WO2021135535A1 - 歌词改编方法、装置、计算机设备及存储介质 - Google Patents

歌词改编方法、装置、计算机设备及存储介质 Download PDF

Info

Publication number
WO2021135535A1
WO2021135535A1 PCT/CN2020/122425 CN2020122425W WO2021135535A1 WO 2021135535 A1 WO2021135535 A1 WO 2021135535A1 CN 2020122425 W CN2020122425 W CN 2020122425W WO 2021135535 A1 WO2021135535 A1 WO 2021135535A1
Authority
WO
WIPO (PCT)
Prior art keywords
preset
words
lyric
target
document
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2020/122425
Other languages
English (en)
French (fr)
Inventor
邓黎明
王龙
纪明阔
庄伯金
王少军
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Ping An Technology Shenzhen Co Ltd
Original Assignee
Ping An Technology Shenzhen Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Ping An Technology Shenzhen Co Ltd filed Critical Ping An Technology Shenzhen Co Ltd
Publication of WO2021135535A1 publication Critical patent/WO2021135535A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00Handling natural language data
    • G06F40/10Text processing
    • G06F40/166Editing, e.g. inserting or deleting
    • G06F40/186Templates

Definitions

  • This application relates to the fields of artificial intelligence, natural language processing and blockchain technology, and in particular to a method, device, computer equipment and storage medium for adapting lyrics.
  • the embodiments of the present application provide a method, device, computer equipment and storage medium for lyric adaptation, aiming to solve the problem of low quality of lyrics after lyric adaptation in the prior art.
  • an embodiment of the present application provides a lyrics adaptation method, which includes:
  • a lyric document to be adapted includes a preset sentence pattern template word, a preset topic word, a preset related word, and a preset weakly related word;
  • the preset topic words and the preset related words in the lyric document to be adapted are filtered out, and the preset topic words and the preset related words are combined with a preset marker. Occluding and covering related words to generate a target lyric template, wherein the target lyric template includes the preset sentence pattern template word and the preset weakly related word;
  • the target lyric template and the target topic word are input into a preset adaptation model to generate a lyric document related to the target topic word.
  • an embodiment of the present application also provides a lyrics adaptation device, which includes:
  • An acquiring unit for acquiring a lyric document to be adapted includes a preset sentence pattern template word, a preset topic word, a preset related word, and a preset weakly related word;
  • the screening and covering unit is configured to screen out the preset topic words and the preset related words in the lyric document to be adapted according to a preset method and a preset screening model, and use a preset marker to mark the preset topic Words and the preset related words are occluded and covered to generate a target lyric template, wherein the target lyric template includes the preset sentence pattern template word and the preset weakly related word;
  • the generating unit is configured to input the target lyric template and the target topic word into a preset adaptation model to generate a lyric document related to the target topic word.
  • an embodiment of the present application also provides a computer device, which includes a memory and a processor, the memory stores a computer program, and the processor implements the above method when the computer program is executed.
  • the embodiments of the present application also provide a computer-readable storage medium, the storage medium stores a computer program, and the computer program can implement the foregoing method when executed by a processor.
  • the preset subject words and the preset related words in the lyric document to be adapted are first screened out according to the preset method and the preset filtering model, and the preset markers are used to cover them to generate the target lyric template.
  • the target lyric template, the target theme word and the preset adaptation model a lyric document related to the target theme word is generated. Since the target lyric template retains the inherent structure of the lyric document to be adapted, it can improve the performance of the adapted lyrics to a certain extent. Lyric quality.
  • FIG. 1 is a schematic diagram of a scene of a method for adapting lyrics provided by an embodiment of the application
  • FIG. 2 is a schematic flowchart of a method for adapting lyrics according to an embodiment of the application
  • FIG. 3 is a schematic diagram of a sub-flow of a method for adapting lyrics according to an embodiment of the application
  • FIG. 4 is a schematic diagram of a sub-flow of a method for adapting lyrics according to an embodiment of the application
  • FIG. 5 is a schematic diagram of a sub-flow of a method for adapting lyrics according to an embodiment of the application
  • FIG. 6 is a schematic diagram of a sub-flow of a method for adapting lyrics according to an embodiment of the application
  • FIG. 7 is a schematic block diagram of a device for adapting lyrics according to an embodiment of the application.
  • FIG. 8 is a schematic block diagram of the screening and covering unit of the lyrics adaptation device provided by an embodiment of the application.
  • FIG. 9 is a schematic block diagram of the first screening and covering subunit of the lyrics adaptation device provided by an embodiment of the application.
  • FIG. 10 is a schematic block diagram of the second screening and covering subunit of the lyrics adaptation device provided by an embodiment of the application.
  • FIG. 11 is a schematic block diagram of the third screening and covering subunit of the lyrics adaptation device provided by an embodiment of the application.
  • FIG. 12 is a schematic block diagram of a generating unit of a lyrics adaptation device provided by an embodiment of the application.
  • FIG. 13 is a schematic block diagram of the first generating subunit of the lyrics adaptation device provided by an embodiment of the application.
  • FIG. 14 is a schematic block diagram of a computer device provided by an embodiment of this application.
  • the term “if” can be interpreted as “when” or “once” or “in response to determination” or “in response to detection” depending on the context .
  • the phrase “if determined” or “if detected [described condition or event]” can be interpreted as meaning “once determined” or “in response to determination” or “once detected [described condition or event]” depending on the context ]” or “in response to detection of [condition or event described]”.
  • FIG. 1 is a schematic diagram of a scene of a method for adapting lyrics provided by an embodiment of the present application.
  • the lyrics adaptation method of the embodiment of the present application can be applied to a server, for example, the lyrics adaptation method can be implemented by a software program configured on the server.
  • the server communicates with the terminal, so that the server calls the lyric file to be adapted uploaded by the user through the terminal and performs a series of processing on the lyric file to be adapted to obtain the adapted lyric file, thereby realizing the adaptation of the lyrics.
  • the terminal can be a desktop computer, a laptop computer, a tablet computer, etc., and there is no specific restriction here.
  • the number of the terminal and the user is one. It can be understood that in the actual application process, the number of the terminal and the user may be multiple, and FIG. 1 only serves as a schematic illustration.
  • FIG. 2 is a schematic flowchart of a method for adapting lyrics according to an embodiment of the present application. As shown in Figure 2, the method includes the following steps S100-S120.
  • Obtain a lyric document to be adapted where the lyric document to be adapted includes a preset sentence pattern template word, a preset topic word, a preset related word, and a preset weakly related word.
  • the server if the server is to implement the adaptation of lyrics, it first needs to obtain the lyrics file to be adapted, where the lyrics file to be adapted includes preset sentence template words, preset topic words, preset related words, and preset words. Set weakly related words, and then perform a series of processing based on the to-be-adapted lyric document before generating the adapted lyric document.
  • the user can upload the lyric file to be adapted through the user terminal.
  • the user can upload the lyric file to be adapted through the lyric adaptation web page of the user terminal to send the lyric file to be adapted to server.
  • the lyrics file to be adapted is a text file.
  • lyrics document to be adapted may also be other types of documents, such as Word documents.
  • the server after the server obtains the lyric document to be adapted, it needs to filter out the preset subject words and the preset related words in the lyric document to be adapted according to the preset method and the preset filtering model, and use the preset marker to predict the Set subject words and preset related words for occlusion and coverage to generate target lyrics template.
  • the preset method is, for example, a part of speech judgment method, which includes nouns, pronouns, verbs, adjectives, and so on.
  • the preset markers are, for example, spaces, #, etc.
  • the generated target lyric template can be uploaded to the blockchain, and the corresponding summary information is obtained based on the target lyric template.
  • the summary information is obtained by hashing the target lyric template , Such as the use of sha256s algorithm processing.
  • Uploading summary information to the blockchain can ensure its security and fairness and transparency to users.
  • the user equipment can download the summary information from the blockchain to verify whether the differentiated information and the updated pre-stored configuration information have been tampered with.
  • the blockchain referred to in this example is a new application mode of computer technology such as distributed data storage, point-to-point transmission, consensus mechanism, and encryption algorithm.
  • Blockchain essentially a decentralized database, is a series of data blocks associated with cryptographic methods. Each data block contains a batch of network transaction information for verification. The validity of the information (anti-counterfeiting) and the generation of the next block.
  • the step S110 may include the following steps S111-S112.
  • the words that are not the preset sentence pattern template in the lyric document to be adapted are identified; all the sentence patterns that are not the preset sentence pattern in the lyric document to be adapted are identified
  • the template words are input into a preset screening model to filter out the preset topic words and the preset related words, and the preset topic words and the preset related words are covered with a preset marker to generate a target lyrics template; wherein,
  • the template word of the preset sentence pattern means that a sentence must be organized according to a certain pattern, and the word that generates this pattern is the template word of the preset sentence pattern.
  • the step S112 may include the following steps S1121-S1122.
  • all words in the lyric document to be adapted that are not the preset sentence pattern template are input into a preset screening model to filter out the preset topic words and the preset related words, and use the preset A marker is set to cover the preset subject words and the preset related words to generate a target lyrics template.
