Method, device, semiconductor chip and computer device for processing streaming data

Abstract

This application relates to a method, apparatus, semiconductor chip, and computer device for streaming data processing. The method includes: dividing the streaming data into multiple data blocks according to a time axis, wherein each data block corresponds to a time interval on the time axis; in a first time interval, a first calculation block performs calculations and classifications on the data blocks in the first time interval to obtain a first identification result and caches the first identification result; in a second time interval, a second calculation block performs calculations on the data blocks in the second time interval to obtain a second calculation result; the second calculation block retrieves the cached first identification result and classifies the first identification result and the second calculation result to obtain a second identification result. This method solves the problem of high repetitive calculation rate in streaming data processing by smart devices, which leads to high power consumption and high cost of smart devices. It reduces the repetitive calculation rate of streaming data, improves computational efficiency, and reduces costs.

Description

Technical Field

[0001] This application relates to the field of information processing technology, and in particular to methods, apparatus, semiconductor chips and computer equipment for streaming data processing. Background Technology

[0002] With the development of science and technology, more and more users are using smart devices to communicate with relatives and friends. During the process of using smart devices, the devices need to process a large amount of audio or video data, which is streaming data.

[0003] In related technologies, smart devices process streaming data through predefined and complete convolutional neural networks. The high rate of data repetition leads to high power consumption and high cost of smart devices.

[0004] Currently, no effective solution has been proposed for the problem of high repetitive computation rate in streaming data processing of intelligent devices, which leads to high power consumption and high cost of intelligent devices. Summary of the Invention

[0005] This application provides a method, apparatus, semiconductor chip, computer device, and computer-readable storage medium for streaming data processing, to at least solve the problem in the related art where the high rate of repetitive calculations in streaming data processing by smart devices leads to high power consumption and high cost of smart devices.

[0006] In a first aspect, embodiments of this application provide a method for streaming data processing, the method comprising:

[0007] Acquire streaming data within a preset time period, and divide the streaming data into multiple data blocks according to the time axis of the streaming data, wherein the data blocks correspond to the time intervals on the time axis;

[0008] The data block is computed using a convolutional neural network, which includes multiple computational blocks. In a first time interval, the first computational block of the convolutional neural network computes and classifies the data block in the first time interval to obtain a first recognition result, and caches the first recognition result.

[0009] During the second time interval, the second computation block of the convolutional neural network performs calculations on the data block in the second time interval to obtain a second calculation result. The second computation block obtains the cache of the first recognition result, classifies the first recognition result and the second calculation result, and obtains a second recognition result.

[0010] In some embodiments, both the first computation block and the second computation block correspond to time intervals on the time axis.

[0011] In some embodiments, after obtaining the second identification result, the method further includes:

[0012] The second identification result is cached in a first-in-first-out memory.

[0013] In some embodiments, the calculation and classification of the data blocks in the first time interval includes:

[0014] When the streaming data is audio data, the audio features of the audio data are extracted, and the voiceprint of the audio data is identified based on the audio features;

[0015] The audio data is classified and identified based on the voiceprint.

[0016] In some embodiments, the calculation and classification of the data blocks in the first time interval further includes:

[0017] When the streaming data is video data, image recognition is performed on the video frames of the video data, feature vectors of the video frames are extracted, and the video frames are classified and identified based on the feature vectors.

[0018] Secondly, embodiments of this application provide an audio data processing apparatus, the apparatus including a microphone and a microprocessor;

[0019] The microphone acquires audio data within a preset time period, and the microprocessor divides the audio data into multiple data blocks according to the time axis of the audio data, wherein the data blocks correspond to the time intervals on the time axis;

[0020] In a first time interval, the microprocessor performs calculations and classifications on the data blocks in the first time interval to obtain a first identification result, and caches the first identification result in a first-in-first-out memory. In a second time interval, the microprocessor performs calculations on the data blocks in the second time interval to obtain a second calculation result. The microprocessor retrieves the cache of the first identification result, classifies and identifies the first identification result and the second calculation result, and obtains a second identification result.

