Comment information generation method and device based on large model, electronic equipment and storage medium

Abstract

The invention provides a comment information generation method and device based on a large model, electronic equipment and a storage medium, and relates to the technical field of artificial intelligence, in particular to the technical fields of deep learning, large models, natural language processing and the like. According to the specific scheme, to-be-commented resources are understood based on a large model, and description information of the to-be-commented resources is obtained; according to the description information, comment information of the to-be-commented resource is obtained, the comment information at least comprises a comment video of the to-be-commented resource, and the comment video is displayed in a comment area. According to the invention, intelligent generation of comment videos and texts is realized, the accuracy of comment information can be improved, comment generation operation is simplified, and comment speed is improved. And moreover, by introducing a video comment form, more diversified comment forms are provided for the user, and the user experience is greatly improved.

Description

Technical Field

[0001] The present disclosure relates to the field of artificial intelligence technology, in particular to technical fields such as deep learning, big models, and natural language processing, and specifically to a method, device, electronic device, and storage medium for generating comment information based on a big model. Background Art

[0002] In information flow applications, user comments are an important part of content consumption. In the prior art, user comments are in text form, and the content format of the comment area is relatively simple, which often leads to poor interactivity of the application and fails to effectively increase the time users spend on the application. Summary of the invention

[0003] The present invention provides a method, device, electronic device and storage medium for generating comment information based on a large model.

[0004] According to one aspect of the present disclosure, a method for generating comment information based on a large model is provided, comprising:

[0005] Understand the resources to be reviewed based on the big model and obtain the description information of the resources to be reviewed;

[0006] Acquire comment information of the resource to be commented on according to the description information, wherein the comment information at least includes a comment video of the resource to be commented on;

[0007] Display the comment video in the comment area.

[0008] According to another aspect of the present disclosure, a device for generating comment information based on a large model is provided, comprising:

[0009] The first acquisition module is used to understand the resources to be reviewed based on the big model and obtain the description information of the resources to be reviewed;

[0010] A second acquisition module is used to acquire comment information of the resource to be commented on according to the description information, wherein the comment information at least includes a comment video of the resource to be commented on;

[0011] The comment display module is used to display the comment video in the comment area.

[0012] According to another aspect of the present disclosure, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the large model-based comment information generation method described in the above-mentioned one aspect embodiment.

[0013] According to another aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, on which a computer program / instructions are stored, and the computer instructions are used to enable the computer to execute the method for generating comment information based on a large model as described in the above-mentioned embodiment.

[0014] According to another aspect of the present disclosure, a computer program product is provided, including a computer program / instruction, which, when executed by a processor, implements the method for generating comment information based on a large model as described in the above-mentioned embodiment.

[0015] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0017] Figure 1 A flowchart of a method for generating comment information based on a large model provided in an embodiment of the present disclosure;

[0018] Figure 2 A flowchart of another method for generating comment information based on a large model provided in an embodiment of the present disclosure;

[0019] Figure 2a A schematic diagram of displaying a comment video provided in an embodiment of the present disclosure.

[0020] Figure 3 A flowchart of another method for generating comment information based on a large model provided in an embodiment of the present disclosure;

[0021] Figure 4 A flowchart of another method for generating comment information based on a large model provided in an embodiment of the present disclosure;

[0022] Figure 5 A flowchart of another method for generating comment information based on a large model provided in an embodiment of the present disclosure;

[0023] Figure 6 A schematic diagram of the structure of a large model-based comment information generating device provided in an embodiment of the present disclosure;

[0024] Figure 7 The present invention is a block diagram of an electronic device for implementing the method for generating comment information based on a large model according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0025] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0026] The following describes a method, device, and electronic device for generating comment information based on a large model according to an embodiment of the present disclosure with reference to the accompanying drawings.

[0027] Artificial Intelligence (AI) is a discipline that studies how computers can simulate certain thought processes and intelligent behaviors of human beings (such as learning, reasoning, thinking, planning, etc.). It includes both hardware-level technologies and software-level technologies. AI hardware technologies generally include computer vision technology, speech recognition technology, natural language processing technology, as well as learning / deep learning, big data processing technology, knowledge graph technology, and other aspects.

[0028] Natural Language Processing (NLP) is an important field in the fields of computer science and artificial intelligence. It studies various theories and methods that can achieve effective communication between humans and computers using natural language. Natural language processing is a science that integrates linguistics, computer science, and mathematics. Natural language processing is mainly used in machine translation, public opinion monitoring, automatic summarization, opinion extraction, text classification, question answering, text semantic comparison, speech recognition, etc.

[0029] Deep Learning (DL) is a new research direction in the field of Machine Learning (ML). It is introduced into machine learning to make it closer to its original goal - artificial intelligence. Deep learning is the internal laws and representation levels of learning sample data. The information obtained in the learning process is of great help in the interpretation of data such as text, images and sounds. Its ultimate goal is to enable machines to have analytical learning capabilities like humans and to recognize data such as text, images and sounds. Deep learning is a complex machine learning algorithm that has achieved results in speech and image recognition that far exceed previous related technologies.