  • all words in the lyric document to be adapted that are not the preset sentence pattern template words are input into the first preset screening model to filter out preset topic words, and the preset topic words are marked with a preset To generate a first lyric document; input the first lyric document and the preset subject words into a second preset model to filter out the preset related words, and put the first lyric document in the first lyric document.
  • the preset related words are occluded and covered with preset markers to generate a second lyric document, and the second lyric document is used as a target lyric template.
  • the first preset model is a TF-IDF (Term Frequency-Inverse Document Frequency) model, and the TF-IDF model is used to evaluate the importance of a word to a document set or a document to a corpus in which it is located .
  • the second preset model is the Word2Vec (Word to vector) model
  • the Word2Vec model can be used to map each word to a vector, which can be used to represent the relationship between word-to-word
  • the preset topic word is the topic word of the lyrics document to be adapted
  • the preset related word is the topic word Strongly associated words, more specifically, in this embodiment, first calculate the word frequency of each word in the lyrics document after the preset sentence pattern template words are extracted, and the inverse document of each word in the entire lyrics document Frequency, then combine the word frequency and the inverse document frequency to calculate the TF-IDF score corresponding to each word, and sort the TF-IDF scores, and select the N words with the highest TF-IDF score as the preset topic words, where N is The value depends on the specific needs.
  • the lyric documents after filtering out the preset topic words are input into the Word2Vec model.
  • the Word2Vec model calculates the word vector corresponding to each preset topic word and performs a weighted average to obtain the lyrics document The topic vector, and finally by calculating the cosine similarity between the topic vector and each word vector in the lyric document, the preset related words and the preset weakly related words can be obtained, and the weakly related words are weakly related or not related to the preset topic words Joint words.
  • the preset topic words and the preset related words are covered with preset markers to form a target with only the preset weakly related words and the preset sentence template words Lyrics template.
  • the preset topic words and preset related words are covered with preset markers to form only preset weakly related words and preset sentences
  • the target lyric template of the template word, the target lyric template and the target topic word are input into the preset adaptation model to generate a lyric document related to the target topic word.
  • the step S120 may include the following steps S121-S122.
  • the target lyric template and the target topic word are input into a preset adaptation model to generate a lyric document related to the target topic word.
  • the target lyrics template and target topic words are input into a preset adaptation model to generate target topic replacement words related to the target topic words; then fill the target topic replacement words into the target lyrics template To generate a lyric document related to the target topic word.
  • the process of filling the target topic replacement words into the target lyric template is to replace the position of the preset markers with The process of replacing words with the target subject.
  • the target topic replacement words are filled into the target lyrics template to generate a lyrics document related to the target theme words.
  • the step S121 may include the following steps S1211-S1212.
  • the target lyrics template and target topic words are input into a preset adaptation model to generate target topic replacement words related to the target topic words.
  • the preset adaptation models include the Bert (Bidirectional Encoder Representations From Transformers) model and the Word2Vec model.
  • the Bert model is a model based on Transformer that uses a bidirectional language. It can extract the syntax and semantic information of the document, and can also extract the document context information.
  • the target topic candidate words related to the target topic words are generated, and then the target topic candidate words and target topic words are input into the Word2Vec model.
  • the similarity comparison determines the target theme replacement words, and the target theme replacement words are filled into the target lyric template to generate a lyric document related to the target theme word, thereby realizing the lyrics adaptation of a specific theme.
  • FIG. 7 is a schematic block diagram of a device 200 for adapting lyrics according to an embodiment of the present application.
  • the present application also provides a lyrics adaptation device 200.
  • the lyric adaptation device 200 includes a unit for executing the above-mentioned lyric adaptation method, and the device can be configured in a server.
  • the lyrics adaptation device 200 includes an acquiring unit 201, a screening and covering unit 202, and a generating unit 203.
  • the obtaining unit 201 is configured to obtain a lyric document to be adapted, wherein the lyric document to be adapted includes preset sentence template words, preset subject words, preset related words, and preset weakly related words; the screening and covering unit 202 uses According to a preset method and a preset filtering model, the preset topic words and the preset related words in the lyric document to be adapted are filtered out, and the preset topic words and the preset related words are combined with a preset marker.
  • the generating unit 203 is configured to combine the target lyric template and target subject words
  • the preset adaptation model is input to generate a lyric document related to the target topic word.
  • the screening and covering unit 202 includes a judging unit 2021 and a first screening and covering subunit 2022.
  • the judging unit 2021 is configured to identify words that are not the preset sentence pattern template in the lyric document to be adapted according to the part of speech judging method; the first screening and covering subunit 2022 is configured to combine the lyric document to be adapted All non-preset sentence pattern template words are input to a preset screening model to filter out the preset topic words and the preset related words, and the preset topic words and the preset related words are blocked by a preset marker Overwrite to generate target lyrics template.
  • the first screening coverage subunit 2022 includes a second screening coverage subunit 2023 and a third screening coverage subunit 2024.
  • the second screening and covering subunit 2023 is used for inputting all words in the lyric document to be adapted that are not the preset sentence pattern template words into the first preset screening model to screen out the preset topic words, and The preset topic words are occluded and covered with preset markers to generate a first lyric document;
  • the third screening and covering subunit 2024 is used to input the first lyric document and the preset topic words into a second preset model for filtering A preset related word is generated, and the preset related word is masked and covered with a preset marker in the first lyric document to generate a second lyric document, and the second lyric document is used as a target lyric template.
  • the second screening and covering sub-unit 2023 includes a fourth screening and covering sub-unit 2025.
  • the fourth screening and covering subunit 2025 is configured to input all non-pre-set sentence template words in the lyric document to be adapted into the TF-IDF model to filter out preset subject words, and to combine the preset The subject words are occluded and covered with preset markers to generate the first lyric document.
  • the third screening and covering subunit 2024 includes a fifth screening and covering subunit 2026.
  • the fifth screening and covering subunit 2026 is configured to input the first lyric document and the preset subject words into the Word2Vec model to filter out preset related words, and to combine the preset related words in the first lyric document
  • the second lyric document is generated by masking and covering with the preset marker, and the second lyric document is used as the target lyric template.
  • the generating unit 203 includes a first generating subunit 2031 and a second generating subunit 2032.
  • the first generation subunit 2031 is used to input the target lyrics template and target topic words into a preset adaptation model to generate target topic replacement words related to the target topic words; the second generation subunit 2032 is used to input all the target topic words.
  • the target topic replacement words are filled into the target lyric template to generate a lyric document related to the target topic words.
  • the first generating subunit 2031 In some embodiments, such as this embodiment, as shown in FIG. 13, the first generating subunit 2031
  • It includes a third generation sub-unit 2033 and a fourth generation sub-unit 2034.
  • the third generating subunit 2033 is used to input the target lyrics template and target topic words into the Bert model to generate target topic candidate words related to the target topic words;
  • the fourth generating subunit 2034 is used to input the target topic words into the Bert model;
  • the topic candidate words and the target topic words are input into the Word2Vec model to generate target topic replacement words related to the target topic words.
  • the above-mentioned lyrics adaptation device can be implemented in the form of a computer program, and the computer program can be run on a computer device as shown in FIG. 14.
  • FIG. 14 is a schematic block diagram of a computer device according to an embodiment of the present application.
  • the computer device 300 is a server.
  • the server may be an independent server or a server cluster composed of multiple servers.
  • the computer device 300 includes a processor 302, a memory, and a network interface 305 connected through a system bus 301, where the memory may include a non-volatile storage medium 303 and an internal memory 304.
  • the non-volatile storage medium 303 can store an operating system 3031 and a computer program 3032.
  • the computer program 3032 When executed, it can cause the processor 302 to execute a lyrics adaptation method.
  • the processor 302 is used to provide calculation and control capabilities to support the operation of the entire computer device 300.
  • the internal memory 304 provides an environment for the operation of the computer program 3032 in the non-volatile storage medium 303.
  • the processor 302 can make the processor 302 execute a lyrics adaptation method.
  • the network interface 305 is used for network communication with other devices.
  • the structure shown in FIG. 14 is only a block diagram of part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device 300 to which the solution of the present application is applied.
  • the specific computer device 300 may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.
