Portrayal model processing method and device, electronic equipment and computer storage medium

Abstract

The invention provides a portrait model processing method and device, electronic equipment and a computer storage medium. The portrait model processing method comprises the following steps: acquiring basic data; wherein the basic data comprises product information; inputting the basic data into a pre-trained portrait model, and outputting a first basic portrait; wherein the portrait model is trained and determined based on a natural language processing model; taking the cleaned first basic portrait as a customer portrait; the data is analyzed through the natural language processing model, the customer portrait is determined, the accuracy of the portrait is improved, the cost of training the model is reduced, and the user experience is improved.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a portrait model processing method, device, electronic equipment and computer storage medium. Background Art

[0002] In recent years, with the continuous development and rapid improvement of the Internet, traffic costs are becoming more and more expensive, the cost of attracting new customers is getting higher and higher, and the loyalty of old customers is getting lower and lower. User portraits can describe user characteristics from multiple dimensions. Different product types have different dimensions of user portraits. Faster and more accurate portrait information can help e-commerce companies accurately locate actual user groups and make product recommendations.

[0003] Traditional user portrait generation mainly relies on industry experts such as sales engineers. However, this method requires a lot of manpower and material resources, and it is difficult to quickly adapt to different products and customer groups. Another commonly used method is based on machine learning, which requires a large amount of labeled data and manpower to set labels, collect data, and adjust them.

[0004] Traditional methods require a lot of human resources, are inefficient, and difficult to expand and maintain. Machine learning-based methods require a lot of labeled data and human participation, involving complex data collection and label setting processes. Summary of the invention

[0005] In view of this, the purpose of the present invention is to provide a portrait model processing method, device, electronic device and computer storage medium, which analyzes data through a natural language processing model to determine the customer portrait, improve the accuracy of the portrait, reduce the cost of training the model, and enhance the user experience.

[0006] In a first aspect, the present invention provides a portrait model processing method, which includes: obtaining basic data; wherein the basic data includes: product information; inputting the basic data into a pre-trained portrait model, and outputting a first basic portrait; wherein the portrait model is determined based on natural language processing model training; and using the cleaned first basic portrait as a customer portrait.

[0007] In some preferred embodiments of the present invention, the basic data also includes: user information; the portrait model processing method also includes: inputting the basic data into the portrait model, outputting a second basic portrait; and using the cleaned second basic portrait as the customer portrait.

[0008] In some preferred embodiments of the present invention, after the step of inputting basic data into the portrait model and outputting the second basic portrait, the portrait model processing method further includes: integrating the first basic portrait and the second basic portrait to determine the third basic portrait; and using the third basic portrait as the customer portrait.

[0009] In some preferred embodiments of the present invention, the portrait model is trained through the following steps: obtaining historical data; wherein the historical data includes: professional articles and texts designed by industry experts; separating the historical data to determine the training data; separating the training data into a training set and a verification set; inputting the training set into the basic model to determine the first loss function value; wherein the basic model is a natural language processing model; if the first loss function value is less than a preset first loss function threshold, using the trained basic model as the model to be verified; and verifying the model to be verified based on the verification set.

[0010] In some preferred embodiments of the present invention, the step of verifying the model to be verified based on the verification set includes: inputting the verification set into the model to be verified, and determining the second loss function value; if the second loss function value is less than or equal to a preset second loss function threshold, using the model to be verified as a portrait model.

[0011] In some preferred embodiments of the present invention, after the step of inputting the verification set into the model to be verified and determining the second loss function value, the method further includes: if the second loss function value is greater than a preset second loss function threshold, continuing to train the basic model based on the training set until the second loss function value is less than or equal to the second loss function threshold.

[0012] In a second aspect, the present invention provides a portrait model processing device, which includes: a basic data acquisition module for acquiring basic data; wherein the basic data includes: product information; a basic portrait output module for inputting the basic data into a pre-trained portrait model and outputting a first basic portrait; wherein the portrait model is determined based on natural language processing model training; a customer portrait determination module for using the cleaned first basic portrait as a customer portrait.