[0021] In some embodiments, the microprocessor includes a convolutional neural network that computes and classifies the data blocks in the time interval, wherein the convolutional neural network includes a plurality of computation blocks that correspond to time intervals on the time axis.

[0022] Thirdly, embodiments of this application provide a semiconductor chip for streaming data processing, the chip including a neural network accelerator:

[0023] The receiver of the semiconductor chip acquires streaming data within a preset time period. The neural network accelerator divides the streaming data into multiple data blocks according to the time axis of the streaming data, wherein the data blocks correspond to the time intervals on the time axis.

[0024] The neural network accelerator calculates and classifies the data block in the first time interval through the first computation block of the convolutional neural network in the first time interval to obtain a first recognition result, and caches the first recognition result;

[0025] In the second time interval, the neural network accelerator performs calculations on the data block in the second time interval through the second calculation block of the convolutional neural network to obtain a second calculation result. The neural network accelerator obtains the cache of the first recognition result, classifies the first recognition result and the second calculation result, and obtains a second recognition result.

[0026] In some embodiments, in the neural network accelerator, the first computation block and the second computation block of the convolutional neural network both correspond to time intervals on the time axis.

[0027] Fourthly, embodiments of this application provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any of the methods described above.

[0028] Fifthly, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements any of the methods described above.

[0029] Compared to related technologies, the streaming data processing method provided in this application, by acquiring streaming data within a preset time period, dividing the streaming data into multiple data blocks according to the time axis of the streaming data, wherein each data block corresponds to a time interval on the time axis, and performing calculations on the data blocks using a convolutional neural network, the convolutional neural network including multiple calculation blocks, in a first time interval, the first calculation block of the convolutional neural network performs calculations and classifications on the data blocks in the first time interval to obtain a first recognition result, and caches the first recognition result, in a second time interval, the second calculation block of the convolutional neural network performs calculations on the data blocks in the second time interval to obtain a second calculation result, the second calculation block obtains the cached first recognition result, classifies the first recognition result and the second calculation result to obtain a second recognition result, solves the problem of high repetitive calculation rate in streaming data processing by smart devices, which leads to high power consumption and high cost of smart devices, reduces the repetitive calculation rate of streaming data, improves calculation efficiency, and reduces costs.

[0030] Details of one or more embodiments of this application are set forth in the following drawings and description to make other features, objects and advantages of this application more readily apparent. Attached Figure Description

[0031] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0032] Figure 1 This is a flowchart of a streaming data processing method according to an embodiment of this application;

[0033] Figure 2 This is a flowchart of an audio data processing method according to an embodiment of this application;

[0034] Figure 3 This is a structural block diagram of an audio data processing apparatus according to an embodiment of this application;

[0035] Figure 4 This is a schematic diagram of the structure of a convolutional neural network according to an embodiment of this application;

[0036] Figure 5 This is a schematic diagram of the internal structure of an electronic device according to an embodiment of this application. Detailed Implementation

[0037] To make the objectives, technical solutions, and advantages of this application clearer, the application is described and illustrated below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application. All other embodiments obtained by those skilled in the art based on the embodiments provided in this application without inventive effort are within the scope of protection of this application.

[0038] Obviously, the accompanying drawings described below are merely some examples or embodiments of this application. Those skilled in the art can apply this application to other similar scenarios based on these drawings without any inventive effort. Furthermore, it is understood that although the efforts made in this development process may be complex and lengthy, for those skilled in the art related to the content disclosed in this application, any changes to design, manufacturing, or production based on the technical content disclosed in this application are merely conventional technical means and should not be construed as insufficient disclosure of the content of this application.

[0039] In this application, the reference to "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment that is mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described in this application may be combined with other embodiments without conflict.