  • the processor 302 is configured to run a computer program 3032 stored in the memory to implement the following steps: Obtain a lyric file to be adapted, where the lyric file to be adapted includes preset sentence template words and presets Subject words, preset related words, and preset weakly related words; according to a preset method and a preset filtering model, the preset subject words and the preset related words in the lyric document to be adapted are filtered out, and the preset tags are used The preset topic words and the preset related words are occluded and covered to generate a target lyrics template, wherein the target lyrics template includes the preset sentence pattern template words and the preset weakly related words; The lyric template and the target topic word are input into a preset adaptation model to generate a lyric document related to the target topic word.
  • the processor 302 filters out the preset topic words and the preset in the lyric document to be adapted according to the preset method and preset screening model.
  • the step of generating a target lyric template by masking and covering the preset topic words and the preset associated words with a preset tag the following steps are specifically implemented: identifying the lyric document to be adapted according to the part of speech judgment method Words that are not the preset sentence pattern template; input all the words that are not the preset sentence pattern template in the lyric document to be adapted into a preset filtering model to filter out the preset topic words and the preset Related words, and masking and covering the preset topic words and the preset related words with a preset marker to generate a target lyrics template.
  • the processor 302 performs the input of all non-pre-set sentence template words in the lyric document to be adapted into a preset screening model to filter out the pre-set
  • the following steps are specifically implemented: the lyric document to be adapted Input all non-preset sentence pattern template words in the first preset screening model to filter out preset topic words, and cover the preset topic words with preset markers to generate a first lyric document; Input the first lyric document and the preset topic words into a second preset model to filter out preset related words, and cover the preset related words with preset markers in the first lyric document To generate a second lyric file, and use the second lyric file as a target lyric template.
  • the processor 302 performs the input of all non-pre-set sentence template words in the lyric document to be adapted into the first preset screening model to filter out the pre-set screening model.
  • the following steps are specifically implemented: all non-presets in the to-be-adapted lyric document
  • the sentence template words are input into the TF-IDF model to filter out the preset topic words, and the preset topic words are covered with a preset marker to generate the first lyrics document.
  • the processor 302 is implementing the input of the first lyric document and the preset subject words into a second preset model to filter out preset related words, and in the In the first lyric document, the preset associated words are covered with preset markers to generate a second lyric document, and when the second lyric document is used as the target lyric template, the following steps are specifically implemented: The lyric document and the preset subject words are input into the Word2Vec model to filter out the preset related words, and the preset related words are covered with a preset marker in the first lyric document to generate a second lyric document.
  • the second lyric file serves as a target lyric template.
  • the processor 302 when the processor 302 implements the step of inputting the target lyric template and target topic word into a preset adaptation model to generate a lyric document related to the target topic word, specifically The following steps are achieved: input the target lyrics template and target topic words into a preset adaptation model to generate target topic replacement words related to the target topic words; fill the target topic replacement words into the target lyrics template to Generate a lyric document related to the target topic word.
  • the processor 302 when the processor 302 implements the step of inputting the target lyrics template and target topic words into a preset adaptation model to generate target topic replacement words related to the target topic words ,
  • the processor 302 may be a central processing unit (Central Processing Unit, CPU), and the processor 302 may also be other general-purpose processors, digital signal processors (Digital Signal Processors, DSPs), Application Specific Integrated Circuit (ASIC), Field-Programmable Gate Array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc.
  • the general-purpose processor may be a microprocessor or the processor may also be any conventional processor.
  • the computer program may be stored in a storage medium, and the storage medium is a computer-readable storage medium.
  • the computer program is executed by at least one processor in the computer system to implement the process steps of the foregoing method embodiment.
  • the storage medium may be a computer-readable storage medium.
  • the storage medium stores a computer program.
  • the processor executes the following steps: acquiring a lyric document to be adapted, wherein the lyric document to be adapted includes preset sentence template words, preset topic words, preset related words, and presets Weakly related words; filter out the preset topic words and the preset related words in the lyric document to be adapted according to the preset method and the preset filtering model, and use the preset marker to combine the preset topic words and all the words
  • the preset related words are occluded and covered to generate a target lyric template, wherein the target lyric template includes the preset sentence pattern template words and the preset weakly related words; the target lyric template and the target topic words are input into the preset
  • the model is adapted to generate a lyric document related to the target topic word.
  • the processor executes the computer program to realize the selection of the preset in the lyrics file to be adapted according to a preset method and a preset screening model.
  • the following steps are specifically implemented: identifying the subject words and the preset related words according to the part-of-speech judging method.
  • the processor is executing the computer program to realize the input of all non-pre-set sentence template words in the to-be-adapted lyric document into preset filtering
  • the model filters out the preset topic words and the preset related words, and uses preset markers to cover the preset topic words and preset related words to generate the target lyrics template step
  • the following steps are specifically implemented: All words in the to-be-adapted lyric document that are not the preset sentence pattern template are input into the first preset screening model to filter out the preset topic words, and the preset topic words are covered by a preset marker
  • input the first lyric document and the preset subject words into a second preset model to filter out preset related words, and use the preset related words in the first lyric document
  • the preset marker is masked and covered to generate a second lyric document, and the second lyric document is used as a target lyric template.
  • the processor executes the computer program to realize the input of all non-pre-set sentence template words in the to-be-adapted lyric document into the first preset.
  • the filter model is set to filter out the preset topic words, and the preset topic words are occluded and covered with preset markers to generate the first lyric document
  • the following steps are specifically implemented: the lyric document to be adapted All non-pre-set sentence template words are input into the TF-IDF model to filter out the preset topic words, and the preset topic words are covered with preset markers to generate the first lyrics document.
  • the processor executes the computer program to realize the input of the first lyric document and the preset subject words into a second preset model to filter out the preset model.
  • Set associated words, and cover the preset associated words with preset markers in the first lyric document to generate a second lyric document, and when the second lyric document is used as a target lyric template step the specific implementation is as follows Step: Input the first lyric document and the preset subject words into the Word2Vec model to filter out preset related words, and cover the preset related words with preset markers in the first lyric document to cover A second lyric file is generated, and the second lyric file is used as a target lyric template.
  • the processor is executing the computer program to realize the input of the target lyrics template and target topic words into a preset adaptation model to generate a data related to the target topic words
  • the lyrics document step the following steps are specifically implemented: input the target lyrics template and target subject words into a preset adaptation model to generate target subject replacement words related to the target subject word; fill the target subject replacement word to The target lyric template is used to generate a lyric document related to the target topic word.
  • the processor is executing the computer program to realize the input of the target lyrics template and target topic words into a preset adaptation model to generate a data related to the target topic words
  • the following steps are specifically implemented: input the target lyrics template and target theme words into the Bert model to generate target theme candidate words related to the target theme words; combine the target theme candidate words and all the target theme words
  • the target subject words are input into the Word2Vec model to generate target subject replacement words related to the target subject words.
  • the storage medium may be a U disk, a mobile hard disk, a read-only memory (Read-Only Memory, ROM), a magnetic disk or an optical disk, and other computer-readable storage media that can store program codes.
  • the computer-readable storage medium may It is non-volatile or volatile.
  • the disclosed device and method may be implemented in other ways.
  • the device embodiments described above are only illustrative.
  • the division of each unit is only a logical function division, and there may be other division methods in actual implementation.
  • multiple units or components can be combined or integrated into another system, or some features can be omitted or not implemented.
  • the steps in the method in the embodiment of the present application can be adjusted, merged, and deleted in order according to actual needs.
  • the units in the devices in the embodiments of the present application may be combined, divided, and deleted according to actual needs.
  • the functional units in the various embodiments of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units may be integrated into one unit.
  • the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a storage medium.
  • the technical solution of this application is essentially or the part that contributes to the existing technology, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium. It includes several instructions to make a computer device (which may be a personal computer, a terminal, or a network device, etc.) execute all or part of the steps of the methods described in the various embodiments of the present application.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Computational Linguistics (AREA)
  • General Health & Medical Sciences (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Machine Translation (AREA)