[0013] In some preferred embodiments of the present invention, the portrait model processing device also includes: a model training module; a model training module for acquiring historical data; wherein the historical data includes: professional articles and texts designed by industry experts; separating the historical data to determine the training data; separating the training data into a training set and a verification set; inputting the training set into the basic model to determine the first loss function value; wherein the basic model is a natural language processing model; if the first loss function value is less than a preset first loss function threshold, using the trained basic model as a model to be verified; verifying the model to be verified based on the verification set.

[0014] In a third aspect, the present invention provides an electronic device comprising a processor and a memory, wherein the memory stores computer executable instructions that can be executed by the processor, and the processor executes the computer executable instructions to implement any of the above-mentioned portrait model processing methods.

[0015] In a fourth aspect, the present invention provides a computer storage medium storing computer executable instructions. When the computer executable instructions are called and executed by a processor, the computer executable instructions prompt the processor to implement any of the above-mentioned portrait model processing methods.

[0016] The present invention brings the following beneficial effects:

[0017] The present invention provides a portrait model processing method, which includes: obtaining basic data; wherein the basic data includes: product information; inputting the basic data into a pre-trained portrait model, and outputting a first basic portrait; wherein the portrait model is determined based on natural language processing model training; using the cleaned first basic portrait as a customer portrait; analyzing the data through the natural language processing model to determine the customer portrait, thereby improving the accuracy of the portrait, reducing the cost of training the model, and improving the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 A flowchart of a portrait model processing method provided by an embodiment of the present invention;

[0020] Figure 2 A flowchart of another portrait model processing method provided by an embodiment of the present invention;

[0021] Figure 3 A flowchart of a method for training a portrait model provided by an embodiment of the present invention;

[0022] Figure 4 A schematic diagram of the structure of a portrait model processing device provided by an embodiment of the present invention;

[0023] Figure 5 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention.

[0024] Icons: 410 - basic data acquisition module; 420 - basic portrait output module; 430 - customer portrait determination module; 500 - memory; 501 - processor; 502 - bus; 503 - communication interface. DETAILED DESCRIPTION

[0025] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0026] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0027] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0028] In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inside", "outside", etc. indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, or the positions or positional relationships in which the product of the invention is usually placed when in use. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific position, be constructed and operated in a specific position, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", "third", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.

[0029] In addition, the terms "horizontal", "vertical", "overhanging" and the like do not mean that the components are required to be absolutely horizontal or overhanging, but can be slightly tilted. For example, "horizontal" only means that its direction is more horizontal than "vertical", and does not mean that the structure must be completely horizontal, but can be slightly tilted.

[0030] In the description of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "set", "install", "connect", and "connect" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0031] In recent years, with the continuous development and rapid improvement of the Internet, traffic costs are becoming more and more expensive, the cost of attracting new customers is getting higher and higher, and the loyalty of old customers is getting lower and lower. User portraits can describe user characteristics from multiple dimensions. Different product types have different dimensions of user portraits. Faster and more accurate portrait information can help e-commerce companies accurately locate actual user groups and make product recommendations. Traditional user portrait generation mainly relies on industry experts such as sales engineers. However, this method requires a lot of manpower and material resources, and it is difficult to quickly make adaptive adjustments for different products and customer groups. Another commonly used method is the machine learning-based method, which requires a large amount of labeled data and manpower to set labels, collect data, and adjust them. Traditional methods require a lot of human resources, are inefficient, and difficult to expand and maintain. Machine learning-based methods require a large amount of labeled data and human participation, involving complex data collection and label setting processes.

[0032] Some embodiments of the present invention are described in detail below in conjunction with the accompanying drawings. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0033] Embodiment 1

[0034] The embodiment of the present invention provides a method for processing a portrait model. Figure 1 The flowchart of a portrait model processing method provided by an embodiment of the present invention is shown, and the portrait model processing method includes:

[0035] Step S102, obtaining basic data; wherein the basic data includes: product information;

[0036] Specifically, users store information about their products in CRM (Customer Relationship Management) and call data in the CRM system when needed. For example, the main business of a company is aquaculture, and the company pre-stores aquaculture as product information in the CRM system. Furthermore, some questions can be designed based on the product information and stored in the CRM as basic data, for example, "Which companies have demand for 'aquaculture' companies?"