[0040] Unless otherwise defined, the technical or scientific terms used in this application shall have the ordinary meaning understood by one of ordinary skill in the art to which this application pertains. The terms “a,” “an,” “an,” “the,” and similar words used in this application do not indicate quantity limitation and may indicate singular or plural. The terms “comprising,” “including,” “having,” and any variations thereof used in this application are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or modules (units) is not limited to the listed steps or units, but may also include steps or units not listed, or may include other steps or units inherent to these processes, methods, products, or devices. The terms “connected,” “linked,” “coupled,” and similar words used in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. “Multiple” used in this application refers to two or more. “And / or” describes the relationship between related objects, indicating that three relationships may exist; for example, “A and / or B” can represent: A alone, A and B simultaneously, and B alone. The character " / " generally indicates that the preceding and following objects are in an "or" relationship. The terms "first," "second," and "third" used in this application are merely to distinguish similar objects and do not represent a specific ordering of the objects.

[0041] This embodiment provides a method for streaming data processing. Figure 1 This is a flowchart of a streaming data processing method according to an embodiment of this application, such as... Figure 1 As shown, the method includes the following steps:

[0042] Step S101: Obtain streaming data within a preset time period, and divide the streaming data into multiple data blocks according to the time axis of the streaming data, wherein each data block corresponds to a time interval on the time axis.

[0043] In this embodiment, the streaming data is a sequence of data arriving in chronological order, in large quantities, rapidly, and continuously, such as audio or video data generated during user interaction through a smart device. The preset time period is a pre-defined time value for processing the streaming data, divided along a time axis. This preset time period includes multiple time intervals. For example, in the case of video data, the preset time period could be 1 second, and the time intervals could be 30, meaning that the 1-second preset time period includes 30 time intervals. The streaming data is divided into blocks according to these time intervals, resulting in multiple data blocks within 1 second. For example, in video data, the image frame data in each time interval can be considered as one data block, and the number of image frames in one data block can be 1, 2, or other values.

[0044] Step S102: The data block is calculated using a convolutional neural network. The convolutional neural network includes multiple calculation blocks. In the first time interval, the first calculation block of the convolutional neural network calculates and classifies the data block in the first time interval to obtain a first recognition result, and caches the first recognition result.

[0045] Convolutional Neural Networks (CNNs) are a type of feedforward neural network that incorporates convolutional computations and has a deep structure. They are one of the representative algorithms of deep learning. CNNs possess representation learning capabilities, enabling translation-invariant classification of input information according to their hierarchical structure. CNNs are built by mimicking the visual perception mechanism of biological systems. They can perform both supervised and unsupervised learning. The shared parameters of the convolutional kernels within their hidden layers and the sparsity of inter-layer connections allow CNNs to learn gridded features with relatively low computational cost, and the learning results are stable. Furthermore, they do not require additional feature engineering for streaming data.

[0046] In this embodiment, within the first time interval, the first computation block in the convolutional neural network only computes the first data block within the first time interval, and then caches the recognition result.

[0047] In step S103, during the second time interval, the second computation block of the convolutional neural network performs computation on the data block in the second time interval to obtain a second computation result. The second computation block obtains the cache of the first recognition result, classifies the first recognition result and the second computation result, and obtains a second recognition result.

[0048] During the second time interval, the second computation block in the convolutional neural network only computes the data in the second data block, and then combines the first recognition result in the cache to classify and recognize the first recognition result and the second computation result together to obtain the second recognition result.

[0049] Through steps S101 to S103 above, based on the time axis, in each time interval, the computation block in the convolutional neural network only calculates the data block in that time interval, and then classifies and recognizes the recognition result together with the recognition result cached in the previous time interval. This eliminates the computation of a large amount of duplicate data, solves the problem of high repetitive computation rate of streaming data processing in smart devices, which leads to high power consumption and high cost of smart devices, reduces the repetitive computation rate of streaming data, improves computational efficiency, and reduces costs.

[0050] In some embodiments, both the first and second computation blocks correspond to time intervals on the time axis. In this embodiment, the convolutional neural network is divided into multiple computation blocks according to the time axis. These computation blocks correspond to time intervals at the classification and recognition levels, and further correspond to data blocks in the streaming data. This facilitates the processing of streaming data by the convolutional neural network, improving its computational efficiency while ensuring its computational accuracy.