Abstract

一种歌词改编方法、装置、计算机设备及存储介质。该方法包括:获取待改编的歌词文档,其中,待改编的歌词文档包括预设句式模板词、预设主题词、预设关联词以及预设弱关联词(S100);根据预设方法及预设筛选模型筛选出待改编的歌词文档中的预设主题词及预设关联词并用预设标记符将预设主题词及预设关联词进行遮挡覆盖以生成目标歌词模板,其中,目标歌词模板包括预设句式模板词和预设弱关联词(S110);将目标歌词模板及目标主题词输入预设改编模型以生成与目标主题词相关的歌词文档(S120)。该方法可在一定程度上提升歌词改编后的歌词质量。

Description

歌词改编方法、装置、计算机设备及存储介质
本申请要求于2020年06月30日提交中国专利局、申请号为202010622899.6,发明名称为“歌词改编方法、装置、计算机设备及存储介质”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
技术领域
本申请涉及人工智能、自然语言处理以及区块链技术领域,尤其涉及一种歌词改编方法、装置、计算机设备及存储介质。
背景技术
歌曲是人们欣赏和喜爱的一种艺术形式,已经深入渗透到人们的生活中。歌词作为歌曲的重要组成部分,具有某种特定的结构用来表达某种特定的情感,这些特定的结构主要表现在段落分布、句式关联以及押韵等方面,而现有歌词改编时文本的生成方法主要有语言模型和Seq2Seq模型,但发明人发现这两种模型生成的字词都只整合了单侧的信息,当生成文本较长时,容易使得生成的字词不匹配,从而使得歌词上下文语境不相关联、歌词质量较低。
发明内容
本申请实施例提供了一种歌词改编方法、装置、计算机设备及存储介质,旨在解决现有歌词改编后歌词质量较低的问题。
第一方面,本申请实施例提供了一种歌词改编方法,其包括:
获取待改编的歌词文档,其中,所述待改编的歌词文档包括预设句式模板词、预设主题词、预设关联词以及预设弱关联词;
根据预设方法及预设筛选模型筛选出所述待改编的歌词文档中的所述预设主题词及所述预设关联词,并用预设标记符将所述预设主题词及所述预设关联词进行遮挡覆盖以生成目标歌词模板,其中,所述目标歌词模板包括所述预设句式模板词和所述预设弱关联词;
将所述目标歌词模板及目标主题词输入预设改编模型以生成与所述目标主题词相关的歌词文档。
第二方面,本申请实施例还提供了一种歌词改编装置,其包括:
获取单元,用于获取待改编的歌词文档,其中,所述待改编的歌词文档包括预设句式模板词、预设主题词、预设关联词以及预设弱关联词;
筛选覆盖单元,用于根据预设方法及预设筛选模型筛选出所述待改编的歌词文档中的所述预设主题词及所述预设关联词,并用预设标记符将所述预设主题词及所述预设关联词进行遮挡覆盖以生成目标歌词模板,其中,所述目标歌词模板包括所述预设句式模板词和所述预设弱关联词;
生成单元,用于将所述目标歌词模板及目标主题词输入预设改编模型以生成与所述目标主题词相关的歌词文档。
第三方面,本申请实施例还提供了一种计算机设备,其包括存储器及处理器,所述存储 器上存储有计算机程序,所述处理器执行所述计算机程序时实现上述方法。
第四方面,本申请实施例还提供了一种计算机可读存储介质,所述存储介质存储有计算机程序,所述计算机程序当被处理器执行时可实现上述方法。
本申请实施例的技术方案,先根据预设方法及预设筛选模型筛选出待改编的歌词文档中的预设主题词及预设关联词并用预设标记符将其遮挡覆盖以生成目标歌词模板,再根据目标歌词模板、目标主题词以及预设改编模型生成与目标主题词相关的歌词文档,由于目标歌词模板保留了待改编歌词文档中固有的结构,因此可在一定程度上提升歌词改编后的歌词质量。
附图说明
为了更清楚地说明本申请实施例技术方案,下面将对实施例描述中所需要使用的附图作简单地介绍,显而易见地,下面描述中的附图是本申请的一些实施例,对于本领域普通技术人员来讲,在不付出创造性劳动的前提下,还可以根据这些附图获得其他的附图。
图1为本申请实施例提供的一种歌词改编方法的场景示意图;
图2为本申请实施例提供的一种歌词改编方法的流程示意图;
图3为本申请实施例提供的一种歌词改编方法的子流程示意图;
图4为本申请实施例提供的一种歌词改编方法的子流程示意图;
图5为本申请实施例提供的一种歌词改编方法的子流程示意图;
图6为本申请实施例提供的一种歌词改编方法的子流程示意图;
图7为本申请实施例提供的一种歌词改编装置的示意性框图;
图8为本申请实施例提供的歌词改编装置的筛选覆盖单元的示意性框图;
图9为本申请实施例提供的歌词改编装置的第一筛选覆盖子单元的示意性框图;
图10为本申请实施例提供的歌词改编装置的第二筛选覆盖子单元的示意性框图;
图11为本申请实施例提供的歌词改编装置的第三筛选覆盖子单元的示意性框图;
图12为本申请实施例提供的歌词改编装置的生成单元的示意性框图;
图13为本申请实施例提供的歌词改编装置的第一生成子单元的示意性框图;以及
图14为本申请实施例提供的一种计算机设备的示意性框图。
具体实施方式
下面将结合本申请实施例中的附图,对本申请实施例中的技术方案进行清楚、完整地描述,显然,所描述的实施例是本申请一部分实施例,而不是全部的实施例。基于本申请中的实施例,本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例,都属于本申请保护的范围。
应当理解,当在本说明书和所附权利要求书中使用时,术语“包括”和“包含”指示所描述特征、整体、步骤、操作、元素和/或组件的存在,但并不排除一个或多个其它特征、整体、步骤、操作、元素、组件和/或其集合的存在或添加。
还应当理解,在此本申请说明书中所使用的术语仅仅是出于描述特定实施例的目的而并不意在限制本申请。如在本申请说明书和所附权利要求书中所使用的那样,除非上下文清楚地指明其它情况,否则单数形式的“一”、“一个”及“该”意在包括复数形式。
还应当进一步理解,在本申请说明书和所附权利要求书中使用的术语“和/或”是指相关联列出的项中的一个或多个的任何组合以及所有可能组合,并且包括这些组合。
如在本说明书和所附权利要求书中所使用的那样,术语“如果”可以依据上下文被解释为“当...时”或“一旦”或“响应于确定”或“响应于检测到”。类似地,短语“如果确定”或“如果检测到[所描述条件或事件]”可以依据上下文被解释为意指“一旦确定”或“响应于确定”或“一旦检测到[所描述条件或事件]”或“响应于检测到[所描述条件或事件]”。
请参阅图1,图1是本申请实施例提供的一种歌词改编方法的场景示意图。本申请实施例的歌词改编方法可应用于服务器中,例如可通过配置于服务器上的软件程序来实现该歌词改编方法。服务器与终端进行通信,以使服务器调用用户通过终端上传的待改编的歌词文档并对待改编的歌词文档进行一系列处理后得到改编后的歌词文档,从而实现歌词的改编。该终端可以为台式电脑、手提电脑、平板电脑等,在此不做具体限制。另外,在图1中,该终端和用户的个数为一个,可以理解的是,在实际应用过程中,该终端和用户的个数可以为多个,图1仅仅起到示意说明的作用。
请参阅图2,图2是本申请实施例提供的一种歌词改编方法的流程示意图。如图2所示,该方法包括以下步骤S100-S120。
S100、获取待改编的歌词文档,其中,所述待改编的歌词文档包括预设句式模板词、预设主题词、预设关联词以及预设弱关联词。
在本申请实施例中,服务器要实现歌词的改编,首先需获取待改编的歌词文档,其中,所述待改编的歌词文档包括预设句式模板词、预设主题词、预设关联词以及预设弱关联词,进而基于该待改编的歌词文档进行一系列处理后才可生成改编后的歌词文档。在本申请实施例中,可由用户通过用户终端上传待改编的歌词文档,具体地,用户可通过用户终端的歌词改编的网页端上传待改编的歌词文档以将所述待改编的歌词文档发送到服务器。其中,在本申请实施例中,所述待改编的歌词文档为文本文档。
需要说明的是,在其它实施例中,待改编的歌词文档也可以为其它类型的文档,例如Word文档。
S110、根据预设方法及预设筛选模型筛选出所述待改编的歌词文档中的所述预设主题词及所述预设关联词,并用预设标记符将所述预设主题词及所述预设关联词进行遮挡覆盖以生成目标歌词模板,其中,所述目标歌词模板包括所述预设句式模板词和所述预设弱关联词。
在本申请实施例中,服务器获取了待改编的歌词文档之后需根据预设方法及预设筛选模型筛选出待改编的歌词文档中的预设主题词及预设关联词并用预设标记符将预设主题词及预设关联词进行遮挡覆盖以生成目标歌词模板。其中,预设方法例如为词性判断方法,词性包括名词、代词、动词、形容词等等。预设标记符例如为空格、#等。
需要说明的是,在本实施例中,可将生成的目标歌词模板上传至区块链中,基于目标歌词模板得到对应的摘要信息,具体来说,摘要信息由目标歌词模板进行散列处理得到,比如利用sha256s算法处理得到。将摘要信息上传至区块链可保证其安全性和对用户的公正透明性。用户设备可以从区块链中下载得该摘要信息,以便查证差异化信息及更新后的预存配置 信息是否被篡改。本示例所指区块链是分布式数据存储、点对点传输、共识机制、加密算法等计算机技术的新型应用模式。区块链(Blockchain),本质上是一个去中心化的数据库,是一串使用密码学方法相关联产生的数据块,每一个数据块中包含了一批次网络交易的信息,用于验证其信息的有效性(防伪)和生成下一个区块。
在某些实施例,例如本实施例中,如图3所示,所述步骤S110可包括以下步骤S111-S112。
S111、根据词性判断方法识别出所述待改编的歌词文档中的非所述预设句式模板词。
S112、将所述待改编的歌词文档中的所有非所述预设句式模板词输入预设筛选模型以筛选出所述预设主题词及所述预设关联词,并用预设标记符将所述预设主题词及预设关联词进行遮挡覆盖以生成目标歌词模板。
在本申请实施例中,根据词性判断方法识别出所述待改编的歌词文档中的非所述预设句式模板词;将所述待改编的歌词文档中的所有非所述预设句式模板词输入预设筛选模型以筛选出所述预设主题词及所述预设关联词,并用预设标记符将所述预设主题词及预设关联词进行遮挡覆盖以生成目标歌词模板;其中,预设句式模板词是指一个句子必须按照一定的模式来组织,以生成这个模式的词即为预设句式模板词。例如古汉语中的判断句,“廉颇者,赵之良将也”中的“者”和“也”即为预设句式模板词。在实际词性判断中,若待改编的歌词文档中的歌词为代词“我”、“你”或者“他”,则可判断为预设句式模板词,预设句式模板词需保留。
在某些实施例,例如本实施例中,如图4所示,所述步骤S112可包括以下步骤S1121-S1122。
S1121、将所述待改编的歌词文档中的所有非所述预设句式模板词输入第一预设筛选模型以筛选出预设主题词,并将所述预设主题词用预设标记符进行遮挡覆盖以生成第一歌词文档。
