Model cooperative processing method and device, electronic equipment and computer storage medium

Abstract

The invention provides a model cooperative processing method and device, electronic equipment and a computer storage medium, and relates to the technical fields of artificial intelligence, natural language processing, computer vision and the like. According to the specific implementation scheme, to-be-processed information in an agricultural scene is acquired; sending the to-be-processed information and the decision information cue word to the large model to obtain small model information and decision parameter information output by the large model; based on the small model information, determining a target small model set, and sending the decision parameter information to related target small models in the target small model set; sending the to-be-processed information to the target small model set; receiving a small model processing result set output by the target small model set; and obtaining a processing result of the to-be-processed information based on the small model processing result set, the to-be-processed information, the small model information, the decision parameter information and the large model.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, and in particular to natural language processing, artificial intelligence, computer vision, and other technical fields. In particular, the present disclosure provides a model collaborative processing method and apparatus, an electronic device, and a computer-readable storage medium. Background Art

[0002] Large models perform well in general knowledge question answering, but they still have shortcomings in specific areas, such as agricultural production processes. Large models lack specific industry knowledge and are prone to misconceptions. Current large models tend to think and plan in a generalized and divergent manner, failing to provide specialized thinking and planning within specific agricultural scenarios. Small agricultural models require numerous input parameters, are susceptible to regional constraints, and have weak generalization capabilities. Existing integration approaches primarily involve building small model tools into APIs and directly invoking them through workflows or agents. This obscures the large model's internal operational logic, preventing optimization of the small models and rendering their results unreliable. Summary of the Invention

[0003] The present disclosure provides a model collaborative processing method and device, an electronic device, and a computer-readable storage medium.

[0004] According to a first aspect, a model collaborative processing method is provided, which includes: obtaining information to be processed in an agricultural scene; sending the information to be processed and decision information prompt words to a large model to obtain small model information and decision parameter information output by the large model; determining a target small model set based on the small model information, and sending the decision parameter information to the relevant target small models in the target small model set; sending the information to be processed to the target small model set; receiving a small model processing result set output by the target small model set; and obtaining a processing result of the information to be processed based on the small model processing result set, the information to be processed, the small model information, the decision parameter information and the large model.

[0005] According to the second aspect, a model collaborative processing device is provided, which includes: an acquisition unit, configured to acquire information to be processed in an agricultural scene; a decision unit, configured to send the information to be processed and the decision information prompt word to the big model, and obtain the small model information and decision parameter information output by the big model; a determination unit, configured to determine the target small model set based on the small model information, and send the decision parameter information to the relevant target small models in the target small model set; a sending unit, configured to send the information to be processed to the target small model set; a receiving unit, configured to receive the small model processing result set output by the target small model set; an obtaining unit, configured to obtain the processing result of the information to be processed based on the small model processing result set, the information to be processed, the small model information, the decision parameter information and the big model.

[0006] According to a third aspect, 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 to enable the at least one processor to execute the method described in any implementation manner of the first aspect.

[0007] According to a fourth aspect, a non-transitory computer-readable storage medium storing computer instructions is provided, where the computer instructions are used to cause a computer to execute the method as described in any implementation of the first aspect.

[0008] The model collaborative processing method and device provided by the embodiments of the present disclosure first obtain the information to be processed in the agricultural scene; secondly, send the information to be processed and the decision information prompt word to the big model to obtain the small model information and decision parameter information output by the big model; thirdly, based on the small model information, determine the target small model set, and send the decision parameter information to the relevant target small models in the target small model set; thirdly, send the information to be processed to the target small model set; then, receive the small model processing result set output by the target small model set; finally, based on the small model processing result set, the information to be processed, the small model information, the decision parameter information and the big model, obtain the processing result of the information to be processed, and collaboratively process the information to be processed by combining the big model with the target small model. This collaborative processing method makes full use of the global decision-making ability of the big model and the refined processing ability of the small model, realizes the efficiency, accuracy and flexibility of information processing in the agricultural scene, can effectively improve the decision-making efficiency and management level of agricultural production, and provides strong support for the development of intelligent agriculture.