[0030] Large models refer to machine learning models with large-scale parameters and complex computational structures. These models are usually built with deep neural networks and have billions or even hundreds of billions of parameters. Large models are designed to improve the expressiveness and predictive performance of the model and to handle more complex tasks and data. Large models are widely used in various fields, including natural language processing, computer vision, speech recognition, and recommendation systems. Large models learn complex patterns and features by training massive amounts of data, have stronger generalization capabilities, and can make accurate predictions on unseen data.

[0031] Figure 1 A flowchart of a method for generating comment information based on a large model provided in an embodiment of the present disclosure.

[0032] like Figure 1 As shown, the method for generating comment information based on the large model may include but is not limited to the following steps:

[0033] S101, understanding the resources to be reviewed based on the big model, and obtaining description information of the resources to be reviewed.

[0034] It should be noted that the execution subject of the comment information generation method in the embodiment of the present disclosure may be a hardware device with data information processing capabilities and / or the necessary software required to drive the hardware device to work. Optionally, the execution subject may include a server, a computer, a user terminal and other intelligent devices. Optionally, the user terminal includes but is not limited to a mobile phone, a computer, an intelligent voice interaction device, etc. Optionally, the server includes but is not limited to a network server, an application server, and may also be a server of a distributed system, or a server combined with a blockchain, etc.

[0035] In some embodiments, the resources to be commented on may be resources on various resource platforms, and the resource platforms may include but are not limited to: news clients, social media platforms, video sharing websites, etc. The resources to be commented on may include but are not limited to: articles, pictures, audio, and video. For example, an article may be a text material on an academic forum. For another example, it may be an image resource published on a social platform. For another example, the audio may be a song, music, radio drama, etc. For another example, the video may be a short video, or it may be a TV series, movie, or other resource.

[0036] In some embodiments, the resource to be commented on can be understood as a resource browsed by a user on a certain platform, for example, it can be a short video currently playing on a social platform.

[0037] In some embodiments, the description information of the resource to be reviewed may include but is not limited to: key information and subject information of the resource to be reviewed. It is understandable that the key information and subject information of the resource to be reviewed can be used as a basis for selecting subsequent review videos and generating review texts.

[0038] In some embodiments, the key information may include, but is not limited to: keywords, titles, and subjects of the resource to be reviewed.

[0039] In some embodiments, semantic analysis and content understanding of the resource to be reviewed can be performed based on a pre-trained large model to extract key information and topic tags of the resource to be reviewed.

[0040] S102: Obtain comment information of the resource to be commented on according to the description information, where the comment information at least includes a comment video of the resource to be commented on.

[0041] S103, displaying the comment video in the comment area.

[0042] In some embodiments, based on the description information and in combination with a pre-trained large model, comment information of the resource to be commented on can be output. In the embodiment of the present disclosure, the comment information at least includes a comment video of the resource to be commented on.

[0043] In some embodiments, the review information of the resource to be reviewed may also include review text, that is, the review information of the resource to be reviewed may be a combination of a review video and a review text.

[0044] In some embodiments, according to the description information, a review video of the resource to be reviewed is obtained from the video library. Optionally, the resource to be reviewed is the i-th episode of a TV series, and the review video can be the i+1th episode, or the -1th episode of the same TV series. Optionally, the review video can include but is not limited to other viewpoint videos under the same event, background extension videos of the event, etc.

[0045] In some embodiments, after obtaining the review video, a first review text of the resource to be reviewed can be generated according to the resource to be reviewed and the review video. It can be understood that the first review text is a transitional review text, which can be used to guide the user to transition from the resource to be reviewed to the review video, thereby improving the coherence and readability of the review.

[0046] In some embodiments, a second review text of the resource to be reviewed may be generated based on the description information, and a review video of the resource to be reviewed may be generated based on the second review text.

[0047] In some embodiments, the comment area corresponding to the user may include at least two types of comment display areas, for example, may include but are not limited to a text display area and a video display area.

[0048] In some embodiments, after obtaining the review information of the resource to be reviewed, each type of review content in the review information may be filled into the corresponding display area.

[0049] According to the comment information generation method of the embodiment of the present disclosure, the intelligent generation of comment videos and texts is realized, which can improve the accuracy and intelligence level of comment information. Furthermore, by introducing the video comment form, more diverse comment forms are provided to users for easy selection, which greatly improves the user experience. Moreover, by intelligently generating comment information based on a large model, the comment generation operation is simplified and the complexity of the comment is reduced, which can not only improve the speed of commenting, but also increase the amount of comments for resources, thereby making it easier to push the resources to be commented on.