[0037] Step S104, inputting the basic data into a pre-trained portrait model, and outputting a first basic portrait; wherein the portrait model is determined based on natural language processing model training;

[0038] Specifically, the basic data is input into the pre-trained portrait model to output the first basic portrait. For example, the question "Which companies have demand for 'aquaculture' companies?" is input into the portrait model, and the output result is as follows:

[0039] The aquaculture industry is in demand by many businesses, including the following:

[0040] 1. Catering industry: Restaurants, hotels, seafood restaurants and fast food chains all need aquatic products as ingredients. These businesses have a continuous demand for various aquaculture products such as fish, shrimp, shellfish, etc.

[0041] 2. Food processing enterprises: Food processing enterprises usually need a large amount of aquatic products as raw materials to produce various processed aquatic products, such as fish balls, fish sauce, fish oil, fish meal, shrimp, etc. These enterprises usually establish long-term cooperative relationships with aquaculture farms.

[0042] 3. Import and export trading companies: Import and export traders can import and export aquatic products from different countries or regions. They need to cooperate with aquaculture farms to meet the demand for aquatic products in the domestic and international markets.

[0043] 4. Feed companies: Aquaculture requires feed as the main source of nutrition for aquatic products. Feed companies can provide the feed needed for a variety of fish and shrimp farming and establish cooperative relationships with farms.

[0044] 5. Biotechnology and pharmaceutical companies: Ingredients and extracts in certain aquatic products have biological activity and medicinal value. Biotechnology and pharmaceutical companies may be interested in these ingredients to conduct research and develop new drugs or health products.

[0045] 6. Fishing gear and equipment suppliers: A fish farm may need a variety of fishing gear and equipment, such as fishing nets, breeding boxes, water pumps, oxygen aeration equipment, etc. Fishing gear and equipment suppliers can meet these needs.

[0046] 7. Aquaculture technology and consulting companies: Companies that provide aquaculture equipment, feed, and technical support to meet the needs of farms.

[0047] The above results are used as the first basic portrait.

[0048] Step S106, using the cleaned first basic portrait as the customer portrait;

[0049] Specifically, after obtaining the first basic portrait, some results do not meet the needs of the user. For example, items 6 and 7 in the above first basic portrait are cleared and the remaining content is output as the customer portrait.

[0050] The present invention provides a portrait model processing method, which includes: obtaining basic data; wherein the basic data includes: product information; inputting the basic data into a pre-trained portrait model, and outputting a first basic portrait; wherein the portrait model is determined based on natural language processing model training; using the cleaned first basic portrait as a customer portrait; analyzing the data through the natural language processing model to determine the customer portrait, thereby improving the accuracy of the portrait, reducing the cost of training the model, and improving the user experience.

[0051] Embodiment 2

[0052] Based on the above embodiment, the basic data also includes: user information; based on this, the embodiment of the present invention provides another portrait model processing method, see Figure 2 The flowchart of another portrait model processing method provided by an embodiment of the present invention is shown, and the method includes:

[0053] Step S202, obtaining basic data; wherein the basic data includes: product information and user information;

[0054] Specifically, the portrait model can not only obtain the first basic portrait based on the product information, but also obtain the second basic portrait based on the user information, wherein the user information is the information of the cooperative enterprises with which the enterprise has already concluded transactions, or the enterprise information of the potential cooperation objects that the enterprise users clearly know. Further, some questions can be designed based on the user information. For example, the business of the user's company is aquaculture, and it has reached cooperation with some restaurants. Then the following questions can be designed: "Our company's main business is "aquaculture". Now we have reached supply and demand relationships with many local restaurants. Which other companies can we provide services to?"

[0055] Step S204, inputting the basic data into a pre-trained portrait model, and outputting a first basic portrait and a second basic portrait;

[0056] Specifically, the user information is input into the portrait model, which decomposes the user information downward, analyzes the characteristics of the enterprise corresponding to the user information, and searches for enterprises with similar characteristics for portraits based on the characteristics. For example, if the user enterprise produces sensors, and the enterprises cooperating with the enterprise include mobile phone manufacturers, the portrait model conducts downward analysis of mobile phone manufacturers. Mobile phone manufacturers also use screens, audio processing equipment, video processing equipment, cameras and other hardware. The portrait model searches for enterprises that use similar hardware as mobile phone manufacturers, such as 3D stage manufacturing enterprises, and analyzes them as the user's portrait of potential customers.