[0009] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present disclosure, nor are they intended to limit the scope of the present disclosure. Other features of the present disclosure will become readily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The accompanying drawings are provided to facilitate a better understanding of the present invention and do not constitute a limitation of the present disclosure. Figure 1 is a flow chart of an embodiment of a model collaborative processing method according to the present disclosure; Figure 2 It is a structural diagram of a system corresponding to the disclosed model collaborative processing method; Figure 3 It is a structural diagram of an embodiment of the model collaborative processing device disclosed in the present invention; Figure 4 It is a block diagram of an electronic device used to implement the model collaborative processing method of an embodiment of the present disclosure. DETAILED DESCRIPTION

[0011] Unless expressly stated otherwise, throughout the specification and claims, the term "comprise" or variations such as "include" or "comprising", etc., will be understood to include the stated elements or components but not to exclude other elements or other components.

[0012] The technical solutions of the present disclosure are described below through specific examples. It should be understood that one or more steps mentioned in the present disclosure do not exclude the existence of other methods and steps before and after the combination step, or other methods and steps may be inserted between these explicitly mentioned steps. It should also be understood that these examples are only used to illustrate the present disclosure and are not used to limit the scope of the present disclosure. Unless otherwise specified, the numbering of each method step is only for the purpose of identifying each method step, and does not limit the order of arrangement of each method or limit the scope of implementation of the present disclosure. Changes or adjustments in their relative relationships can also be regarded as the scope of implementation of the present disclosure without substantial changes in the technical content.

[0013] The sources of the raw materials and instruments used in the examples are not particularly limited and can be purchased from the market or prepared according to conventional methods known to those skilled in the art.

[0014] In existing technologies, agricultural models face two core challenges: Limitations of large models: Although general large models have natural language interaction and common sense reasoning capabilities, the general knowledge base has problems such as insufficient knowledge coverage in agricultural professional fields and poor data timeliness. In professional scenarios such as crop growth models and pest and disease diagnosis, logical gaps or knowledge hallucinations are easily generated, and there is a data acquisition bottleneck when a large number of input parameters are required.

[0015] Pain points of small agricultural models: Crop growth models based on expert experience (such as DSSAT, WOFOST, and APSIM) rely on high-precision soil parameters, meteorological data, and other inputs. Model parameters are region-specific, and their accuracy decreases when applied across regions (for example, a Northeast China corn model is difficult to directly apply to the mountainous areas of the Southwest). Combination of large and small models: The large model has its own thinking and reasoning logic, and the small model has its own theoretical rules. The current combination mainly directly calls the model operation result data through workflows or intelligent agents. The comprehensive analysis of the result data cannot reveal the internal operation logic of the small model, and cannot fundamentally resolve the contradiction between the data dependence of the small model and the professional credibility of the large model.

[0016] In response to the defects in traditional technologies, this paper proposes a model collaborative processing method. This collaborative processing method fully utilizes the global decision-making capabilities of large models and the refined processing capabilities of small models, achieves high efficiency, accuracy and flexibility in information processing, can effectively improve production decision-making efficiency and management level, and provide strong support for intelligent development. Figure 1 A process 100 according to an embodiment of the model collaborative processing method of the present disclosure is shown. The model collaborative processing method includes the following steps: Step 101: Obtain information to be processed in the agricultural scene.

[0017] In this embodiment, the information to be processed refers to the data on unresolved issues related to agricultural scenarios collected from the terminal and / or scene information collected from data sources such as agricultural Internet of Things sensor networks, satellite remote sensing, and weather stations. The scene information includes soil moisture, temperature, light intensity, crop growth status, pest and disease conditions, and meteorological data. After preliminary processing of the data on unresolved issues and the scene information, structured information to be processed is formed. The information to be processed can be multimodal data such as images and texts. After preliminary processing of the information to be processed, information that can be used for further analysis can be formed, such as soil moisture, temperature, crop growth status, pest and disease problems, etc.

