集合也可以包括用于存储数据,计算机程序的存储部件;
(3)ASIC,例如调制解调器(Modem);
(4)可嵌入在其他设备内的模块;
(5)接收机、终端设备、智能终端设备、蜂窝电话、无线设备、手持机、移动单元、车载设备、网络设备、云设备、人工智能设备等等;
(6)其他等等。
对于通信装置可以是芯片或芯片系统的情况,可参见图15所示的芯片的结构示意图。图15所示的芯片包括处理器1501和接口1502。其中,处理器1501的数量可以是一个或多个,接口1502的数量可以是多个。
可选的,芯片还包括存储器1503,存储器1503用于存储必要的计算机程序和数据。
本领域技术人员还可以了解到本申请实施例列出的各种说明性逻辑块(illustrative logical block)和步骤(step)可以通过电子硬件、电脑软件,或两者的结合进行实现。这样的功能是通过硬件还是软件来实现取决于特定的应用和整个系统的设计要求。本领域技术人员可以对于每种特定的应用,可以使用各种方法实现所述的功能,但这种实现不应被理解为超出本申请实施例保护的范围。
本申请还提供一种可读存储介质,其上存储有指令,该指令被计算机执行时实现上述任一方法实施例的功能。
本申请还提供一种计算机程序产品,该计算机程序产品被计算机执行时实现上述任一方法实施例的功能。
在上述实施例中,可以全部或部分地通过软件、硬件、固件或者其任意组合来实现。当使用软件实现时,可以全部或部分地以计算机程序产品的形式实现。所述计算机程序产品包括一个或多个计算机程序。在计算机上加载和执行所述计算机程序时,全部或部分地产生按照本申请实施例所述的流程或功能。所述计算机可以是通用计算机、专用计算机、计算机网络、或者其他可编程装置。所述计算机程序可以存储在计算机可读存储介质中,或者从一个计算机可读存储介质向另一个计算机可读存储介质传输,例如,所述计算机程序可以从一个网站站点、计算机、服务器或数据中心通过有线(例如同轴电缆、光纤、数字用户线(digital subscriber line,DSL))或无线(例如红外、无线、微波等)方式向另一个网站站点、计算机、服务器或数据中心进行传输。所述计算机可读存储介质可以是计算机能够存取的任何可用介质或者是包含一个或多个可用介质集成的服务器、数据中心等数据存储设备。所述可用介质可以是磁性介质(例如,软盘、硬盘、磁带)、光介质(例如,高密度数字视频光盘(digital video disc,DVD))、或者半导体介质(例如,固态硬盘(solid state disk,SSD))等。
本领域普通技术人员可以理解:本申请中涉及的第一、第二等各种数字编号仅为描述方便进行的区分,并不用来限制本申请实施例的范围,也表示先后顺序。
本申请中的至少一个还可以描述为一个或多个,多个可以是两个、三个、四个或者更多个,本申请不做限制。在本申请实施例中,对于一种技术特征,通过“第一”、“第二”、“第三”、“A”、“B”、“C”和“D”等区分该种技术特征中的技术特征,该“第一”、“第二”、“第三”、“A”、“B”、“C”和“D”描述的技术特征间无先后顺序或者大小顺序。
本申请中各表所示的对应关系可以被配置,也可以是预定义的。各表中的信息的取值仅仅是举例,可以配置为其他值,本申请并不限定。在配置信息与各参数的对应关系时,并不一定要求必须配置各表中示意出的所有对应关系。例如,本申请中的表格中,某些行示出的对应关系也可以不配置。又例如,可以基于上述表格做适当的变形调整,例如,拆分,合并等等。上述各表中标题示出参数的名称也可以采用通信装置可理解的其他名称,其参数的取值或表示方式也可以通信装置可理解的其他取值或表示方式。上述各表在实现时,也可以采用其他的数据结构,例如可以采用数组、队列、容器、栈、线性表、指针、链表、树、图、结构体、类、堆、散列表或哈希表等。
本申请中的预定义可以理解为定义、预先定义、存储、预存储、预协商、预配置、固化、或预烧制。
本领域普通技术人员可以意识到,结合本文中所公开的实施例描述的各示例的单元及算法步骤,能够以电子硬件、或者计算机软件和电子硬件的结合来实现。这些功能究竟以硬件还是软件方式来执行,取决于技术方案的特定应用和设计约束条件。专业技术人员可以对每个特定的应用来使用不同方法来实现所描述的功能,但是这种实现不应认为超出本申请的范围。