S1122、将所述第一歌词文档及所述预设主题词输入第二预设模型以筛选出预设关联词,并在所述第一歌词文档中将所述预设关联词用预设标记符进行遮挡覆盖以生成第二歌词文档,将所述第二歌词文档作为目标歌词模板。
在本申请实施例中,将所述待改编的歌词文档中的所有非所述预设句式模板词输入预设筛选模型以筛选出所述预设主题词及所述预设关联词,并用预设标记符将所述预设主题词及预设关联词进行遮挡覆盖以生成目标歌词模板。具体地,将所述待改编的歌词文档中的所有非所述预设句式模板词输入第一预设筛选模型以筛选出预设主题词,并将所述预设主题词用预设标记符进行遮挡覆盖以生成第一歌词文档;将所述第一歌词文档及所述预设主题词输入第二预设模型以筛选出预设关联词,并在所述第一歌词文档中将所述预设关联词用预设标记符进行遮挡覆盖以生成第二歌词文档,将所述第二歌词文档作为目标歌词模板。其中,所述第一预设模型为TF-IDF(Term Frequency-Inverse Document Frequency)模型,TF-IDF模型用以评估一字词对于一个文件集或一份文件对于所在的一个语料库中的重要程度。字词的重要性随着它在文件中出现的次数成正比增加,但同时会随着它在语料库中出现的频率成反比下降;所述第二预设模型为Word2Vec(Word to vector)模型,Word2Vec模型可用来映射 每个词到一个向量,可用来表示词对词之间的关系;所述预设主题词为待改编的歌词文档的主题词;所述预设关联词为与所述主题词强相关联的词,更为具体地,在本实施例中,首先计算提取所述预设句式模板词之后的歌词文档中每个词的词频和每个词在整个歌词文档中的逆文档频率,然后结合词频和逆文档频率计算出每个词对应的TF-IDF分数,并对TF-IDF分数进行排序,选择TF-IDF分数最高的N个词作为预设主题词,其中,N的值根据具体需求而定。在确定预设主题词之后,将筛选出所述预设主题词之后的歌词文档输入Word2Vec模型中,Word2Vec模型会计算每个预设主题词所对应的词向量并进行加权求平均得到歌词文档的主题向量,最后通过计算主题向量与歌词文档中每个词向量的余弦相似性,可以得到预设关联词和预设弱关联词,所述弱关联词为与所述预设主题词弱相关联或不相关联的词。在本实施例中,在经过TF-IDF模型和Word2Vec模型之后,会将预设主题词及预设关联词用预设标记符遮挡覆盖以形成只有预设弱关联词和预设句式模板词的目标歌词模板。
需要说明的是,在其它实施例中,根据实际需求也可采用其它模型对提取预设句式模板词之后的歌词文档进行筛选以筛选出预设弱关联词,例如One-Hot模型和Glove模型。
S120、将所述目标歌词模板及目标主题词输入预设改编模型以生成与所述目标主题词相关的歌词文档。
在本申请实施例中,在经过词性判断方法、TF-IDF模型以及Word2Vec模型之后,会将预设主题词及预设关联词用预设标记符遮挡覆盖以形成只有预设弱关联词和预设句式模板词的目标歌词模板,将目标歌词模板及目标主题词输入预设改编模型以生成与目标主题词相关的歌词文档。
在某些实施例,例如本实施例中,如图5所示,所述步骤S120可包括以下步骤S121-S122。
S121、将所述目标歌词模板及目标主题词输入预设改编模型以生成与所述目标主题词相关的目标主题替换词。
S122、将所述目标主题替换词填充至所述目标歌词模板中以生成与所述目标主题词相关的歌词文档。
在本申请实施例中,在形成目标歌词模板之后,会将所述目标歌词模板及目标主题词输入预设改编模型以生成与所述目标主题词相关的歌词文档。具体地,先将所述目标歌词模板及目标主题词输入预设改编模型以生成与所述目标主题词相关的目标主题替换词;再将所述目标主题替换词填充至所述目标歌词模板中以生成与所述目标主题词相关的歌词文档。可理解地,由于目标歌词模板中预设主题词及预设关联词被预设标记符遮挡覆盖,因此将目标主题替换词填充至目标歌词模板的过程,即是将预设标记符的位置替换成目标主题替换词的过程。需要说明的是,在生成目标主题替换词的过程中,会经过多轮迭代预测,以使生成的目标主题替换词更为准确。最后将目标主题替换词填充至目标歌词模板中以生成与目标主题词相关的歌词文档。
在某些实施例,例如本实施例中,如图6所示,所述步骤S121可包括以下步骤S1211-S1212。
S1211、将所述目标歌词模板及目标主题词输入Bert模型以生成与所述目标主题词相关 的目标主题候选词。
S1212、将所述目标主题候选词及所述目标主题词输入Word2Vec模型以生成与所述目标主题词相关的目标主题替换词。
在本申请实施例中,将所述目标歌词模板及目标主题词输入预设改编模型以生成与所述目标主题词相关的目标主题替换词。其中,预设改编模型包括Bert(Bidirectional Encoder Representations From Transformers)模型和Word2Vec模型。Bert模型是一种基于Transformer采用了双向语言的一种模型,能够提取到文档的语法语义信息,并且还能够结合文档上下文信息进行提取。在本实施例中,基于目标歌词模板及目标主题词并且采用Bert模型整合双向的语义信息生成与目标主题词相关的目标主题候选词,之后再将目标主题候选词和目标主题词输入Word2Vec模型进行相似性比较从而确定目标主题替换词,将目标主题替换词填充至目标歌词模板中即可生成与目标主题词相关的歌词文档,进而实现特定主题的歌词改编。
图7是本申请实施例提供的一种歌词改编装置200的示意性框图。如图7所示,对应于以上歌词改编方法,本申请还提供一种歌词改编装置200。该歌词改编装置200包括用于执行上述歌词改编方法的单元,该装置可以被配置于服务器中。具体地,请参阅图7,该歌词改编装置200包括获取单元201、筛选覆盖单元202以及生成单元203。
其中,获取单元201用于获取待改编的歌词文档,其中,所述待改编的歌词文档包括预设句式模板词、预设主题词、预设关联词以及预设弱关联词;筛选覆盖单元202用于根据预设方法及预设筛选模型筛选出所述待改编的歌词文档中的所述预设主题词及所述预设关联词,并用预设标记符将所述预设主题词及所述预设关联词进行遮挡覆盖以生成目标歌词模板,其中,所述目标歌词模板包括所述预设句式模板词和所述预设弱关联词;生成单元203用于将所述目标歌词模板及目标主题词输入预设改编模型以生成与所述目标主题词相关的歌词文档。
在某些实施例,例如本实施例中,如图8所示,所述筛选覆盖单元202包括判断单元2021及第一筛选覆盖子单元2022。
其中,判断单元2021用于根据词性判断方法识别出所述待改编的歌词文档中的非所述预设句式模板词;第一筛选覆盖子单元2022用于将所述待改编的歌词文档中的所有非所述预设句式模板词输入预设筛选模型以筛选出所述预设主题词及所述预设关联词,并用预设标记符将所述预设主题词及预设关联词进行遮挡覆盖以生成目标歌词模板。
在某些实施例,例如本实施例中,如图9所示,所述第一筛选覆盖子单元2022包括第二筛选覆盖子单元2023及第三筛选覆盖子单元2024。
其中,第二筛选覆盖子单元2023用于将所述待改编的歌词文档中的所有非所述预设句式模板词输入第一预设筛选模型以筛选出预设主题词,并将所述预设主题词用预设标记符进行遮挡覆盖以生成第一歌词文档;第三筛选覆盖子单元2024用于将所述第一歌词文档及所述预设主题词输入第二预设模型以筛选出预设关联词,并在所述第一歌词文档中将所述预设关联词用预设标记符进行遮挡覆盖以生成第二歌词文档,将所述第二歌词文档作为目标歌词模板。
在某些实施例,例如本实施例中,如图10所示,所述第二筛选覆盖子单元2023包括第 四筛选覆盖子单元2025。
其中,第四筛选覆盖子单元2025用于将所述待改编的歌词文档中的所有非所述预设句式模板词输入TF-IDF模型以筛选出预设主题词,并将所述预设主题词用预设标记符进行遮挡覆盖以生成第一歌词文档。
在某些实施例,例如本实施例中,如图11所示,所述第三筛选覆盖子单元2024包括第五筛选覆盖子单元2026。
其中,第五筛选覆盖子单元2026用于将所述第一歌词文档及所述预设主题词输入Word2Vec模型以筛选出预设关联词,并在所述第一歌词文档中将所述预设关联词用预设标记符进行遮挡覆盖以生成第二歌词文档,将所述第二歌词文档作为目标歌词模板。
在某些实施例,例如本实施例中,如图12所示,所述生成单元203包括第一生成子单元2031及第二生成子单元2032。
其中,第一生成子单元2031用于将所述目标歌词模板及目标主题词输入预设改编模型以生成与所述目标主题词相关的目标主题替换词;第二生成子单元2032用于将所述目标主题替换词填充至所述目标歌词模板中以生成与所述目标主题词相关的歌词文档。
在某些实施例,例如本实施例中,如图13所示,所述第一生成子单元2031
包括第三生成子单元2033及第四生成子单元2034。
其中,第三生成子单元2033用于将所述目标歌词模板及目标主题词输入Bert模型以生成与所述目标主题词相关的目标主题候选词;第四生成子单元2034用于将所述目标主题候选词及所述目标主题词输入Word2Vec模型以生成与所述目标主题词相关的目标主题替换词。
上述歌词改编装置可以实现为一种计算机程序的形式,该计算机程序可以在如图14所示的计算机设备上运行。
请参阅图14,图14是本申请实施例提供的一种计算机设备的示意性框图。该计算机设备300为服务器,具体地,服务器可以是独立的服务器,也可以是多个服务器组成的服务器集群。
参阅图14,该计算机设备300包括通过系统总线301连接的处理器302、存储器和网络接口305,其中,存储器可以包括非易失性存储介质303和内存储器304。
该非易失性存储介质303可存储操作系统3031和计算机程序3032。该计算机程序3032被执行时,可使得处理器302执行一种歌词改编方法。
该处理器302用于提供计算和控制能力,以支撑整个计算机设备300的运行。
该内存储器304为非易失性存储介质303中的计算机程序3032的运行提供环境,该计算机程序3032被处理器302执行时,可使得处理器302执行一种歌词改编方法。
该网络接口305用于与其它设备进行网络通信。本领域技术人员可以理解,图14中示出的结构,仅仅是与本申请方案相关的部分结构的框图,并不构成对本申请方案所应用于其上的计算机设备300的限定,具体的计算机设备300可以包括比图中所示更多或更少的部件,或者组合某些部件,或者具有不同的部件布置。
其中,所述处理器302用于运行存储在存储器中的计算机程序3032,以实现如下步骤: 获取待改编的歌词文档,其中,所述待改编的歌词文档包括预设句式模板词、预设主题词、预设关联词以及预设弱关联词;根据预设方法及预设筛选模型筛选出所述待改编的歌词文档中的所述预设主题词及所述预设关联词,并用预设标记符将所述预设主题词及所述预设关联词进行遮挡覆盖以生成目标歌词模板,其中,所述目标歌词模板包括所述预设句式模板词和所述预设弱关联词;将所述目标歌词模板及目标主题词输入预设改编模型以生成与所述目标主题词相关的歌词文档。