[0057] For example, if we input the question "Our company's main business is aquaculture, and we have already established supply and demand relationships with many local restaurants, which other companies can we provide services to?" into the portrait model, the following results will be output:

[0058] As an aquaculture company, your main business is to provide aquatic products supply, mainly with local restaurants to establish a supply and demand relationship. In addition to restaurants, you can also consider providing services to the following businesses:

[0059] 1. Hotels and resorts: Hotels and resorts usually require a large amount of fresh aquatic products for supplying restaurants, buffets, banquets and room services, etc.

[0060] 2. Food processing plants: Food processing plants may require aquatic products as raw materials for processing various aquatic products, such as frozen seafood, surimi, fish chunks, fish balls, etc.

[0061] 3. Importers and exporters: By establishing cooperative relationships with importers and exporters, you can expand your market reach and sell your aquatic products to domestic and foreign supermarkets, retailers and overseas markets.

[0062] 4. Supermarkets and Retailers: Cooperate with local supermarkets and retailers to provide your aquatic products to consumers. You can provide them with a variety of fresh or frozen aquatic product options.

[0063] 5. Food delivery service: Partner with or set up your own food delivery service to deliver your seafood directly to consumers’ doorsteps to meet their needs.

[0064] 6. Feed manufacturers: If you also produce aquatic feed, you can provide the feed to other aquaculture companies to help them improve their breeding efficiency.

[0065] 7. Farming equipment supplier: In addition to the supply of aquatic products, you can also consider providing services such as farming equipment, fish fry and technical support to other farming companies.

[0066] 8. Restaurant chains: In addition to local restaurants, you can establish cooperative relationships with restaurant chains and provide seafood supplies for their branches. Use the above content as the second basic portrait. If necessary, you can clean the second basic portrait and use the cleaned content as the customer portrait.

[0067] Step S206, integrating the first basic portrait and the second basic portrait to determine a third basic portrait, and using the third basic portrait as the customer portrait;

[0068] Specifically, in this embodiment, the first basic portrait and the second basic portrait are integrated, that is, duplicate information is merged and filtered, and the integrated content is used as the customer portrait.

[0069] Exemplarily, the first basic portrait and the second basic portrait determined in the first embodiment and this embodiment are merged and closed to filter out duplicate content, and the following results are obtained:

[0070] 1. Catering industry: Restaurants, hotels, seafood restaurants and fast food chains all need aquatic products as ingredients. These businesses have a continuous demand for various aquaculture products such as fish, shrimp, shellfish, etc.

[0071] 2. Food processing enterprises: Food processing enterprises usually need a large amount of aquatic products as raw materials to produce various processed aquatic products, such as fish balls, fish sauce, fish oil, fish meal, shrimp, etc. These enterprises usually establish long-term cooperative relationships with aquaculture farms.

[0072] 3. Import and export trading companies: Import and export traders can import and export aquatic products from different countries or regions. They need to cooperate with aquaculture farms to meet the demand for aquatic products in the domestic and international markets.

[0073] 4. Feed companies: Aquaculture requires feed as the main source of nutrition for aquatic products. Feed companies can provide the feed needed for a variety of fish and shrimp farming and establish cooperative relationships with farms.

[0074] 5. Biotechnology and pharmaceutical companies: Ingredients and extracts in certain aquatic products have biological activity and medicinal value. Biotechnology and pharmaceutical companies may be interested in these ingredients to conduct research and develop new drugs or health products.

[0075] 6. Fishing gear and equipment suppliers: A fish farm may need a variety of fishing gear and equipment, such as fishing nets, breeding boxes, water pumps, oxygen aeration equipment, etc. Fishing gear and equipment suppliers can meet these needs.

[0076] 7. Aquaculture technology and consulting companies: Companies that provide aquaculture equipment, feed, and technical support to meet the needs of farms.

[0077] 8. Hotels and resorts: Hotels and resorts usually require a large amount of fresh aquatic products for supplying restaurants, buffets, banquets and room services, etc.