所属领域的技术人员可以清楚地了解到,为描述的方便和简洁,上述描述的系统、装置和单元的具体工作过程,可以参考前述方法实施例中的对应过程,在此不再赘述。
以上所述,仅为本申请的具体实施方式,但本申请的保护范围并不局限于此,任何熟悉本技术领域的技术人员在本申请揭露的技术范围内,可轻易想到变化或替换,都应涵盖在本申请的保护范围之内。因此,本申请的保护范围应以所述权利要求的保护范围为准。
Claims (29)
- 一种数据类型确定方法,其特征在于,多个设备共同参与执行模型任务,所述方法由所述多个设备中的终端设备执行,包括:确定执行所述模型任务时所要使用的数据类型,其中,所述终端设备确定的数据类型与所述多个设备中除所述终端设备外的其他设备确定的数据类型相同。
- 如权利要求1所述的方法,其特征在于,所述方法还包括:分别向所述其他设备发送第一数据支持能力信息,所述第一数据支持能力信息用于指示所述终端设备支持的数据类型。
- 如权利要求2所述的方法,其特征在于,所述方法还包括:接收所述其他设备发送的第二数据支持能力信息,所述第二数据支持能力信息用于指示所述其他设备支持的数据类型。
- 如权利要求3所述的方法,其特征在于,所述确定执行所述模型任务时所要使用的数据类型,包括:基于所述第一数据支持能力信息和第二数据支持能力信息确定所述终端设备和其他设备共同支持的至少一种数据类型;从所述终端设备和其他设备共同支持的至少一种数据类型中确定执行所述模型任务时所要使用的数据类型。
- 如权利要求4所述的方法,其特征在于,所述从所述终端设备和其他设备共同支持的至少一种数据类型中确定执行所述模型任务时所要使用的数据类型,包括以下至少一种:将所述共同支持的至少一种数据类型中精度最大的第一数据类型确定为执行所述模型任务时所要使用的数据类型;将所述共同支持的至少一种数据类型中精度小于所述第一数据类型的第二数据类型确定为执行所述模型任务时所要使用的数据类型;基于所述模型任务的期望资源开销从所述共同支持的至少一种数据类型中确定执行所述模型任务时所要使用的数据类型;基于所述模型任务的业务需求从所述共同支持的至少一种数据类型中确定执行所述模型任务时所要使用的数据类型。
- 如权利要求2所述的方法,其特征在于,所述方法还包括:分别向所述其他设备发送第一能力信息,所述第一能力信息用于指示以下至少一种:所述终端设备的存储能力;所述终端设备的功耗能力。
- 如权利要求6所述的方法,其特征在于,所述方法还包括:接收所述其他设备发送的第二数据支持能力信息和第二能力信息,所述第二数据支持能力信息用于指示所述其他设备支持的数据类型,所述第二能力信息用于指示以下至少一种:所述其他设备的存储能力;所述其他设备的功耗能力。
- 如权利要求7所述的方法,其特征在于,所述确定执行所述模型任务时所要使用的数据类型,包括:基于所述第一数据支持能力信息和第二数据支持能力信息确定所述终端设备和其他设备共同支持的至少一种数据类型;基于所述第一能力信息和所述第二能力信息从所述共同支持的至少一种数据类型中确定执行所述模型任务时所要使用的数据类型。
- 如权利要求8所述的方法,其特征在于,所述基于所述第一能力信息和所述第二能力信息从所述共同支持的至少一种数据类型中确定执行所述模型任务时所要使用的数据类型,包括:从所述共同支持的至少一种数据类型中选择精度匹配于所述第一能力信息和/或所述第二能力信息所指示的能力的第三数据类型;其中,所述第三数据类型的精度与所述存储能力和所述功耗能力均呈正相关;将所述第三数据类型确定为执行所述模型任务时所要使用的数据类型。
- 如权利要求1所述的方法,其特征在于,所述方法还包括:向所述网络设备发送第一数据支持能力信息和/或第一能力信息。
- 如权利要求10所述的方法,其特征在于,所述确定执行所述模型任务时所要使用的数据类型,包括:接收网络设备发送的指示信息,所述指示信息用于指示执行所述模型任务时所要使用的数据类型;基于所述指示信息确定执行所述模型任务时所要使用的数据类型。
- 如权利要求1所述的方法,其特征在于,所述方法还包括:基于确定的所述数据类型执行所述模型任务。
- 如权利要求1-12任一所述的方法,其特征在于,所述模型任务包括:模型训练任务,和/或,模型推理任务。