[0050] Figure 2 A flowchart of a method for generating comment information based on a large model provided in an embodiment of the present disclosure.

[0051] like Figure 2 As shown, the method for generating comment information based on the large model may include but is not limited to the following steps:

[0052] S201, obtaining description information of the resource to be reviewed.

[0053] The optional implementation of step S201 can be found in Figure 1 Optional implementations of step S101, and Figure 1 Other related parts in the embodiments involved will not be described in detail here.

[0054] S202, obtaining a review video of the resource to be reviewed from a video library according to the description information.

[0055] In some embodiments, candidate videos related to the resource to be reviewed are obtained from a video library based on the description information, and further, a video with the highest relevance is selected from the videos as a review video for the resource to be reviewed.

[0056] Optionally, the description information of each video in the video library is obtained, and the correlation is calculated based on the description information of the resource to be evaluated and the description information of the video, so as to obtain the correlation between the resource to be commented on and each video. Further, the videos in the database are screened based on the correlation of the videos to obtain candidate videos related to the resource to be commented on. Optionally, videos whose correlation is greater than or equal to a set value are selected as candidate videos.

[0057] Optionally, the candidate videos are sorted from high to low, and the candidate video ranked at the top is further selected as the comment video for the video to be commented on, so that the comment video is most matched with the resource to be commented on, which can make the comment video's evaluation of the resource to be commented on more accurately and improve the user experience.

[0058] In the disclosed embodiment, the comment videos obtained from the video library may include the next episode of a film or TV series, other viewpoint videos on the same event, event background extension videos, etc., which can not only enhance the interactivity and attractiveness of the comments, but also realize the presentation of multimodal comments. Through the multimodal comment format, multi-dimensional comments on the comment resources can be realized, making the evaluation of the resources to be commented on more comprehensive and accurate.

[0059] S203: Generate a first comment text of the resource to be commented on according to the resource to be commented on and the comment video.

[0060] In some embodiments, after obtaining the review video, speech analysis and content understanding are performed on the review resource and the review video based on the large language model, and a first review text is generated based on the understood content. It is understandable that the first review text is a transitional review text, which can be used to guide the user to transition from the resource to be reviewed to the review video, thereby improving the coherence and readability of the review. Optionally, the first review text may also include a brief evaluation or statement of opinion on the resource to be reviewed and the review video, thereby further commenting on the review video, so that the user can better understand the content of the resource to be reviewed through the comments on the review video.

[0061] In some embodiments, the user's historical comment data is obtained, which may include but is not limited to the historical resources that have been commented on, and the corresponding comment content, etc. Further, the generated first comment text is polished according to the historical comment data and the public knowledge base. In other words, the first comment text is polished in combination with the user's historical comment data and the public knowledge base, so that the expression logic, sentences and grammar of the first comment text not only fit the human expression mode, but also can meet the user's comment habits, thereby improving the quality of the comment and making the comment more accurate.

[0062] In some embodiments, the expression logic, sentences, grammar and other contents of the first comment text can be analyzed in combination with the public knowledge base. If there are abnormalities in the above aspects of the first comment text, the above aspects are adjusted and optimized to obtain an optimized comment text, so that the expression logic, sentences and grammar of the optimized comment text are consistent with human expression patterns. Further, based on historical comment data, the user's comment like information can be analyzed, for example, it can include but is not limited to comment style, comment length and other information. And based on the comment like information, the optimized comment text is optimized in style and length to obtain the final first comment text.

[0063] S204, displaying the comment video and the first comment text in the comment area.

[0064] In some embodiments, the comment area corresponding to the user may include a text display area and a video display area.

[0065] In some embodiments, after obtaining the review information of the resource to be reviewed, each type of review content in the review information can be filled into the corresponding display area, that is, the review video is filled into the video display area, and the first review text is filled into the text display area.

[0066] In some embodiments, in response to other users sliding to comment on the video, the comment video can be played.

[0067] In some embodiments, other users can click the play button of the comment video to trigger the play of the comment video, such as Figure 2a shown.

[0068] It can be understood that in the embodiment of the present application, the comment area in the front end not only has a text display area, but also adds a video display area, which can support the display and playback of comment videos and ensure that multimodal video comments are displayed normally to users.

[0069] In the disclosed embodiment, compared with users manually searching or entering comment content, a large model is used to deeply understand the content of the resource to be evaluated, and comment videos that match the resource to be evaluated are selected from a massive video library, that is, comment videos related to the resource to be evaluated can be automatically recommended to users. At the same time, high-quality transitional comment texts are generated in combination with the resource to be evaluated and the comment video to improve the coherence and readability of the comment. In this embodiment, the intelligent generation of comment videos and comment texts can improve the accuracy and intelligence level of comment information. Furthermore, by introducing the form of video comments, users are provided with more diverse comment forms to facilitate user selection, which greatly improves the user experience. Moreover, by intelligently generating comment information based on a large model, the comment generation operation is simplified and the complexity of the comment is reduced, which can not only increase the speed of commenting, but also increase the amount of comments for resources, thereby making it easier to push the resources to be commented on.