[0051] In some embodiments, it is necessary to train the weight parameters of multiple computation blocks in the convolutional neural network to find the optimal values ​​of the weight parameters. During the training process, the weight parameters of multiple computation blocks are trained simultaneously.

[0052] In some embodiments, after obtaining the second recognition result, the streaming data processing method further includes: caching the second recognition result in a first-in, first-out (FIFO) memory. FIFO buffers are typically used for data transmission between different clock domains. During data caching, if the rate at which data enters the FIFO buffer interface exceeds the interface's transmission rate, the FIFO buffer allows data to enter the queue in the order it arrives at the interface. Simultaneously, the FIFO buffer dequeues data in the order it entered the buffer, with earlier data being dequeued first and later data being dequeued last, until all recognition results are cached. This embodiment uses FIFO for caching recognition results, simplifying data processing and reducing processing costs.

[0053] This embodiment also provides a method for audio data processing. Figure 2 This is a flowchart of an audio data processing method according to an embodiment of this application, such as... Figure 2 As shown, when the streaming data is audio data, the method further includes the following steps:

[0054] Step S201: Extract the audio features of the audio data and identify the voiceprint of the audio data based on the audio features. The audio data can be real-time sound data acquired by a microphone or sound data cached in the mobile phone. Audio features include signal amplitude, frequency, continuity, etc., and the voiceprint is the spectral representation of the sound signal. In this embodiment, one frame of audio data can be extracted every 10 milliseconds.

[0055] Step S202: Based on the voiceprint, classify and identify the audio data. The voiceprint contains a large amount of information about the sound data. By calculating, classifying, and identifying this information, the processing result of the input audio data can be obtained.

[0056] Through the above steps S201 and S202, after acquiring the audio data, the audio data is filtered and scaled by the computation blocks in the convolutional neural network to extract the audio features in each frame of audio data. Then, the audio features are classified and identified to obtain the final audio data result, which improves the processing efficiency of audio data.

[0057] In some embodiments, when the streaming data is video data, image recognition is performed on the video frames of the video data to extract the feature vectors of the video frames, and the video frames are classified and identified based on the feature vectors. For example, if there is only one video frame in a time interval, the computation block in the convolutional neural network extracts features from that video frame to achieve classification and identification of the video frame, thereby improving the processing efficiency of the video data.

[0058] It should be noted that the steps shown in the above process or in the flowchart of the accompanying figures can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0059] This embodiment also provides an audio data processing apparatus for implementing the above embodiments and preferred embodiments, which will not be repeated hereafter. As used below, the terms "module," "unit," "subunit," etc., can refer to a combination of software and / or hardware that performs a predetermined function. Although the apparatus described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0060] In some embodiments, Figure 3 This is a structural block diagram of an audio data processing apparatus according to an embodiment of this application, such as... Figure 3 As shown, the device includes: a microphone 31 and a microprocessor 32;

[0061] Microphone 31 acquires audio data within a preset time period, converts the audio data into a digital signal, and sends it to microprocessor 32 for processing. Microprocessor 32 divides the audio data into multiple data blocks according to the time axis of the audio data, wherein each data block corresponds to a time interval on the time axis. In the first time interval, microprocessor 32 calculates and classifies the data blocks in the first time interval to obtain a first recognition result, and caches the first recognition result in a first-in-first-out memory. In the second time interval, microprocessor 32 calculates the data blocks in the second time interval to obtain a second calculation result. Microprocessor 32 retrieves the cached first recognition result from the first-in-first-out memory, classifies and recognizes the first recognition result and the second calculation result, and obtains a second recognition result.

[0062] In related technologies, during classification and recognition at four time intervals, each recognition result requires querying data from 32 time intervals. In calculating the streaming data in [t-31, t] and [t-27, t+4], the data duplication rate reaches 87%, which is unacceptable for low-power devices and mobile devices. In this embodiment, based on the time axis, within each time interval, the computational block in the convolutional neural network only calculates the data block within that time interval. Then, the recognition result is combined with the recognition result cached in the previous time interval for classification and recognition. This eliminates a large amount of redundant data computation, solving the problem of high power consumption and high cost in smart devices due to high redundant computation rates in streaming data processing. It reduces the redundant computation rate of streaming data, improves computational efficiency, and lowers costs.