[0018] In this embodiment, the process of acquiring pending information in agricultural scenarios can be described as follows: First, a sensor network is deployed across farmland using IoT technology. These sensors can collect key data such as soil moisture, temperature, light intensity, and crop growth status in real time. Simultaneously, satellite remote sensing technology is used to acquire large-scale farmland imagery for analysis of crop distribution, growth, and pest and disease conditions. Furthermore, weather station data can be combined to obtain local weather information such as temperature, precipitation, and wind speed. This data is transmitted to an agricultural information management system for storage and preliminary screening. Next, the system processes the collected data using data analysis algorithms to identify abnormal data points, such as soil moisture below the optimum range or crop growth indicators deviating from normal values. This abnormal data is marked as pending information. Combined with historical data and expert knowledge bases, corresponding warning information and recommended measures are generated, such as irrigation reminders, fertilization recommendations, or pest and disease control plans. Finally, this pending information is pushed to farm managers via mobile applications or text messages, allowing them to take timely action to optimize agricultural production processes and improve crop yield and quality.

[0019] Optionally, the above step 101 includes: receiving agricultural scenario problems or agricultural scenario data processing demand information sent by users, and treating the agricultural scenario problems or agricultural scenario data processing demand information as information to be processed.

[0020] Step 102: Send the information to be processed and the decision information prompt words to the large model to obtain the small model information and decision parameter information output by the large model.

[0021] In this embodiment, the small model information is information related to the small model, including the small model function, small model model, small model parameters, etc. The decision parameter information is the parameters of the small model that are closely related to the decision during the decision-making process. For example, the decision parameter information is cross-regional high-precision soil parameters, meteorological data, etc., where cross-region refers to a region that is different from the regional data currently analyzed by the small model. For example, the small model is a Northeast corn analysis model, and the decision parameter information includes corn analysis data in the southwest region.

[0022] In this embodiment, when the large model executes logic according to the rules of the small agricultural model, if it encounters information or parameters that cannot be obtained by simply calling existing resources through the small model, it can obtain decision parameter information through internet knowledge (such as a crop database), such as querying local historical meteorological disasters and pest and disease information. If the results of this step differ significantly from the actual results, the parameters need to be fine-tuned to achieve a reasonable result. Among them, the above-mentioned existing resources can be a pre-maintained variety database, a weather query tool interface, or a tool interface for querying soil nutrient and texture information. Optionally, when decision parameter information cannot be obtained from internet knowledge, the decision parameter information can also be information directly generated by the large model.

[0023] In this embodiment, decision information prompts are used to prompt the large model to determine the small model information related to the information to be processed and to search for decision parameter information for the determined small model. First, the large model extracts the information to be processed from the agricultural IoT system, including soil moisture, temperature, crop growth status, pest and disease conditions, and meteorological data. This information is combined with preset decision information prompts (e.g., "Determine whether irrigation is needed based on the current soil moisture and crop growth stage" or "Recommend appropriate control measures based on the pest and disease situation") to generate complete input content. This information is then sent to the large language model. Leveraging its powerful natural language processing capabilities and knowledge base, the large language model analyzes and infers the input information to generate targeted small model information (e.g., a simplified irrigation decision model or pest and disease control model) as well as specific decision parameter information (e.g., irrigation amount, irrigation time, pesticide type, and dosage). Finally, the small model information and decision parameter information output by the large language model are stored in the agricultural management system's database for reference and implementation by farm managers and users, thereby achieving accurate and efficient agricultural decision support.

[0024] Step 103: Determine a target small model set based on the small model information, and send the decision parameter information to the relevant target small models in the target small model set.