[0070] Figure 3 A flowchart of a method for generating comment information based on a large model provided in an embodiment of the present disclosure.

[0071] like Figure 3 As shown, the method for generating comment information based on the large model may include but is not limited to the following steps:

[0072] S301, obtaining description information of the resource to be reviewed.

[0073] The optional implementation of step S301 can be found in Figure 1 Optional implementations of step S101, and Figure 1 Other related parts in the embodiments involved will not be described in detail here.

[0074] S302: Generate a second review text for the resource to be reviewed according to the description information.

[0075] In some embodiments, the user's historical comment data is obtained, which may include but is not limited to the historical resources that have been commented on, and the corresponding comment content, etc. Further, a second comment text is generated based on the description information, historical comment data, and public knowledge base. In other words, a deep understanding of the comment resource is combined with information such as the user's historical comment data and the public knowledge base, and a large model is used to generate a second comment text related to the resource to be commented on, so that the expression logic, sentences, and grammar of the second comment text not only fit the human expression mode, but also can meet the user's comment habits, thereby improving the quality of the comment and making the comment more accurate.

[0076] In some embodiments, an initial third comment text is generated based on the description information, and further, the initial third comment text is polished based on the historical comment data and the public knowledge base to generate a second comment text. Optionally, the expression logic, sentences, grammar and other contents of the initial third comment text can be analyzed in combination with the public knowledge base. When there are abnormalities in the above aspects of the initial third comment text, the above aspects are adjusted and optimized to obtain an optimized fourth comment text, so that the expression logic, sentences and grammar of the fourth comment text are consistent with human expression patterns. Further, based on the historical comment data, the user's comment like information can be analyzed, for example, it can include but is not limited to comment style, comment length and other information. And based on the comment like information, the optimized fourth comment text is optimized in style and length to obtain the final second comment text.

[0077] Optionally, the second comment text may be an explanation of the viewpoint, emotional expression or extended discussion of the resource to be evaluated, which can further express the user's emotions.

[0078] S303: Generate a review video of the resource to be reviewed based on the second review text.

[0079] In some embodiments, the second comment text can be input into a pre-trained text-to-video generation model to obtain a comment video of the resource to be commented on. The comment video is generated by the second comment text, so that the acquisition of the comment video is no longer based on the existing video resource, which can enrich the content of the comment video and reduce the probability of duplication of the comment video of the same resource to be commented on.

[0080] In some embodiments, the second comment text can be input into a pre-trained text-to-video generation model to obtain an initial first video. Furthermore, in order to increase the display effect and quality of the comment video, at least one of background audio and special effects can be added to the first video to obtain a second video, and the second video can be enhanced to obtain a comment video of the resource to be commented on. In other words, only the background audio of the first video can be added, or only the characteristics of the first video can be added, or both the background audio and characteristics of the first video can be added. Optionally, the background audio can be background music, dubbing, or narration, etc.

[0081] In some embodiments, according to at least one of the first video and the second comment text, target background audio and target special effects are selected from candidate background audio and candidate special effects, and added to the first video to obtain the second video. That is, at least one of the first video and the second comment text is understood based on the large model, and the target background audio and target special effects are selected from the candidate background audio and candidate special effects in combination with the understood content.

[0082] In some embodiments, a user's selection operation is received, and target background audio and target special effects are determined based on the selection operation and added to the first video to obtain a second video.

[0083] In some embodiments, historical or current operations of image quality enhancement are obtained, and target image quality parameters corresponding to the first video are determined based on the historical or current operations of image quality enhancement. Further, the image quality of the second video is enhanced based on the target image quality parameters to obtain a comment video for the resource to be commented on.

[0084] In some embodiments, based on historical operations of image quality enhancement, the image quality that the user likes can be determined, and the most frequently used image quality parameters can be selected from the historical operations as target image quality parameters, so that the image quality is more in line with the user's habits, which not only improves the quality of the comment video, but also improves the user experience.

[0085] In some embodiments, initial image quality parameters of the first video are obtained, and in response to the initial image quality parameters not meeting image quality requirements, image quality enhancement is performed on the second video to obtain a review video of the resource to be reviewed.

[0086] In the disclosed embodiment, by adding background audio and features to the commentary video, the elements of the commentary video are made more comprehensive and the image quality can be enhanced, which not only increases the quality of the commentary video but also enables the commentary video to have a higher level of intelligence.

[0087] S304, displaying the comment video and the second comment text in the comment area.

[0088] The optional implementation of step S304 can be found in Figure 2 Optional implementations of step S204, and Figure 1 Other related parts in the embodiments involved will not be described in detail here.