在某些实施例,例如本实施例中,处理器302在实现所述根据预设方法及预设筛选模型筛选出所述待改编的歌词文档中的所述预设主题词及所述预设关联词,并用预设标记符将所述预设主题词及所述预设关联词进行遮挡覆盖以生成目标歌词模板步骤时,具体实现如下步骤:根据词性判断方法识别出所述待改编的歌词文档中的非所述预设句式模板词;将所述待改编的歌词文档中的所有非所述预设句式模板词输入预设筛选模型以筛选出所述预设主题词及所述预设关联词,并用预设标记符将所述预设主题词及预设关联词进行遮挡覆盖以生成目标歌词模板。
在某些实施例,例如本实施例中,处理器302在实现所述将所述待改编的歌词文档中的所有非所述预设句式模板词输入预设筛选模型以筛选出所述预设主题词及所述预设关联词,并用预设标记符将所述预设主题词及预设关联词进行遮挡覆盖以生成目标歌词模板步骤时,具体实现如下步骤:将所述待改编的歌词文档中的所有非所述预设句式模板词输入第一预设筛选模型以筛选出预设主题词,并将所述预设主题词用预设标记符进行遮挡覆盖以生成第一歌词文档;将所述第一歌词文档及所述预设主题词输入第二预设模型以筛选出预设关联词,并在所述第一歌词文档中将所述预设关联词用预设标记符进行遮挡覆盖以生成第二歌词文档,将所述第二歌词文档作为目标歌词模板。
在某些实施例,例如本实施例中,处理器302在实现所述将所述待改编的歌词文档中的所有非所述预设句式模板词输入第一预设筛选模型以筛选出预设主题词,并将所述预设主题词用预设标记符进行遮挡覆盖以生成第一歌词文档步骤时,具体实现如下步骤:将所述待改编的歌词文档中的所有非所述预设句式模板词输入TF-IDF模型以筛选出预设主题词,并将所述预设主题词用预设标记符进行遮挡覆盖以生成第一歌词文档。
在某些实施例,例如本实施例中,处理器302在实现所述将所述第一歌词文档及所述预设主题词输入第二预设模型以筛选出预设关联词,并在所述第一歌词文档中将所述预设关联词用预设标记符进行遮挡覆盖以生成第二歌词文档,将所述第二歌词文档作为目标歌词模板步骤时,具体实现如下步骤:将所述第一歌词文档及所述预设主题词输入Word2Vec模型以筛选出预设关联词,并在所述第一歌词文档中将所述预设关联词用预设标记符进行遮挡覆盖以生成第二歌词文档,将所述第二歌词文档作为目标歌词模板。
在某些实施例,例如本实施例中,处理器302在实现所述将所述目标歌词模板及目标主题词输入预设改编模型以生成与所述目标主题词相关的歌词文档步骤时,具体实现如下步骤:将所述目标歌词模板及目标主题词输入预设改编模型以生成与所述目标主题词相关的目标主题替换词;将所述目标主题替换词填充至所述目标歌词模板中以生成与所述目标主题词相关 的歌词文档。
在某些实施例,例如本实施例中,处理器302在实现所述将所述目标歌词模板及目标主题词输入预设改编模型以生成与所述目标主题词相关的目标主题替换词步骤时,具体实现如下步骤:将所述目标歌词模板及目标主题词输入Bert模型以生成与所述目标主题词相关的目标主题候选词;将所述目标主题候选词及所述目标主题词输入Word2Vec模型以生成与所述目标主题词相关的目标主题替换词。
应当理解,在本申请实施例中,处理器302可以是中央处理单元(Central Processing Unit,CPU),该处理器302还可以是其他通用处理器、数字信号处理器(Digital Signal Processor,DSP)、专用集成电路(Application Specific Integrated Circuit,ASIC)、现成可编程门阵列(Field-Programmable Gate Array,FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件等。其中,通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。
本领域普通技术人员可以理解的是实现上述实施例的方法中的全部或部分流程,是可以通过计算机程序来指令相关的硬件来完成。该计算机程序可存储于一存储介质中,该存储介质为计算机可读存储介质。该计算机程序被该计算机系统中的至少一个处理器执行,以实现上述方法的实施例的流程步骤。
因此,本申请还提供一种存储介质。该存储介质可以为计算机可读存储介质。该存储介质存储有计算机程序。该计算机程序被处理器执行时使处理器执行如下步骤:获取待改编的歌词文档,其中,所述待改编的歌词文档包括预设句式模板词、预设主题词、预设关联词以及预设弱关联词;根据预设方法及预设筛选模型筛选出所述待改编的歌词文档中的所述预设主题词及所述预设关联词,并用预设标记符将所述预设主题词及所述预设关联词进行遮挡覆盖以生成目标歌词模板,其中,所述目标歌词模板包括所述预设句式模板词和所述预设弱关联词;将所述目标歌词模板及目标主题词输入预设改编模型以生成与所述目标主题词相关的歌词文档。
在某些实施例,例如本实施例中,所述处理器在执行所述计算机程序而实现所述根据预设方法及预设筛选模型筛选出所述待改编的歌词文档中的所述预设主题词及所述预设关联词,并用预设标记符将所述预设主题词及所述预设关联词进行遮挡覆盖以生成目标歌词模板步骤时,具体实现如下步骤:根据词性判断方法识别出所述待改编的歌词文档中的非所述预设句式模板词;将所述待改编的歌词文档中的所有非所述预设句式模板词输入预设筛选模型以筛选出所述预设主题词及所述预设关联词,并用预设标记符将所述预设主题词及预设关联词进行遮挡覆盖以生成目标歌词模板。
在某些实施例,例如本实施例中,所述处理器在执行所述计算机程序而实现所述将所述待改编的歌词文档中的所有非所述预设句式模板词输入预设筛选模型以筛选出所述预设主题词及所述预设关联词,并用预设标记符将所述预设主题词及预设关联词进行遮挡覆盖以生成目标歌词模板步骤时,具体实现如下步骤:将所述待改编的歌词文档中的所有非所述预设句式模板词输入第一预设筛选模型以筛选出预设主题词,并将所述预设主题词用预设标记符进 行遮挡覆盖以生成第一歌词文档;将所述第一歌词文档及所述预设主题词输入第二预设模型以筛选出预设关联词,并在所述第一歌词文档中将所述预设关联词用预设标记符进行遮挡覆盖以生成第二歌词文档,将所述第二歌词文档作为目标歌词模板。
在某些实施例,例如本实施例中,所述处理器在执行所述计算机程序而实现所述将所述待改编的歌词文档中的所有非所述预设句式模板词输入第一预设筛选模型以筛选出预设主题词,并将所述预设主题词用预设标记符进行遮挡覆盖以生成第一歌词文档步骤时,具体实现如下步骤:将所述待改编的歌词文档中的所有非所述预设句式模板词输入TF-IDF模型以筛选出预设主题词,并将所述预设主题词用预设标记符进行遮挡覆盖以生成第一歌词文档。
在某些实施例,例如本实施例中,所述处理器在执行所述计算机程序而实现所述将所述第一歌词文档及所述预设主题词输入第二预设模型以筛选出预设关联词,并在所述第一歌词文档中将所述预设关联词用预设标记符进行遮挡覆盖以生成第二歌词文档,将所述第二歌词文档作为目标歌词模板步骤时,具体实现如下步骤:将所述第一歌词文档及所述预设主题词输入Word2Vec模型以筛选出预设关联词,并在所述第一歌词文档中将所述预设关联词用预设标记符进行遮挡覆盖以生成第二歌词文档,将所述第二歌词文档作为目标歌词模板。
在某些实施例,例如本实施例中,所述处理器在执行所述计算机程序而实现所述将所述目标歌词模板及目标主题词输入预设改编模型以生成与所述目标主题词相关的歌词文档步骤时,具体实现如下步骤:将所述目标歌词模板及目标主题词输入预设改编模型以生成与所述目标主题词相关的目标主题替换词;将所述目标主题替换词填充至所述目标歌词模板中以生成与所述目标主题词相关的歌词文档。
在某些实施例,例如本实施例中,所述处理器在执行所述计算机程序而实现所述将所述目标歌词模板及目标主题词输入预设改编模型以生成与所述目标主题词相关的目标主题替换词步骤时,具体实现如下步骤:将所述目标歌词模板及目标主题词输入Bert模型以生成与所述目标主题词相关的目标主题候选词;将所述目标主题候选词及所述目标主题词输入Word2Vec模型以生成与所述目标主题词相关的目标主题替换词。
所述存储介质可以是U盘、移动硬盘、只读存储器(Read-Only Memory,ROM)、磁碟或者光盘等各种可以存储程序代码的计算机可读存储介质,所述计算机可读存储介质可以是非易失性,也可以是易失性。
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、计算机软件或者二者的结合来实现,为了清楚地说明硬件和软件的可互换性,在上述说明中已经按照功能一般性地描述了各示例的组成及步骤。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。
在本申请所提供的几个实施例中,应该理解到,所揭露的装置和方法,可以通过其它的方式实现。例如,以上所描述的装置实施例仅仅是示意性的。例如,各个单元的划分,仅仅为一种逻辑功能划分,实际实现时可以有另外的划分方式。例如多个单元或组件可以结合或 者可以集成到另一个系统,或一些特征可以忽略,或不执行。
本申请实施例方法中的步骤可以根据实际需要进行顺序调整、合并和删减。本申请实施例装置中的单元可以根据实际需要进行合并、划分和删减。另外,在本申请各个实施例中的各功能单元可以集成在一个处理单元中,也可以是各个单元单独物理存在,也可以是两个或两个以上单元集成在一个单元中。
该集成的单元如果以软件功能单元的形式实现并作为独立的产品销售或使用时,可以存储在一个存储介质中。基于这样的理解,本申请的技术方案本质上或者说对现有技术做出贡献的部分,或者该技术方案的全部或部分可以以软件产品的形式体现出来,该计算机软件产品存储在一个存储介质中,包括若干指令用以使得一台计算机设备(可以是个人计算机,终端,或者网络设备等)执行本申请各个实施例所述方法的全部或部分步骤。
在上述实施例中,对各个实施例的描述都各有侧重,某个实施例中没有详细描述的部分,可以参见其他实施例的相关描述。
显然,本领域的技术人员可以对本申请进行各种改动和变型而不脱离本申请的精神和范围。这样,尚且本申请的这些修改和变型属于本申请权利要求及其等同技术的范围之内,则本申请也意图包含这些改动和变型在内。
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,可轻易想到各种等效的修改或替换,这些修改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权利要求的保护范围为准。