[0025] In this embodiment, the target small model set refers to a set of small models specially trained for specific agricultural tasks (such as irrigation, pest control, fertilizer management, etc.), wherein the target small model set includes at least one target small model, each target small model has a corresponding specific agricultural task, and these target small models can handle specific agricultural problems and generate targeted decision-making recommendations. For example, if the information to be processed is: What varieties of spring corn are suitable for planting in Northeast China in 2025, the target small model set includes: a meteorological analysis small model and a soil analysis small model; wherein the meteorological analysis small model is used to analyze local meteorological conditions and determine the accumulated temperature during the planting season; the soil analysis small model is used to analyze local soil conditions and determine the soil nutrient conditions. Figure 2 As shown, each target model in the target model set has its own theoretical rules. After determining the target model set, the theoretical rules of all target models in the target model set are combined in the order of reasoning to obtain a standardized rule chain. Figure 2 In the process, the standardized rule chain has three steps: the first step, the second step and the third step. The large model can obtain the standardized rule chain of the target small model set, and can also process the information to be processed according to the theoretical rules corresponding to the standardized rule chain.

[0026] In this embodiment, sending the decision parameter information to the relevant target small model in the target small model set includes: determining the target small model that needs the decision parameter information, and sending the decision parameter information to the target small model.

[0027] In this embodiment, in an intelligent agricultural system, based on the small model information output by the large language model, the system further analyzes and selects a set of target small models that best match the current information to be processed and the decision-making requirements. These target small models are typically pre-trained, specialized models for specific agricultural scenarios (such as irrigation optimization, pest control, and fertilization management), and they can more accurately handle specific tasks. The specific implementation process is as follows: the system first parses the small model information output by the large language model to identify model features relevant to the current agricultural task (such as model type, applicable scenario, input and output parameters, etc.). It then retrieves a set of target small models that meet these features from a pre-set small model library. The decision parameter information (such as irrigation amount, fertilization rate, and control measures) output by the large language model is then transmitted to these target small models. Based on the received decision parameter information and combined with their own training data and algorithms, the target small models further optimize the decision plan and generate final execution instructions, such as a specific irrigation schedule and pesticide spraying path, thereby achieving accurate and efficient agricultural decision-making and automated execution.

[0028] Step 104: Send the information to be processed to the target small model set.

[0029] In this embodiment, the execution entity on which the model collaborative processing method runs can send the information to be processed to each target model in the target model set (each target model is a parallel processing task), or send the information to be processed to a specific target model in the target model set (the target models in the target model set are a serial processing task). The target models in the target model set will analyze and process the received information to be processed and specific decision parameter information, such as irrigation amount, irrigation time, fertilizer type and amount, and pest and disease control measures, based on their own algorithms and training data, to generate a set of small model processing results, thereby providing accurate decision support for agricultural managers.

[0030] like Figure 2 As shown, after the information to be processed is sent to the target small model set, the small model processing result set output by the target small model set can be obtained. The small model processing result set includes the small model processing results of each target small model in the target small model set, such as Figure 2 In the example, the small model processing results include: result 1, result 2, and result 3 executed in sequence.

[0031] Step 105: Receive the small model processing result set output by the target small model set.

[0032] In this embodiment, the small model processing result set is the result set output after each target small model processes the information to be processed. The small model processing result set includes at least one small model processing result. Each small model processing result corresponds to a target small model. Each small model processing result can include specific processing results, optimization suggestions or warning information, etc.

[0033] In this embodiment, the specific content of the processing results varies depending on the problem to be solved. For example, the processing results may include: irrigation system startup time and water volume, fertilizer type and dosage, pest and disease control measures, etc. The execution entity running the model collaborative processing method integrates and verifies the received processing results to ensure their accuracy and consistency. These processing results are then presented to agricultural managers in the form of a visual interface, SMS notifications, or automated instructions, or sent directly to relevant agricultural equipment, thereby achieving precise agricultural decision support and automated operations.

[0034] Step 106: Based on the small model processing result set, the information to be processed, the small model information, the decision parameter information and the large model, a processing result of the information to be processed is obtained.