[0063] In some embodiments, the microprocessor 32 includes a convolutional neural network that performs calculations and classifications on the data blocks in the time interval, wherein the convolutional neural network includes a plurality of computation blocks that correspond to time intervals on the time axis.

[0064] In related technologies, the repetition rate can be reduced by increasing the time interval between classification and recognition. For example, changing the classification and recognition process from four time intervals to eight time intervals will decrease the repetition rate to 75%. However, at this point, the recognition accuracy decreases because the data signal starts from the middle of the time interval. In this embodiment, the convolutional neural network is divided into multiple computational blocks according to the time axis. These computational blocks correspond to the time intervals at the classification and recognition level, and further correspond to the data blocks in the streaming data. This facilitates the processing of streaming data by the convolutional neural network, improving its computational efficiency while ensuring its computational accuracy.

[0065] In some embodiments, a semiconductor chip for streaming data processing is also provided, the semiconductor chip including a neural network accelerator; the receiver of the semiconductor chip acquires streaming data within a preset time period, the neural network accelerator divides the streaming data into multiple data blocks according to the time axis of the streaming data, wherein the data blocks correspond to time intervals on the time axis; in a first time interval, the neural network accelerator performs calculations and classifications on the data blocks in the first time interval through a first computation block of a convolutional neural network to obtain a first recognition result, and caches the first recognition result; in a second time interval, the neural network accelerator performs calculations on the data blocks in the second time interval through a second computation block of the convolutional neural network to obtain a second calculation result, the neural network accelerator acquires the cached first recognition result, classifies the first recognition result and the second calculation result to obtain a second recognition result.

[0066] In this embodiment, based on the time axis, within each time interval, the computation block in the convolutional neural network only computes the data block within that time interval, and then classifies and identifies the recognition result together with the recognition result cached in the previous time interval. This eliminates the computation of a large amount of duplicate data, solves the problem of high repetitive computation rate of streaming data processing in smart devices, which leads to high power consumption and high cost of smart devices, reduces the repetitive computation rate of streaming data, improves computational efficiency, reduces the power consumption of hardware devices, and thus reduces costs.

[0067] In some embodiments, in a neural network accelerator of a semiconductor chip, the first and second computation blocks of a convolutional neural network both correspond to time intervals on the time axis. Figure 4 This is a schematic diagram of the structure of a convolutional neural network according to an embodiment of this application, such as... Figure 4 As shown, along the time axis, the convolutional neural network is divided into multiple computational blocks, named Block0, Block1, Block2, Block3, Block…, Block-n+3, Block-n+2, Block-n+1, ​​Block-n, etc. Following these computational blocks are flattened layers, fully connected layers (FC), and classes. In this embodiment, the convolutional neural network is divided into multiple computational blocks along the time axis. These blocks correspond to time intervals at the classification and recognition levels, and further correspond to data blocks in the streaming data. This facilitates the processing of streaming data by the convolutional neural network, improving its computational efficiency while maintaining its computational accuracy.

[0068] It should be noted that the above modules can be functional modules or program modules, and can be implemented through software or hardware. For modules implemented through hardware, the above modules can reside in the same processor; or the above modules can be located in different processors in any combination.

[0069] In one embodiment, a computer device is provided, which may be a terminal. The computer device includes a processor, memory, a network interface, a display screen, and an input device connected via a system bus. The processor of the computer device provides computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, it implements a streaming data processing method. The display screen of the computer device may be a liquid crystal display (LCD) or an electronic ink display. The input device of the computer device may be a touch layer covering the display screen, or buttons, a trackball, or a touchpad located on the casing of the computer device, or an external keyboard, touchpad, or mouse, etc.