[0078] 9. Importers and exporters: By establishing cooperative relationships with importers and exporters, you can expand your market reach and sell your aquatic products to domestic and foreign supermarkets, retailers and overseas markets.

[0079] 10. Food delivery service: Partner with or set up your own food delivery service to deliver your seafood directly to consumers’ doorsteps to meet their needs.

[0080] Use the above content as your customer profile.

[0081] The embodiment of the present invention provides a portrait model processing method, which uses a natural language processing model to perform customer portrait, can quickly, efficiently and accurately extract key information from user information and customer information, and has a high extraction accuracy and effect guarantee; only a small amount of manpower is required, which greatly reduces the human resource and time costs compared to the label and template setting method of the traditional method; compared to real-time reliance on industry experts, the natural language processing model reasoning efficiency is higher, greatly saving time and human resource consumption, and once processed, it is always available; the natural language processing model has powerful expression and generalization capabilities, and is suitable for different types of products and customers. The automated process avoids errors and deviations caused by human factors and improves the accuracy and robustness of the portrait.

[0082] Embodiment 3

[0083] Based on the above embodiments, the present invention provides a method for training a portrait model. Figure 3 The flowchart of a method for training a portrait model provided by an embodiment of the present invention is shown, and the method includes:

[0084] Step S302, obtaining historical data; wherein the historical data includes: professional articles and texts designed by industry experts;

[0085] Specifically, historical data is obtained. Historical data is generally industry-related articles. For example, for the aquatic industry, it can generally be obtained from magazines such as "Journal of Fisheries of China", "Chinese Fisheries Science", and "Catering Manager". Of course, it also includes texts designed by some industry experts, such as: "1. Supply chain relationships in the catering industry: The catering supply chain mainly includes procurement, processing, circulation and other links. Through the integration of resources, it helps companies provide safe and stable raw materials, improve efficiency and reduce costs. The development of catering companies needs to rely on the support of the supply chain. An efficient supply chain can help catering companies ensure food safety and quality, improve procurement and logistics efficiency, control food costs to a certain extent, and ultimately improve the standardization of catering companies; 2. Supply chain relationships in the aquaculture industry: The upstream of the aquaculture industry chain is the breeding and farming link, which mainly includes feed production, hatching and breeding, aquaculture additives and drug production; the midstream of the industry chain is the processing and production link, which mainly involves the processing of fresh aquatic products; the downstream of the industry chain is the sales terminal. The processed aquatic products flow into the hands of consumers through e-commerce, supermarkets and other channels, including fish, shrimp, crabs and shellfish that appear on residents' tables, as well as canned foods, snack foods and other products." In general, the ratio of professional articles to texts designed by industry experts is about 99:1.

[0086] Step S304, separating the historical data to determine the training data; separating the training data into a training set and a validation set;

[0087] Specifically, the historical data is preprocessed, the symbols and English in the text are removed, the remaining text is separated, and some words are replaced; for example, "The development of catering enterprises needs to rely on the support of the supply chain" can be separated into "Catering enterprises' development needs to rely on the support of the supply chain", and "need" is replaced with "to be predicted", and "Catering enterprises' development 'to be predicted' relies on the support of the supply chain" is used as historical data, and the historical data is divided according to a preset ratio. Generally, the ratio of training set to validation set is 8:2.

[0088] Step S306, inputting the training set into the basic model to determine the first loss function value; wherein the basic model is a natural language processing model;

[0089] Specifically, the historical data is input into the natural language processing model, and the first loss function value is output. Furthermore, in model building, a common method is to use a generative adversarial network to fight against adversarial training between the generator and the discriminator; another method is to use reinforcement learning, in which the model learns to generate answers by interacting with the environment and obtaining rewards; you can choose a suitable training method according to your needs.

[0090] Step S308, if the first loss function value is less than the preset first loss function threshold, the trained basic model is used as the model to be verified;

[0091] For example, if the result of the first loss function value is 0.5, and the first loss function threshold is set to 0.4, it is obvious that the training result does not meet the requirements, and the training is continued. If the result of the first loss function value is 0.38, 0.4>0.38, it meets the requirements, and the trained basic model is used as the model to be verified.