[0035] In this embodiment, the execution entity on which the model collaborative processing method runs first receives a set of small model processing results output by the target small model set. These results are specific decision recommendations or optimization solutions generated by the small models based on the information to be processed. At the same time, the execution entity combines the scenario information in the information to be processed (such as sensor data, meteorological data, etc.) with the small model information (such as the applicable scenarios and functional descriptions of the small models) and decision parameter information (such as specific irrigation and fertilizer amounts, etc.) output by the large language model. By comparing and verifying this information, the execution entity utilizes the comprehensive analysis capabilities of the large model to further optimize and calibrate the processing results of the small models. Ultimately, the processing results of the information to be processed are generated, such as precise irrigation plans, pest and disease control measures, or crop growth optimization plans, and are fed back to agricultural managers or sent directly to relevant agricultural equipment to achieve intelligent agricultural decision-making and automated operations.

[0036] Optionally, the above step 106 also includes: determining the local data of the target area based on the information to be processed, combining the local data with the small model processing result set to obtain the processing result of the information to be processed. For example, the information to be processed is: whether spring corn in Northeast China is suitable for local planting. The above determination of the local data of the target area based on the information to be processed includes: local target yield analysis, and obtaining the yield information of the area by searching the Internet information. Other planting management, irrigation type, sowing date, and harvest date are obtained from the Internet information data to obtain the planting information of the area. Based on the Internet information, the variety data of the area for many years are obtained to obtain the local variety data. When the large model optimizes and calibrates the processing of the small model, the yield information, planting information and local variety data are sent to the large model, so that the large model can fully refer to the local information, give accurate optimization and calibration results, and improve the accuracy of the processing results of the information to be processed.

[0037] Optionally, a small model can also be used to verify the reasoning results of the large model. The reasoning results of the large model need to be verified by the small model. If they exceed the limit, it will trigger iterative correction (such as adjusting parameters or re-retrieval of knowledge). The above step 106 includes: based on the small model information, determining the standardized rule chain of the target small model set; using the large model to execute the standardized rule chain to obtain multiple execution results; sending the decision parameter information and the information to be processed to the target small model set, and collecting the processing results of each target small model in the target small model set, and determining whether the small model processing result set is qualified by comparing the small model processing result set with the processing results of all the collected target small models. When the small model processing result set is qualified, the small model processing result set is used to verify the various execution results output by the large model.

[0038] When the large model executes a standardized rule chain (i.e., following the rules of the smaller model), each step produces an output, and the accuracy of each step needs to be verified. For example, the crop model simulates the growth and development of crops. Soil moisture simulation uses meteorological and initial soil information, randomizing some initial parameters to dynamically simulate soil moisture balance. This process can be verified using real data measured by field moisture instruments. If there are deviations, the initialization parameters are corrected. Growth period simulation uses accumulated temperature and field management information to initialize other parameters of the smaller model, calculate the crop growth period, and verify the authenticity of the growth period simulation using actual field observation records. Leaf area and biomass simulations are also verified using field sampling data.

[0039] The model collaborative processing method provided in this embodiment implements rule-driven and data-complete approaches. Specifically, it breaks down agricultural expert rules into rule chains understandable to the large model, dynamically completes small model parameters with internet knowledge, and reduces data collection costs. It also implements dual trusted verification, specifically, cross-validating the large and small model processes to reduce the illusion of the large model while improving the regional adaptation accuracy of the small model. It can be generalized and applied, using the large model to generate virtual data, making the small model applicable to other regions within a reasonable range. It is suitable for multi-level scenarios from family farms to large-scale planting bases.

[0040] The model collaborative processing method provided by the embodiment of the present disclosure first obtains the information to be processed in the agricultural scene; secondly, the information to be processed and the decision information prompt word are sent to the large model to obtain the small model information and decision parameter information output by the large model; thirdly, based on the small model information, the target small model set is determined, and the decision parameter information is sent to the relevant target small model in the target small model set; thirdly, the information to be processed is sent to the target small model set; then, the small model processing result set output by the target small model set is received; finally, based on the small model processing result set, the information to be processed, the small model information, the decision parameter information and the large model, the processing result of the information to be processed is obtained. This collaborative processing method fully utilizes the global decision-making ability of the large model and the refined processing ability of the small model, realizes the efficiency, accuracy and flexibility of information processing in the agricultural scene, can effectively improve the decision-making efficiency and management level of agricultural production, and provides strong support for the development of intelligent agriculture.