- 一种数据类型确定方法,其特征在于,多个设备共同参与执行模型任务,所述方法由所述多个设备中的网络设备执行,包括:确定执行所述模型任务的各个设备执行所述模型任务时所要使用的数据类型,其中,所述网络设备针对所述各个设备确定的执行所述模型任务时使用的数据类型相同。
- 如权利要求14所述的方法,其特征在于,所述方法还包括:确定所述多个设备中除所述网络设备外的其他设备支持的数据类型。
- 如权利要求15所述的方法,其特征在于,所述确定所述多个设备中除所述网络设备外的其他设备支持的数据类型,包括以下任意一种:接收其他设备发送的数据支持能力信息,所述数据支持能力信息用于指示所述其他设备支持的数据类型;从核心网设备处确定所述其他设备支持的数据类型。
- 如权利要求16所述的方法,其特征在于,所述确定执行所述模型任务的各个设备执行所述模型任务时所要使用的数据类型,包括:确定所述网络设备支持的数据类型;基于所述网络设备支持的数据类型和其他设备支持的数据类型确定所述网络设备和所述其他设备共同支持的至少一种数据类型;从所述网络设备和其他设备共同支持的至少一种数据类型中确定执行所述模型任务时所要使用的数据类型。
- 如权利要求17所述的方法,其特征在于,从所述网络设备和其他设备共同支持的至少一种数据类型中确定执行所述模型任务时所要使用的数据类型,包括以下至少一种:将所述共同支持的至少一种数据类型中精度最大的第一数据类型确定为执行所述模型任务时所要使用的数据类型;将所述共同支持的至少一种数据类型中精度小于所述第一数据类型的第二数据类型确定为执行所述模型任务时所要使用的数据类型;基于所述模型任务的期望资源开销从所述共同支持的至少一种数据类型中确定执行所述模型任务时所要使用的数据类型;基于所述模型任务的业务需求从所述共同支持的至少一种数据类型中确定执行所述模型任务时所要使用的数据类型。
- 如权利要求15所述的方法,其特征在于,所述方法还包括:接收所述其他设备发送的能力信息,所述能力信息包括以下至少一种:所述其他设备的存储能力;所述其他设备的功耗能力。
- 如权利要求19所述的方法,其特征在于,所述确定执行所述模型任务的各个设备执行所述模型任务时所要使用的数据类型,包括:确定所述网络设备的能力,所述网络设备的能力包括存储能力和/或功耗能力;基于所述网络设备支持的数据类型和其他设备支持的数据类型确定所述网络设备和所述其他设备共同支持的至少一种数据类型;基于所述能力信息和所述网络设备的能力从所述共同支持的至少一种数据类型中确定执行所述模型任务时所要使用的数据类型。
- 如权利要求20所述的方法,其特征在于,所述基于所述能力信息和所述网络设备的能力从所述共同支持的至少一种数据类型中确定执行所述模型任务时所要使用的数据类型,包括:从所述共同支持的至少一种数据类型中选择出精度匹配于所述能力信息所指示的能力和/或所述网络设备的能力的第三数据类型;其中,所述第三数据类型的精度与所述存储能力和所述功耗能力均呈正相关;将所述第三数据类型确定为执行所述模型任务时所要使用的数据类型。
- 如权利要求14-21任一所述的方法,其特征在于,所述方法还包括:向所述多个设备中除所述网络设备外的其他设备发送指示信息,所述指示信息用于指示执行模型任务时所要使用的数据类型。
- 如权利要求14所述的方法,其特征在于,所述方法还包括:基于确定的所述数据类型执行所述模型任务。
- 如权利要求14-23任一所述的方法,其特征在于,所述模型任务包括:模型训练任务,和/或,模型推理任务。
- 一种通信装置,被配置在终端设备中,包括:确定模块,用于确定执行所述模型任务时所要使用的数据类型,其中,所述终端设备确定的数据类型与所述多个设备中的其他设备确定的数据类型相同。
- 一种通信装置,被配置在网络设备中,包括:确定模块,用于确定执行所述模型任务的各个设备执行所述模型任务时所要使用的数据类型,其中,所述网络设备针对所述各个设备确定的执行所述模型任务时使用的数据类型相同。
- 一种通信装置,其特征在于,所述装置包括处理器和存储器,其中,所述存储器中存储有计算机程序,所述处理器执行所述存储器中存储的计算机程序,以使所述装置执行如权利要求1至13中任一项所述的方法,或所述处理器执行所述存储器中存储的计算机程序,以使所述装置执行如权利要求14至24中任一项所述的方法。