Claims (20)

  1. 一种歌词改编方法,包括:
    获取待改编的歌词文档,其中,所述待改编的歌词文档包括预设句式模板词、预设主题词、预设关联词以及预设弱关联词;
    根据预设方法及预设筛选模型筛选出所述待改编的歌词文档中的所述预设主题词及所述预设关联词,并用预设标记符将所述预设主题词及所述预设关联词进行遮挡覆盖以生成目标歌词模板,其中,所述目标歌词模板包括所述预设句式模板词和所述预设弱关联词;
    将所述目标歌词模板及目标主题词输入预设改编模型以生成与所述目标主题词相关的歌词文档。
  2. 根据权利要求1所述的方法,其中,所述根据预设方法及预设筛选模型筛选出所述待改编的歌词文档中的所述预设主题词及所述预设关联词,并用预设标记符将所述预设主题词及所述预设关联词进行遮挡覆盖以生成目标歌词模板,包括:
    根据词性判断方法识别出所述待改编的歌词文档中的非所述预设句式模板词;
    将所述待改编的歌词文档中的所有非所述预设句式模板词输入预设筛选模型以筛选出所述预设主题词及所述预设关联词,并用预设标记符将所述预设主题词及预设关联词进行遮挡覆盖以生成目标歌词模板。
  3. 根据权利要求2所述的方法,其中,所述将所述待改编的歌词文档中的所有非所述预设句式模板词输入预设筛选模型以筛选出所述预设主题词及所述预设关联词,并用预设标记符将所述预设主题词及预设关联词进行遮挡覆盖以生成目标歌词模板,包括:
    将所述待改编的歌词文档中的所有非所述预设句式模板词输入第一预设筛选模型以筛选出预设主题词,并将所述预设主题词用预设标记符进行遮挡覆盖以生成第一歌词文档;
    将所述第一歌词文档及所述预设主题词输入第二预设模型以筛选出预设关联词,并在所述第一歌词文档中将所述预设关联词用预设标记符进行遮挡覆盖以生成第二歌词文档,将所述第二歌词文档作为目标歌词模板。
  4. 根据权利要求3所述的方法,其中,所述将所述待改编的歌词文档中的所有非所述预设句式模板词输入第一预设筛选模型以筛选出预设主题词,并将所述预设主题词用预设标记符进行遮挡覆盖以生成第一歌词文档,包括:
    将所述待改编的歌词文档中的所有非所述预设句式模板词输入TF-IDF模型以筛选出预设主题词,并将所述预设主题词用预设标记符进行遮挡覆盖以生成第一歌词文档。
  5. 根据权利要求3所述的方法,其中,所述将所述第一歌词文档及所述预设主题词输入第二预设模型以筛选出预设关联词,并在所述第一歌词文档中将所述预设关联词用预设标记符进行遮挡覆盖以生成第二歌词文档,将所述第二歌词文档作为目标歌词模板,包括:
    将所述第一歌词文档及所述预设主题词输入Word2Vec模型以筛选出预设关联词,并在所述第一歌词文档中将所述预设关联词用预设标记符进行遮挡覆盖以生成第二歌词文档,将所述第二歌词文档作为目标歌词模板。
  6. 根据权利要求1所述的方法,其中,所述将所述目标歌词模板及目标主题词输入预设 改编模型以生成与所述目标主题词相关的歌词文档,包括:
    将所述目标歌词模板及目标主题词输入预设改编模型以生成与所述目标主题词相关的目标主题替换词;
    将所述目标主题替换词填充至所述目标歌词模板中以生成与所述目标主题词相关的歌词文档。
  7. 根据权利要求6所述的方法,其中,所述将所述目标歌词模板及目标主题词输入预设改编模型以生成与所述目标主题词相关的目标主题替换词,包括:
    将所述目标歌词模板及目标主题词输入Bert模型以生成与所述目标主题词相关的目标主题候选词;
    将所述目标主题候选词及所述目标主题词输入Word2Vec模型以生成与所述目标主题词相关的目标主题替换词。
  8. 一种歌词改编装置,包括:
    获取单元,用于获取待改编的歌词文档,其中,所述待改编的歌词文档包括预设句式模板词、预设主题词、预设关联词以及预设弱关联词;
    筛选覆盖单元,用于根据预设方法及预设筛选模型筛选出所述待改编的歌词文档中的所述预设主题词及所述预设关联词,并用预设标记符将所述预设主题词及所述预设关联词进行遮挡覆盖以生成目标歌词模板,其中,所述目标歌词模板包括所述预设句式模板词和所述预设弱关联词;
    生成单元,用于将所述目标歌词模板及目标主题词输入预设改编模型以生成与所述目标主题词相关的歌词文档。
  9. 一种计算机设备,其中,所述计算机设备包括存储器及处理器,所述存储器上存储有计算机程序,所述处理器执行所述计算机程序时实现如下所述的步骤:
    获取待改编的歌词文档,其中,所述待改编的歌词文档包括预设句式模板词、预设主题词、预设关联词以及预设弱关联词;
    根据预设方法及预设筛选模型筛选出所述待改编的歌词文档中的所述预设主题词及所述预设关联词,并用预设标记符将所述预设主题词及所述预设关联词进行遮挡覆盖以生成目标歌词模板,其中,所述目标歌词模板包括所述预设句式模板词和所述预设弱关联词;
    将所述目标歌词模板及目标主题词输入预设改编模型以生成与所述目标主题词相关的歌词文档。
  10. 根据权利要求9所述的计算机设备,其中,所述根据预设方法及预设筛选模型筛选出所述待改编的歌词文档中的所述预设主题词及所述预设关联词,并用预设标记符将所述预设主题词及所述预设关联词进行遮挡覆盖以生成目标歌词模板,包括:
    根据词性判断方法识别出所述待改编的歌词文档中的非所述预设句式模板词;
    将所述待改编的歌词文档中的所有非所述预设句式模板词输入预设筛选模型以筛选出所述预设主题词及所述预设关联词,并用预设标记符将所述预设主题词及预设关联词进行遮挡覆盖以生成目标歌词模板。
  11. 根据权利要求10所述的计算机设备,其中,所述将所述待改编的歌词文档中的所有非所述预设句式模板词输入预设筛选模型以筛选出所述预设主题词及所述预设关联词,并用预设标记符将所述预设主题词及预设关联词进行遮挡覆盖以生成目标歌词模板,包括:
    将所述待改编的歌词文档中的所有非所述预设句式模板词输入第一预设筛选模型以筛选出预设主题词,并将所述预设主题词用预设标记符进行遮挡覆盖以生成第一歌词文档;
    将所述第一歌词文档及所述预设主题词输入第二预设模型以筛选出预设关联词,并在所述第一歌词文档中将所述预设关联词用预设标记符进行遮挡覆盖以生成第二歌词文档,将所述第二歌词文档作为目标歌词模板。
  12. 根据权利要求11所述的计算机设备,其中,所述将所述待改编的歌词文档中的所有非所述预设句式模板词输入第一预设筛选模型以筛选出预设主题词,并将所述预设主题词用预设标记符进行遮挡覆盖以生成第一歌词文档,包括:
    将所述待改编的歌词文档中的所有非所述预设句式模板词输入TF-IDF模型以筛选出预设主题词,并将所述预设主题词用预设标记符进行遮挡覆盖以生成第一歌词文档。
  13. 根据权利要求11所述的计算机设备,其中,所述将所述第一歌词文档及所述预设主题词输入第二预设模型以筛选出预设关联词,并在所述第一歌词文档中将所述预设关联词用预设标记符进行遮挡覆盖以生成第二歌词文档,将所述第二歌词文档作为目标歌词模板,包括:
    将所述第一歌词文档及所述预设主题词输入Word2Vec模型以筛选出预设关联词,并在所述第一歌词文档中将所述预设关联词用预设标记符进行遮挡覆盖以生成第二歌词文档,将所述第二歌词文档作为目标歌词模板。
  14. 根据权利要求9所述的计算机设备,其中,所述将所述目标歌词模板及目标主题词输入预设改编模型以生成与所述目标主题词相关的歌词文档,包括:
    将所述目标歌词模板及目标主题词输入预设改编模型以生成与所述目标主题词相关的目标主题替换词;
    将所述目标主题替换词填充至所述目标歌词模板中以生成与所述目标主题词相关的歌词文档。
  15. 根据权利要求14所述的计算机设备,其中,所述将所述目标歌词模板及目标主题词输入预设改编模型以生成与所述目标主题词相关的目标主题替换词,包括:
    将所述目标歌词模板及目标主题词输入Bert模型以生成与所述目标主题词相关的目标主题候选词;
    将所述目标主题候选词及所述目标主题词输入Word2Vec模型以生成与所述目标主题词相关的目标主题替换词。
  16. 一种计算机可读存储介质,其中,所述存储介质存储有计算机程序,所述计算机程序当被处理器执行时可实现如下所述的步骤:
    获取待改编的歌词文档,其中,所述待改编的歌词文档包括预设句式模板词、预设主题词、预设关联词以及预设弱关联词;
    根据预设方法及预设筛选模型筛选出所述待改编的歌词文档中的所述预设主题词及所述预设关联词,并用预设标记符将所述预设主题词及所述预设关联词进行遮挡覆盖以生成目标歌词模板,其中,所述目标歌词模板包括所述预设句式模板词和所述预设弱关联词;
    将所述目标歌词模板及目标主题词输入预设改编模型以生成与所述目标主题词相关的歌词文档。
  17. 根据权利要求16所述的计算机可读存储介质,其中,所述根据预设方法及预设筛选模型筛选出所述待改编的歌词文档中的所述预设主题词及所述预设关联词,并用预设标记符将所述预设主题词及所述预设关联词进行遮挡覆盖以生成目标歌词模板,包括:
    根据词性判断方法识别出所述待改编的歌词文档中的非所述预设句式模板词;
    将所述待改编的歌词文档中的所有非所述预设句式模板词输入预设筛选模型以筛选出所述预设主题词及所述预设关联词,并用预设标记符将所述预设主题词及预设关联词进行遮挡覆盖以生成目标歌词模板。
  18. 根据权利要求17所述的计算机可读存储介质,其中,所述将所述待改编的歌词文档中的所有非所述预设句式模板词输入预设筛选模型以筛选出所述预设主题词及所述预设关联词,并用预设标记符将所述预设主题词及预设关联词进行遮挡覆盖以生成目标歌词模板,包括:
    将所述待改编的歌词文档中的所有非所述预设句式模板词输入第一预设筛选模型以筛选出预设主题词,并将所述预设主题词用预设标记符进行遮挡覆盖以生成第一歌词文档;
    将所述第一歌词文档及所述预设主题词输入第二预设模型以筛选出预设关联词,并在所述第一歌词文档中将所述预设关联词用预设标记符进行遮挡覆盖以生成第二歌词文档,将所述第二歌词文档作为目标歌词模板。
  19. 根据权利要求18所述的计算机可读存储介质,其中,所述将所述待改编的歌词文档中的所有非所述预设句式模板词输入第一预设筛选模型以筛选出预设主题词,并将所述预设主题词用预设标记符进行遮挡覆盖以生成第一歌词文档,包括:
    将所述待改编的歌词文档中的所有非所述预设句式模板词输入TF-IDF模型以筛选出预设主题词,并将所述预设主题词用预设标记符进行遮挡覆盖以生成第一歌词文档。
  20. 根据权利要求18所述的计算机设备,其中,所述将所述第一歌词文档及所述预设主题词输入第二预设模型以筛选出预设关联词,并在所述第一歌词文档中将所述预设关联词用预设标记符进行遮挡覆盖以生成第二歌词文档,将所述第二歌词文档作为目标歌词模板,包括:
    将所述第一歌词文档及所述预设主题词输入Word2Vec模型以筛选出预设关联词,并在所述第一歌词文档中将所述预设关联词用预设标记符进行遮挡覆盖以生成第二歌词文档,将所述第二歌词文档作为目标歌词模板。
PCT/CN2020/122425 2020-06-30 2020-10-21 歌词改编方法、装置、计算机设备及存储介质 Ceased WO2021135535A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010622899.6 2020-06-30
CN202010622899.6A CN111783413B (zh) 2020-06-30 2020-06-30 歌词改编方法、装置、计算机设备及存储介质