[0041] In some optional implementations of the present disclosure, the above-mentioned processing results of the information to be processed based on the small model processing result set, the information to be processed, the small model information, the decision parameter information and the large model include: verifying the small model processing result set through the large model based on the information to be processed, the small model information and the decision parameter information to obtain the verification result; in response to detecting that the verification result indicates that the verification is qualified, extracting the processing results of the information to be processed in the small model processing result set.

[0042] In this optional implementation, the above-mentioned extraction of the processing results of the information to be processed in the small model processing result set includes: in response to the target small model in the target small model set being a model that sequentially executes the tasks corresponding to the information to be processed, determining the last target small model in the target small model set, and extracting the processing result of the last target small model in the small model processing result set as the processing result of the information to be processed.

[0043] In this optional implementation, the small model processing result set has a corresponding processing specification or result threshold. When the small model processing result set meets the processing specification or does not exceed the result threshold, the verification result indicates that the verification is qualified; when the small model processing result set does not meet the processing specification or exceeds the result threshold, the verification result indicates that the verification is unqualified.

[0044] In this optional implementation, the information to be processed, the small model information, and the decision parameter information are input into the large model. The large model uses its powerful comprehensive analysis capabilities and knowledge base to verify the small model processing result set. The verification process includes checking whether the output of the small model conforms to the expected decision logic, whether it is consistent with the information to be processed, and whether it meets the preset decision parameter requirements. If the verification result shows that the processing result of the small model is qualified, that is, the result is accurate and meets expectations, the execution entity on which the model collaborative processing method runs will extract the specific processing results for the information to be processed from the verified qualified small model processing result set, such as irrigation plans, fertilization recommendations, or pest and disease control measures. These verified processing results can be directly used for decision-making and execution in agricultural production, thereby improving the efficiency and accuracy of agricultural management.

[0045] Optionally, the above-mentioned processing result of the information to be processed based on the small model processing result set, the information to be processed, the small model information, the decision parameter information and the large model also includes: in response to detecting that the verification result indicates that the verification is unqualified, outputting new small model information and new decision parameter information through the large model, determining a new target small model set based on the new small model information, and sending the new decision parameter information to the relevant target small models in the new target small model set; sending the information to be processed to the new target small model set; receiving the new small model processing result set output by the new target small model set; verifying the new small model processing result set through the large model based on the information to be processed, the new small model information and the new decision parameter information to obtain a verification result; and continuing to detect whether the verification result indicates that the verification is qualified until the verification result indicates that the verification is qualified.

[0046] In some optional implementations of the present disclosure, the above-mentioned verification of the small model processing result set through the big model based on the information to be processed, the small model information and the decision parameter information, and the obtained verification result includes: determining the standardized rule chain of the target small model based on the small model information; filling the standardized rule chain based on the decision parameter information to obtain the processing logic chain; controlling the big model to use the processing logic chain to process the information to be processed to obtain the big model processing result set; detecting whether the small model processing result set and the big model processing result set are consistent; in response to detecting that the small model processing result set and the big model processing result set are consistent, obtaining a qualified verification result; in response to detecting that the small model processing result set and the big model processing result set are inconsistent, obtaining an unqualified verification result.

[0047] In this optional implementation, the decision-making process of the target small model is deconstructed into multiple executable steps. For example, the executable steps of the pest control small model include: soil moisture monitoring → water demand calculation → irrigation plan generation → disease symptom identification → environmental factor analysis → control plan recommendation. The executable steps are converted into a standardized rule chain through a semantic parser. The standardized rule chain includes at least two standardized rules, each of which corresponds to an execution step (the parameters in the execution step can be fixed values, or values ​​that need to be calculated by the small model). At least two standardized rules form a chain of small model rules. For example, the standardized rule chain of the plant protection small model is: 1. According to the climate conditions, determine the soil conditions, variety conditions, and the types of pests and diseases that are prone to occur; 2. Determine the occurrence patterns of conventional pests and diseases; 3. Analyze the meteorological conditions this year; 4. Available pest and disease control measures; 5. Pesticide ingredients; 6. Products available in the database.