- 一种通信装置,其特征在于,包括:处理器和接口电路,其中所述接口电路,用于接收代码指令并传输至所述处理器;所述处理器,用于运行所述代码指令以执行如权利要求1至13中任一项所述的方法,或用于运行所述代码指令以执行如权利要求14至24中任一项所述的方法。
- 一种计算机可读存储介质,用于存储有指令,当所述指令被执行时,使如权利要求1至13中任一项所述的方法被实现,或当所述指令被执行时,使如权利要求14至24中任一项所述的方法被实现。
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202280003073.9A CN117859304A (zh) | 2022-08-05 | 2022-08-05 | 一种数据类型确定方法/装置/设备及存储介质 |
PCT/CN2022/110708 WO2024026889A1 (zh) | 2022-08-05 | 2022-08-05 | 一种数据类型确定方法/装置/设备及存储介质 |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/CN2022/110708 WO2024026889A1 (zh) | 2022-08-05 | 2022-08-05 | 一种数据类型确定方法/装置/设备及存储介质 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2024026889A1 true WO2024026889A1 (zh) | 2024-02-08 |
Family
ID=89848403
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/CN2022/110708 WO2024026889A1 (zh) | 2022-08-05 | 2022-08-05 | 一种数据类型确定方法/装置/设备及存储介质 |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN117859304A (zh) |
WO (1) | WO2024026889A1 (zh) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101364232A (zh) * | 2008-10-06 | 2009-02-11 | 深圳华为通信技术有限公司 | 一种非易失性数据备份方法、终端与系统 |
CN111837425A (zh) * | 2020-06-10 | 2020-10-27 | 北京小米移动软件有限公司 | 一种接入方法、接入装置及存储介质 |
CN111885404A (zh) * | 2020-07-16 | 2020-11-03 | RealMe重庆移动通信有限公司 | 数据传输方法、设备及存储介质 |
CN113938911A (zh) * | 2020-07-13 | 2022-01-14 | 华为技术有限公司 | 一种通信方法、设备及系统 |
CN114430920A (zh) * | 2020-07-27 | 2022-05-03 | 北京小米移动软件有限公司 | 能力上报方法和装置、能力确定方法和装置 |
-
2022
- 2022-08-05 CN CN202280003073.9A patent/CN117859304A/zh active Pending
- 2022-08-05 WO PCT/CN2022/110708 patent/WO2024026889A1/zh active Application Filing
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101364232A (zh) * | 2008-10-06 | 2009-02-11 | 深圳华为通信技术有限公司 | 一种非易失性数据备份方法、终端与系统 |
CN111837425A (zh) * | 2020-06-10 | 2020-10-27 | 北京小米移动软件有限公司 | 一种接入方法、接入装置及存储介质 |