[0089] In the disclosed embodiment, compared with users manually searching or entering comment content, a large model is used to deeply understand the content of the resource to be evaluated, generate high-quality comment text, and then generate a comment video matching the resource to be evaluated based on the comment text. In this embodiment, the video is generated through text, which makes the generation of comment videos more efficient. Especially on social media platforms, the number of comments is often related to the popularity of resources. Intelligent comments can simplify the comment operation, increase the comment speed, and quickly increase the number of comments, thereby making it easier to push the resources to be commented on. Moreover, the comment video can be optimized to further improve the quality of the comment video, thereby improving user satisfaction and experience. Furthermore, by introducing the form of video comments, users are provided with more diverse comment forms to facilitate user selection, further improving the user's comment experience.

[0090] It is understandable that the above two comment video generation methods can be freely selected by the user, and a selection operation can be set in the front-end interface. Through the selection operation, the user can select the generation method he wants to use from the two comment video generation methods, thereby increasing the user's selection freedom and improving the user experience. Alternatively, when no suitable comment video is selected from the video library, the comment video can be generated through the second comment text. Generating videos through text can not only generate high-quality comment videos, but also make the comment video generation process more simplified and efficient, thereby improving the intelligence level of the comment video.

[0091] Based on the above embodiment, the comment information of the resource to be commented on can be stored, such as Figure 4 As shown, the storage process of the comment information of the resource to be commented on may include but is not limited to the following steps:

[0092] S401, determining a target storage node for the comment information from storage nodes in a distributed storage system based on a load balancing strategy.

[0093] In some embodiments, the backend server may store the comment video in a distributed storage system, which may include multiple storage nodes, and may determine a target storage node suitable for storing the comment information of the resource to be commented on from the multiple storage nodes according to a load balancing strategy. For example, load balancing may be performed based on the remaining storage space of the storage node to obtain the target storage node. The distributed storage system may ensure the storage security of the comment information of the resource to be commented on, and avoid loss due to competition for storage space.

[0094] S402, obtaining current network status information, and determining conversion information of the comment information according to the network status information.

[0095] S403: Send the review information of the resource to be reviewed to the target storage node.

[0096] In some embodiments, the current network status information is obtained, and the conversion information of the comment information is determined based on the network status information. Optionally, the network status information may include but is not limited to network bandwidth, traffic size, network bandwidth occupancy rate, etc. Optionally, the conversion information of the comment information may include but is not limited to format conversion information and compression type, etc.

[0097] Further, the comment information is converted and compressed based on the conversion information to obtain target comment information, and the target comment information is sent to the target storage node.

[0098] It is understandable that, in subsequent browsing of the comment information of the resource to be commented on, the relevant target comment information can be read from the target storage node and restored to obtain the original comment information of the resource to be commented on.

[0099] In the disclosed embodiment, the storage security of comment information can be ensured through distributed storage, and the appropriate format can be selected for conversion and compression based on the network and device conditions, thereby saving resources during uploading and further ensuring the security of comment information during the uploading process.

[0100] Figure 5 A flowchart of a method for generating comment information based on a large model provided in an embodiment of the present disclosure.

[0101] like Figure 5 As shown, the method for generating comment information based on the large model may include but is not limited to the following steps:

[0102] S501, obtaining description information of the resource to be reviewed.

[0103] The optional implementation of step S501 can be found in Figure 1 Optional implementations of step S101, and Figure 1 Other related parts in the embodiments involved will not be described in detail here.

[0104] After the description information is obtained, branch 1 steps S502 to 504 may be executed to determine the review information of the resource to be reviewed, or branch 2 steps 505 to 509 may be executed to determine the review information of the resource to be reviewed.

[0105] S502: Obtain candidate videos related to the resource to be reviewed from a video library according to the description information.

[0106] S503: Select the video with the highest relevance from the candidate videos as the review video for the resource to be reviewed.

[0107] S504: Generate a first comment text of the resource to be commented on according to the resource to be commented on and the comment video.

[0108] The optional implementation of steps S502 to S504 can be found in Figure 2 Optional implementations of steps S202 to S203, and Figure 2 Other related parts in the embodiments involved will not be described in detail here.

[0109] S505: Generate an initial third comment text according to the description information.

[0110] S506: Based on the historical review data and the public knowledge base, the third review text is polished to generate a second review text.

[0111] S507: Generate an initial first video based on the second comment text.

[0112] S508, adding background audio and special effects to the first video to obtain a second video.

[0113] S509: Enhance the image quality of the second video to obtain a comment video of the resource to be commented on.

[0114] The optional implementation of steps S505 to S509 can be found in Figure 3 Optional implementations of steps S302 to S303, and Figure 3 Other related parts in the embodiments involved will not be described in detail here.

[0115] S510: Displaying comment information of the resource to be commented on in the comment area.

[0116] Optional implementations of step S510 can be found in Figure 2 Optional implementations of step S204, and Figure 2 Other related parts in the embodiments involved will not be described in detail here.