[0092] Step S310, inputting the verification set into the model to be verified to determine the value of the second loss function;

[0093] Step S312, if the second loss function value is less than or equal to the preset second loss function threshold, the model to be verified is used as a portrait model;

[0094] Step S314: if the second loss function value is greater than the preset second loss function threshold, continue to train the basic model based on the training set until the second loss function value is less than or equal to the second loss function threshold;

[0095] Specifically, when the training set is small, it is possible to determine whether the model to be verified is qualified based on a second loss function value; generally, the verification set contains multiple samples, and multiple second loss function values ​​can be determined. The average value can be taken and compared with the second loss function threshold. If the average value is less than or equal to the second loss function threshold, the model to be verified is used as a portrait model; if the average value is greater than the preset second loss function threshold, continue to train the basic model based on the training set until the second loss function value is less than or equal to the second loss function threshold.

[0096] Furthermore, the ratio of multiple second loss values ​​that meet the conditions can also be taken. If the ratio exceeds a certain ratio threshold, the model to be verified is used as a portrait model. Otherwise, the basic model is trained based on the training set until the conditions are met. Exemplarily, 100 second loss function values ​​are determined, of which 97% are less than or equal to the second loss function threshold, and the preset ratio threshold is 95%. Because 97%>95%, the model to be verified can be used as a portrait model.

[0097] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the training method of the portrait model described above can refer to the corresponding process in the embodiment of the aforementioned portrait model processing method, and will not be repeated here.

[0098] Embodiment 4

[0099] Based on the above embodiments, the present invention provides a portrait model processing device. Figure 4The structural diagram of a portrait model processing device provided by an embodiment of the present invention is shown, and the portrait model processing device includes:

[0100] The basic data acquisition module 410 is used to acquire basic data; wherein the basic data includes: product information;

[0101] A basic portrait output module 420 is used to input basic data into a pre-trained portrait model and output a first basic portrait; wherein the portrait model is determined based on natural language processing model training;

[0102] The customer portrait determination module 430 is used to use the cleaned first basic portrait as the customer portrait.

[0103] In some preferred embodiments of the present invention, the portrait model processing device also includes: a model training module; a model training module for acquiring historical data; wherein the historical data includes: professional articles and texts designed by industry experts; separating the historical data to determine the training data; separating the training data into a training set and a verification set; inputting the training set into the basic model to determine the first loss function value; wherein the basic model is a natural language processing model; if the first loss function value is less than a preset first loss function threshold, using the trained basic model as a model to be verified; verifying the model to be verified based on the verification set.

[0104] In some preferred embodiments of the present invention, the basic portrait output module 420 is also used to input basic data into the portrait model and output a second basic portrait.

[0105] In some preferred embodiments of the present invention, the customer portrait determination module 430 is further configured to use the cleaned second basic portrait as the customer portrait.

[0106] In some preferred embodiments of the present invention, the customer portrait determination module 430 is further used to integrate the first basic portrait and the second basic portrait to determine a third basic portrait; and use the third basic portrait as the customer portrait.

[0107] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the portrait model processing device described above can refer to the corresponding process in the aforementioned embodiment of the portrait model processing method, and will not be repeated here.

[0108] Embodiment 5

[0109] The embodiment of the present invention also provides an electronic device for executing the portrait model processing method; see Figure 5The structural schematic diagram of an electronic device provided by an embodiment of the present invention is shown, and the electronic device includes a memory 500 and a processor 501, wherein the memory 500 is used to store one or more computer instructions, and the one or more computer instructions are executed by the processor 501 to implement the above-mentioned portrait model processing method.

[0110] Further, Figure 5 The electronic device shown further includes a bus 502 and a communication interface 503 , and the processor 501 , the communication interface 503 and the memory 500 are connected via the bus 502 .

[0111] The memory 500 may include a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk storage. The communication connection between the system network element and at least one other network element is realized through at least one communication interface 503 (which may be wired or wireless), and the Internet, wide area network, local area network, metropolitan area network, etc. may be used. The bus 502 may be an ISA bus, a PCI bus, or an EISA bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 5 Only one bidirectional arrow is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0112] The processor 501 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit or software instructions in the processor 501. The above processor 501 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present invention can be implemented or executed. The general processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in the embodiment of the present invention can be directly embodied as a hardware decoding processor for execution, or a combination of hardware and software modules in the decoding processor for execution. The software module may be located in a storage medium mature in the art, such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory 500, and the processor 501 reads the information in the memory 500 and completes the steps of the method of the above embodiment in combination with its hardware.