CN113938911A (zh) * | 2020-07-13 | 2022-01-14 | 华为技术有限公司 | 一种通信方法、设备及系统 |
CN111885404A (zh) * | 2020-07-16 | 2020-11-03 | RealMe重庆移动通信有限公司 | 数据传输方法、设备及存储介质 |
CN114430920A (zh) * | 2020-07-27 | 2022-05-03 | 北京小米移动软件有限公司 | 能力上报方法和装置、能力确定方法和装置 |
Also Published As
Publication number | Publication date |
---|---|
CN117859304A (zh) | 2024-04-09 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2023010471A1 (zh) | 一种传输配置指示tci状态配置的方法及其装置 | |
WO2023206179A1 (zh) | 一种确定传输配置指示状态的方法及装置 | |
WO2024026799A1 (zh) | 数据传输方法和装置 | |
WO2024021128A1 (zh) | 同步参考源的确定方法及装置 | |
WO2024026801A1 (zh) | 一种侧行链路sl波束配置方法、装置、设备及存储介质 | |
WO2024050776A1 (zh) | 一种信息确定方法/装置/设备及存储介质 | |
WO2024007171A1 (zh) | 一种算力负载均衡方法及装置 | |
WO2023197187A1 (zh) | 一种信道状态信息的处理方法及装置 | |
WO2024026889A1 (zh) | 一种数据类型确定方法/装置/设备及存储介质 | |
WO2024000202A1 (zh) | 一种信道状态信息csi反馈的确定方法及其装置 | |
WO2023168575A1 (zh) | 一种天线切换能力上报方法及其装置 | |
WO2024182948A1 (zh) | 一种定位方法、装置、设备及存储介质 | |
WO2023230794A1 (zh) | 一种定位方法及装置 | |
WO2024036456A1 (zh) | 一种基于人工智能ai提供服务的方法、装置、设备及存储介质 | |
WO2023201755A1 (zh) | 一种移动性管理的配置方法及装置 | |
WO2024021130A1 (zh) | 一种信道估计方法及其装置 | |
WO2023236123A1 (zh) | 一种系统消息传输方法/装置/设备及存储介质 | |
WO2023168611A1 (zh) | 一种附加解调参考信号dmrs的发送方法及其装置 | |
WO2022266963A1 (zh) | 资源分配方法及其装置 | |
WO2023236124A1 (zh) | 一种人工智能ai模型训练方法/装置/设备及存储介质 | |
WO2024000530A1 (zh) | 一种信道占用时间的共享方法及其装置 | |
WO2024168917A1 (zh) | 一种ai模型的注册方法/装置/设备及存储介质 | |
WO2023178622A1 (zh) | 一种dmrs端口指示方法及其装置 | |
WO2023168574A1 (zh) | 一种天线切换能力上报方法及其装置 | |
WO2024164123A1 (zh) | 一种反馈方法、装置、设备及存储介质 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
WWE | Wipo information: entry into national phase |
Ref document number: 202280003073.9 Country of ref document: CN |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 22953689 Country of ref document: EP Kind code of ref document: A1 |