[0117] S511, determining a target storage node for the comment information from storage nodes in a distributed storage system based on a load balancing strategy.

[0118] S512, obtaining current network status information, and determining conversion and compression of comment information according to the network status information.

[0119] S513: Send the review information of the resource to be reviewed to the target storage node.

[0120] In the disclosed embodiment, the determination of comment videos and text generation are realized, which can improve the accuracy and intelligence level of comment information. Furthermore, by introducing the video comment form, more diverse comment forms are provided to users for easy selection, which greatly improves the user experience. Moreover, by intelligently generating comment information based on a large model, the comment generation operation is simplified and the complexity of comments is reduced, which can not only improve the speed of comments, but also increase the amount of comments for resources, thereby making it easier to push the resources to be commented on.

[0121] Corresponding to the big model-based comment information generating methods provided in the above-mentioned embodiments, an embodiment of the present disclosure also provides a big model-based comment information generating device. Since the information processing device provided in the embodiment of the present disclosure corresponds to the big model-based comment information generating methods provided in the above-mentioned embodiments, the implementation method of the big model-based comment information generating method is also applicable to the information processing device provided in the embodiment of the present disclosure, and will not be described in detail in the following embodiments.

[0122] Figure 6 A schematic diagram of the structure of a large model-based comment information generating device provided in an embodiment of the present disclosure.

[0123] like Figure 6 As shown, the large model-based comment information generating device 600 of the embodiment of the present disclosure includes a first acquisition module 601 , a second acquisition module 602 and a comment display module 603 .

[0124] The first acquisition module 601 is used to acquire description information of the resource to be reviewed;

[0125] A second acquisition module 602 is used to acquire comment information of the resource to be commented on according to the description information, wherein the comment information at least includes a comment video of the resource to be commented on;

[0126] The comment display module 603 is used to display the comment video in the comment area.

[0127] In some embodiments, the second acquisition module 602 is further configured to acquire a review video of the resource to be reviewed from a video library according to the description information.

[0128] In some embodiments, the second acquisition module 602 is further used to acquire candidate videos related to the resource to be commented on from the video library according to the description information; and select the video with the highest correlation from the candidate videos as the comment video for the resource to be commented on.

[0129] In some embodiments, the second acquisition module 602 is further used to generate a first comment text for the resource to be commented on according to the resource to be commented on and the comment video after acquiring the comment video from the video library according to the description information.

[0130] In some embodiments, the second acquisition module 602 is further used to generate a second review text of the resource to be reviewed according to the description information; and generate a review video of the resource to be reviewed according to the second review text.

[0131] In some embodiments, the second acquisition module 602 is further used to acquire the user's historical comment data and a public knowledge base; and polish the first comment text or the second comment text according to the historical comment data and the public knowledge base.

[0132] In some embodiments, the second acquisition module 602 is further used to generate an initial third comment text according to the description information; and polish the third comment text based on the historical comment data and the public knowledge base to generate the second comment text.

[0133] In some embodiments, the second acquisition module 602 is also used to input the second comment text into a pre-trained text-to-video generation model to obtain an initial first video; add at least one of background audio and special effects to the first video to obtain a second video; and enhance the image quality of the second video to obtain a comment video of the resource to be commented on.

[0134] In some embodiments, the second acquisition module 602 is also used to select target background audio from candidate background audios and / or select target special effects from candidate special effects based on at least one content in the first video and the second comment text, and add them to the first video to obtain the second video.

[0135] In some embodiments, the second acquisition module 602 is further used to receive a selection operation, determine the target background audio and target special effects based on the selection operation, and add them to the first video to obtain the second video.

[0136] In some embodiments, the second acquisition module 602 is also used to obtain historical operations or current operations of image quality enhancement; based on the historical operations or current operations of image quality enhancement, determine the target image quality parameters corresponding to the first video; based on the target image quality parameters, enhance the image quality of the second video to obtain a comment video of the resource to be commented on.

[0137] In some embodiments, the second acquisition module 602 is further used to obtain initial image quality parameters of the first video, and in response to the initial image quality parameters not meeting image quality requirements, perform image quality enhancement on the second video to obtain a review video of the resource to be reviewed.

[0138] In some embodiments, the comment display module 603 is further used to fill the comment video into the video display area of ​​the comment area.

[0139] In some embodiments, the second acquisition module 602 is further used to determine a target storage node for the comment information from storage nodes in a distributed storage system based on a load balancing strategy, and send the comment information of the resource to be commented on to the target storage node.

[0140] In some embodiments, the second acquisition module 602 is also used to obtain network status information, and determine the target format and compression information of the comment video in the comment information based on the network status information; convert and compress the comment video based on the target format and compression information to obtain target comment information, and send the target comment information to the target storage node.