Publications (1)

Publication Number Publication Date
WO2021135535A1 true WO2021135535A1 (zh) 2021-07-08

Family

ID=72760460

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/122425 Ceased WO2021135535A1 (zh) 2020-06-30 2020-10-21 歌词改编方法、装置、计算机设备及存储介质

Country Status (2)

Country Link
CN (1) CN111783413B (zh)
WO (1) WO2021135535A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111783413B (zh) * 2020-06-30 2024-12-31 平安科技(深圳)有限公司 歌词改编方法、装置、计算机设备及存储介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20140358539A1 (en) * 2013-05-29 2014-12-04 Tencent Technology (Shenzhen) Company Limited Method and apparatus for building a language model
CN109815493A (zh) * 2019-01-09 2019-05-28 厦门大学 一种智能嘻哈音乐歌词生成的建模方法
CN111061867A (zh) * 2019-10-29 2020-04-24 平安科技(深圳)有限公司 基于质量感知的文本生成方法、设备、存储介质及装置
CN111259665A (zh) * 2020-01-14 2020-06-09 成都嗨翻屋科技有限公司 一种基于神经网络的交互式歌词生成方法及系统
CN111783413A (zh) * 2020-06-30 2020-10-16 平安科技(深圳)有限公司 歌词改编方法、装置、计算机设备及存储介质

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107122492A (zh) * 2017-05-19 2017-09-01 北京金山安全软件有限公司 基于图片内容的歌词生成方法和装置
CN109815363A (zh) * 2018-12-12 2019-05-28 平安科技(深圳)有限公司 歌词内容的生成方法、装置、终端及存储介质

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20140358539A1 (en) * 2013-05-29 2014-12-04 Tencent Technology (Shenzhen) Company Limited Method and apparatus for building a language model
CN109815493A (zh) * 2019-01-09 2019-05-28 厦门大学 一种智能嘻哈音乐歌词生成的建模方法
CN111061867A (zh) * 2019-10-29 2020-04-24 平安科技(深圳)有限公司 基于质量感知的文本生成方法、设备、存储介质及装置
CN111259665A (zh) * 2020-01-14 2020-06-09 成都嗨翻屋科技有限公司 一种基于神经网络的交互式歌词生成方法及系统
CN111783413A (zh) * 2020-06-30 2020-10-16 平安科技(深圳)有限公司 歌词改编方法、装置、计算机设备及存储介质

Also Published As

Publication number Publication date
CN111783413A (zh) 2020-10-16
CN111783413B (zh) 2024-12-31

Similar Documents

Publication Publication Date Title
CN108984529B (zh) 实时庭审语音识别自动纠错方法、存储介质及计算装置
CN107436922B (zh) 文本标签生成方法和装置
US11017178B2 (en) Methods, devices, and systems for constructing intelligent knowledge base
CN106570180B (zh) 基于人工智能的语音搜索方法及装置
US20240331704A1 (en) Caching scheme for voice recognition engines
CN111026319B (zh) 一种智能文本处理方法、装置、电子设备及存储介质
CN110097085A (zh) 歌词文本生成方法、训练方法、装置、服务器及存储介质
CN104866511B (zh) 一种添加多媒体文件的方法及设备
WO2020186828A1 (zh) 应用程序快捷跳转方法、装置、电子设备和存储介质
CN107992585A (zh) 通用标签挖掘方法、装置、服务器及介质
CN111552799A (zh) 信息处理方法、装置、电子设备及存储介质
CN112100374B (zh) 文本聚类方法、装置、电子设备及存储介质
CN112041809A (zh) 将音效自动添加到音频文件中
CN110109888A (zh) 一种文件处理方法及装置
WO2017121355A1 (zh) 搜索处理方法以及装置
US10902215B1 (en) Social hash for language models
US10664664B2 (en) User feedback for low-confidence translations
CN112099870B (zh) 文档处理方法、装置、电子设备及计算机可读存储介质
WO2021135535A1 (zh) 歌词改编方法、装置、计算机设备及存储介质
CN108460077B (zh) 一种指标分析方法、系统以及计算机可读存储介质
CN110390085B (zh) 文本相似度的分析方法、分析装置、存储介质和电子设备
CN116702770A (zh) 长文本的生成方法、装置、终端及存储介质
CN111597310A (zh) 一种敏感内容检测方法、装置、设备、介质
CN104572906A (zh) 一种事件特征的获取方法和设备
CN118898242B (zh) 合规因子表单动态生成方法、装置、设备及存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20910013

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20910013

Country of ref document: EP

Kind code of ref document: A1