[0113] An embodiment of the present invention further provides a computer-readable storage medium, which stores computer-executable instructions. When the computer-executable instructions are called and executed by a processor, the computer-executable instructions prompt the processor to implement the above-mentioned business recommendation method. The specific implementation can be found in the method embodiment, which will not be repeated here.

[0114] The computer program product of the portrait model processing method, device and electronic device provided in the embodiments of the present invention includes a computer-readable storage medium storing program code. The instructions included in the program code can be used to execute the method in the previous method embodiment. The specific implementation can be found in the method embodiment, which will not be repeated here.

[0115] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the system and / or device described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0116] In addition, in the description of the embodiments of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0117] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0118] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A portrait model processing method, characterized in that: The portrait model processing method includes: Acquire basic data; wherein the basic data includes: product information; Inputting the basic data into a pre-trained portrait model to output a first basic portrait; wherein the portrait model is determined based on natural language processing model training; The first basic portrait after cleaning is used as the customer portrait.

2. The image model processing method according to claim 1, characterized in that: The basic data also includes: user information; The portrait model processing method also includes: Inputting the basic data into the portrait model and outputting a second basic portrait; The cleaned second basic portrait is used as the customer portrait.

3. The image model processing method according to claim 2, characterized in that: After the step of inputting the basic data into the portrait model and outputting the second basic portrait, the portrait model processing method further includes: Integrate the first basic portrait and the second basic portrait to determine a third basic portrait; The third basic portrait is used as the customer portrait.

4. The image model processing method according to any one of claims 1 to 3, characterized in that: The image model is trained through the following steps: Acquire historical data; wherein the historical data includes: professional articles and texts designed by industry experts; Separate and process the historical data to determine training data; Separating the training data into a training set and a validation set; Inputting the training set into a basic model to determine a first loss function value; wherein the basic model is a natural language processing model; If the first loss function value is less than a preset first loss function threshold, the trained basic model is used as a model to be verified; The model to be verified is verified based on the verification set.

5. The image model processing method according to claim 4, characterized in that: The step of verifying the model to be verified based on the verification set includes: Inputting the verification set into the model to be verified to determine a second loss function value; If the second loss function value is less than or equal to a preset second loss function threshold, the model to be verified is used as the portrait model.

6. The image model processing method according to claim 5, characterized in that: After the step of inputting the verification set into the model to be verified and determining the value of the second loss function, the method further includes: If the second loss function value is greater than a preset second loss function threshold, continue training the basic model based on the training set until the second loss function value is less than or equal to the second loss function threshold.

7. A portrait model processing device, characterized in that: The portrait model processing device comprises: A basic data acquisition module is used to acquire basic data; wherein the basic data includes: product information; A basic portrait output module, used to input the basic data into a pre-trained portrait model and output a first basic portrait; wherein the portrait model is determined based on natural language processing model training; The customer portrait determination module is used to use the first basic portrait after cleaning as the customer portrait.

8. The image model processing device according to claim 7, characterized in that: The portrait model processing device also includes: a model training module; The model training module is used to obtain historical data; wherein the historical data includes: professional articles and texts designed by industry experts; the historical data is separated and processed to determine the training data; the training data is separated into a training set and a verification set; the training set is input into the basic model to determine the first loss function value; wherein the basic model is a natural language processing model; if the first loss function value is less than a preset first loss function threshold, the trained basic model is used as a model to be verified; and the model to be verified is verified based on the verification set.

9. An electronic device, characterized in that: It includes a processor and a memory, wherein the memory stores computer executable instructions that can be executed by the processor, and the processor executes the computer executable instructions to implement the portrait model processing method described in any one of claims 1 to 6 above.

10. A computer storage medium, characterized in that: The computer storage medium stores computer executable instructions. When the computer executable instructions are called and executed by the processor, the computer executable instructions prompt the processor to implement the portrait model processing method described in any one of claims 1 to 6.

Images