[0141] According to the comment information generation device of the embodiment of the present disclosure, the determination of the comment video and the generation of the text are realized, which can improve the accuracy and intelligence level of the comment information. Furthermore, by introducing the video comment form, a more diverse comment form is provided to users for easy selection, which greatly improves the user experience. Moreover, by intelligently generating comment information based on a large model, the comment generation operation is simplified and the complexity of the comment is reduced, which can not only improve the speed of the comment, but also increase the amount of comments for resources, thereby making it easier to push the resources to be commented on.

[0142] In the technical solution disclosed herein, the acquisition, storage and application of user personal information involved are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0143] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0144] Figure 7A schematic block diagram of an example electronic device 700 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0145] like Figure 7 As shown, the device 700 includes a computing unit 701, which can perform various appropriate actions and processes according to a computer program / instruction stored in a read-only memory (ROM) 702 or a computer program / instruction loaded from a storage unit 706 to a random access memory (RAM) 703. In the RAM 703, various programs and data required for the operation of the device 700 can also be stored. The computing unit 701, the ROM 702, and the RAM 703 are connected to each other via a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0146] A number of components in the device 700 are connected to the I / O interface 705, including: an input unit 706 such as a keyboard, a mouse, etc.; an output unit 707 such as various types of displays, speakers, etc.; a storage unit 708 such as a disk, an optical disk, etc.; and a communication unit 709 such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 709 allows the device 700 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0147] The computing unit 701 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 701 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 701 performs the various methods and processes described above, such as a method for generating comment information based on a large model. For example, in some embodiments, the method for generating comment information based on a large model may be implemented as a computer software program, which is tangibly included in a machine-readable medium, such as a storage unit 706. In some embodiments, part or all of the computer program / instructions may be loaded and / or installed on the device 700 via the ROM 702 and / or the communication unit 709. When the computer program / instructions are loaded into the RAM 703 and executed by the computing unit 701, one or more steps of the method for generating comment information based on a large model described above may be performed. Alternatively, in other embodiments, the computing unit 701 may be configured to execute the large model-based comment information generating method in any other appropriate manner (eg, by means of firmware).

[0148] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs / instructions that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0149] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0150] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0151] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0152] The systems and techniques described herein can be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), the Internet, and a blockchain network.

[0153] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs / instructions running on respective computers and having a client-server relationship with each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.

[0154] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in the disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in the disclosure can be achieved, and this document does not limit them here.

[0155] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A method for generating review information based on a large model, wherein: The method comprises: Understand the resources to be reviewed based on the big model and obtain the description information of the resources to be reviewed; Acquire comment information of the resource to be commented on according to the description information, wherein the comment information at least includes a comment video of the resource to be commented on; Display the comment video in the comment area.

2. The method according to claim 1, wherein: The step of obtaining the review information of the resource to be reviewed according to the description information includes: According to the description information, a review video of the resource to be reviewed is obtained from a video library.

3. The method according to claim 2, wherein: The step of obtaining a review video of the resource to be reviewed from a video library according to the description information includes: According to the description information, obtaining candidate videos related to the resource to be reviewed from the video library; A video with the highest relevance is selected from the candidate videos as a review video for the resource to be reviewed.

4. The method according to claim 2, wherein: After obtaining the comment video from the video library according to the description information, the method further includes: A first comment text of the resource to be commented on is generated according to the resource to be commented on and the comment video.

5. The method according to claim 1, wherein: The step of obtaining the review information of the resource to be reviewed according to the description information includes: Generating a second review text of the resource to be reviewed according to the description information; A review video of the resource to be reviewed is generated according to the second review text.

6. The method according to claim 4 or 5, wherein: The method further comprises: Obtain users’ historical review data and public knowledge base; The first comment text or the second comment text is polished according to the historical comment data and the public knowledge base.

7. The method according to claim 6, wherein: The generation process of the second comment text includes: Generating an initial third comment text according to the description information; Based on the historical review data and the public knowledge base, the third review text is polished to generate the second review text.

8. The method according to claim 5, wherein: Generating a review video of the resource to be reviewed according to the second review text includes: Inputting the second comment text into a pre-trained text-to-video generation model to obtain an initial first video; Add at least one of background audio and special effects to the first video to obtain a second video; The image quality of the second video is enhanced to obtain a comment video of the resource to be commented on.

9. The method according to claim 8, wherein: The adding at least one of background audio and special effects to the first video to obtain the second video includes: According to at least one of the first video and the second comment text, target background audio is selected from candidate background audios, and / or target special effects are selected from candidate special effects and added to the first video to obtain the second video.

10. The method according to claim 8, wherein: The adding at least one of background audio and special effects to the first video to obtain the second video includes: A selection operation is received, and the target background audio and target special effects are determined based on the selection operation, and are added to the first video to obtain the second video.