[0070] In one embodiment, Figure 5 This is a schematic diagram of the internal structure of an electronic device according to an embodiment of this application, such as... Figure 5 As shown, an electronic device is provided, which can be a server, and its internal structure diagram can be as follows. Figure 5 As shown, this electronic device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores data. The network interface communicates with external terminals via a network connection. When the computer program is executed by the processor, it implements a streaming data processing method.

[0071] Those skilled in the art will understand that Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the electronic device to which the present application is applied. The specific electronic device may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements.

[0072] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the streaming data processing method provided in the above embodiments.

[0073] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of the streaming data processing method provided in the above embodiments.

[0074] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0075] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0076] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method for streaming data processing, characterized in that, The method includes: Acquire streaming data within a preset time period, and divide the streaming data into multiple data blocks according to the time axis of the streaming data, wherein the data blocks correspond to the time intervals on the time axis; The data block is computed using a convolutional neural network, which includes multiple computational blocks. In a first time interval, the first computational block of the convolutional neural network computes and classifies the data block in the first time interval to obtain a first recognition result, and caches the first recognition result. During the second time interval, the second computation block of the convolutional neural network performs calculations on the data block in the second time interval to obtain a second calculation result. The second computation block obtains the cache of the first recognition result, classifies and recognizes the first recognition result together with the second calculation result, eliminates duplicate data, and obtains the second recognition result.

2. The method according to claim 1, characterized in that, Both the first calculation block and the second calculation block correspond to the time interval on the time axis.

3. The method according to claim 1, characterized in that, After obtaining the second identification result, the method further includes: The second identification result is cached in a first-in-first-out memory.

4. The method according to claim 1, characterized in that, The calculation and classification of the data blocks in the first time interval includes: When the streaming data is audio data, the audio features of the audio data are extracted, and the voiceprint of the audio data is identified based on the audio features; The audio data is classified and identified based on the voiceprint.

5. The method according to claim 1, characterized in that, The calculation and classification of the data blocks in the first time interval also includes: When the streaming data is video data, image recognition is performed on the video frames of the video data, feature vectors of the video frames are extracted, and the video frames are classified and identified based on the feature vectors.

6. An audio data processing apparatus, characterized in that, The device includes a microphone and a microprocessor; The microphone acquires audio data within a preset time period, and the microprocessor divides the audio data into multiple data blocks according to the time axis of the audio data, wherein the data blocks correspond to the time intervals on the time axis; In a first time interval, the microprocessor performs calculations and classifications on the data blocks in the first time interval to obtain a first identification result, and caches the first identification result in a first-in-first-out memory. In a second time interval, the microprocessor performs calculations on the data blocks in the second time interval to obtain a second calculation result. The microprocessor retrieves the cache of the first identification result, classifies and identifies the first identification result and the second calculation result, and obtains a second identification result.

7. The apparatus according to claim 6, characterized in that, The microprocessor includes a convolutional neural network that performs calculations and classifications on the data blocks in the time interval. The convolutional neural network includes multiple computation blocks, which correspond to time intervals on the time axis.

8. A semiconductor chip for streaming data processing, characterized in that, The chip includes a neural network accelerator: The receiver of the semiconductor chip acquires streaming data within a preset time period. The neural network accelerator divides the streaming data into multiple data blocks according to the time axis of the streaming data, wherein the data blocks correspond to the time intervals on the time axis. The neural network accelerator calculates and classifies the data block in the first time interval through the first computation block of the convolutional neural network in the first time interval to obtain a first recognition result, and caches the first recognition result; In the second time interval, the neural network accelerator performs calculations on the data block in the second time interval through the second calculation block of the convolutional neural network to obtain a second calculation result. The neural network accelerator obtains the cache of the first recognition result, classifies and recognizes the first recognition result together with the second calculation result, deletes duplicate data, and obtains the second recognition result.

9. The semiconductor chip according to claim 8, characterized in that, In the neural network accelerator, the first computation block and the second computation block of the convolutional neural network both correspond to the time interval on the time axis.

10. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1 to 5.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps of the method as described in any one of claims 1 to 5.

Images