11. The method according to claim 8, wherein: The step of enhancing the image quality of the second video to obtain a comment video of the resource to be commented on includes: Get the historical or current operations of image quality enhancement; Determining a target image quality parameter corresponding to the first video based on a historical operation or a current operation of the image quality enhancement; The image quality of the second video is enhanced based on the target image quality parameter to obtain a comment video of the resource to be commented on.

12. The method according to claim 8 or 11, wherein: The step of enhancing the image quality of the second video to obtain a comment video of the resource to be commented on includes: Initial image quality parameters of the first video are obtained, and in response to the initial image quality parameters not meeting image quality requirements, image quality enhancement is performed on the second video to obtain a review video of the resource to be reviewed.

13. The method according to any one of claims 1 to 5 or 8 to 11, wherein: The displaying of the comment video in the comment area includes: The video display area of ​​the comment area is filled with the comment video.

14. The method according to any one of claims 1 to 5 or 8 to 11, wherein: The method further comprises: Based on the load balancing strategy, a target storage node for the review information is determined from storage nodes in the distributed storage system, and the review information of the resource to be reviewed is sent to the target storage node.

15. The method according to claim 14, wherein: The sending the review information of the resource to be reviewed to the target storage node includes: Acquire network status information, and determine the target format and compression information of the comment video in the comment information according to the network status information; The comment video is converted and compressed based on the target format and compression information to obtain target comment information, and the target comment information is sent to the target storage node.

16. A review information generating device based on a large model, wherein: The device comprises: The first acquisition module is used to understand the resources to be reviewed based on the big model and obtain the description information of the resources to be reviewed; A second acquisition module is used to acquire comment information of the resource to be commented on according to the description information, wherein the comment information at least includes a comment video of the resource to be commented on; The comment display module is used to display the comment video in the comment area.

17. The device according to claim 16, wherein: The second acquisition module is further used for: According to the description information, a review video of the resource to be reviewed is obtained from a video library.

18. The device according to claim 17, wherein: The second acquisition module is further used for: According to the description information, obtaining candidate videos related to the resource to be reviewed from the video library; A video with the highest relevance is selected from the candidate videos as a review video for the resource to be reviewed.

19. The device according to claim 17, wherein: The second acquisition module is further used for: A first comment text of the resource to be commented on is generated according to the resource to be commented on and the comment video.

20. The device according to claim 16, wherein: The second acquisition module is further used for: Generating a second review text of the resource to be reviewed according to the description information; A review video of the resource to be reviewed is generated according to the second review text.

21. The device according to claim 19 or 20, wherein: The second acquisition module is further used for: Obtain users’ historical review data and public knowledge base; The first comment text or the second comment text is polished according to the historical comment data and the public knowledge base.

22. The device according to claim 21, wherein The second acquisition module is further used for: Generating an initial third comment text according to the description information; Based on the historical review data and the public knowledge base, the third review text is polished to generate the second review text.

23. The device according to claim 20, wherein: The second acquisition module is further used for: Inputting the second comment text into a pre-trained text-to-video generation model to obtain an initial first video; Add at least one of background audio and special effects to the first video to obtain a second video; The image quality of the second video is enhanced to obtain a comment video of the resource to be commented on.

24. The device according to claim 23, wherein: The second acquisition module is further used for: According to at least one of the first video and the second comment text, target background audio is selected from candidate background audios, and / or target special effects are selected from candidate special effects and added to the first video to obtain the second video.

25. The device according to claim 23, wherein: The second acquisition module is further used for: A selection operation is received, and the target background audio and target special effects are determined based on the selection operation, and are added to the first video to obtain the second video.

26. The device according to claim 23, wherein The second acquisition module is further used for: Get the historical or current operations of image quality enhancement; Determining a target image quality parameter corresponding to the first video based on a historical operation or a current operation of the image quality enhancement; The image quality of the second video is enhanced based on the target image quality parameter to obtain a comment video of the resource to be commented on.

27. The device according to claim 23 or 26, wherein: The second acquisition module is further used for: Initial image quality parameters of the first video are obtained, and in response to the initial image quality parameters not meeting image quality requirements, image quality enhancement is performed on the second video to obtain a review video of the resource to be reviewed.

28. The device according to any one of claims 16-20 or 23-26, wherein The comment display module is also used for: The video display area of ​​the comment area is filled with the comment video.

29. The device according to any one of claims 16-20 or 23-26, wherein The second acquisition module is further used for: Based on the load balancing strategy, a target storage node for the review information is determined from storage nodes in the distributed storage system, and the review information of the resource to be reviewed is sent to the target storage node.

30. The device according to claim 29, wherein: The second acquisition module is further used for: Acquire network status information, and determine the target format and compression information of the comment video in the comment information according to the network status information; The comment video is converted and compressed based on the target format and compression information to obtain target comment information, and the target comment information is sent to the target storage node.

31. An electronic device comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 15.

32. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-15.

33. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the method steps of any one of claims 1-15 are implemented.

Images