[0048] In this optional implementation, the standardized rule chain of the variety recommendation model includes: 1) Approval area adaptability: Screen the nationally approved varieties in the region to ensure that the varieties are adapted to the climate and soil conditions of the region. According to the historical meteorological data of the region, the accumulated temperature range of the planting season in the region is calculated. 2) Growth period and planting pattern: According to the local planting pattern (one-season, two-season intercropping), the number of days of the crop growth cycle is determined to screen suitable varieties. 3) Stress resistance and disease resistance: According to the local historical disease or meteorological disaster occurrence, disease-resistant and stress-resistant varieties are screened. 4) Seed quality and field performance: Screen varieties with a germination rate ≥90% and select varieties that meet local target yields. 5) Soil and fertility adaptation: For heavy clay soils, select varieties with well-developed root systems and resistance to barrenness, and for sandy loam soils, select varieties that tolerate dense planting and prefer fertilizer and water.

[0049] In this optional implementation, the processing logic chain includes at least two processing logics, each corresponding to an execution step and representing the actual processing method employed in conjunction with the decision parameter information when executing the step. For example, if the decision parameter information for the first standardized rule above is the current climate at location X, the processing logic corresponding to this standardized rule would be: Under the current climate at location X, determine the soil conditions, crop varieties, and the types of pests and diseases that are likely to occur at location X.

[0050] In this optional implementation, the execution entity running the model collaborative processing method determines the standardized rule chain for the target small model based on the small model's information (including its functionality, applicable scenarios, and processing logic). This standardized rule chain is a series of predefined execution steps, each of which standardizes the processing of the small model. The execution entity then populates the standardized rule chain based on decision parameter information (such as specific irrigation and fertilizer application amounts) to generate a complete processing logic chain. This processing logic chain represents the specific decision logic that guides the large model in processing the information to be processed.

[0051] The execution entity controls the large model to process the information to be processed using the processing logic chain and generate a large model processing result set. The large model processing result set is the decision recommendation or optimization solution generated by the large model based on the processing logic chain.

[0052] In this optional implementation, the execution entity checks whether the small model processing result set is consistent with the large model processing result set. If the two are consistent, it indicates that the processing results of the small model are reliable, and the system obtains a qualified verification result. Conversely, if the two are inconsistent, it indicates that the processing results of the small model may be biased, and the system obtains a failed verification result.

[0053] The method for obtaining verification results provided by this optional implementation method can ensure that the processing results of the small model are consistent with the large model in logic and decision-making by controlling the large model to adopt the processing logic chain after filling in the decision parameter information, thereby improving the accuracy and reliability of the agricultural intelligent system.

[0054] In some optional implementations of the present disclosure, the above method also includes: in response to detecting that the verification result characterizes that the verification is unqualified, determining an alternative rule chain through the large model, and using the alternative rule chain to generate a new processing result of the information to be processed; or, in response to detecting that the verification result characterizes that the verification is unqualified, detecting whether there is a small model processing result in the small model processing result set that exceeds the range of the corresponding large model processing result, in response to detecting that a small model processing result exceeds the range of the corresponding large model processing result, limiting the range of the small model processing result to obtain a new processing result of the information to be processed and a range limitation strategy for the target small model corresponding to the small model processing result; or, in response to detecting that the verification result characterizes that the verification is unqualified, and the information to be processed is characterized as new variety information, obtaining the processing result of the information to be processed through the large model processing result set, and extracting similar area features corresponding to the information to be processed through the large model to generate a virtual training data set, and using the virtual training data set to train the target small model in the target small model set.

[0055] In this optional implementation, if the validation results indicate that the small model's processing results are unsatisfactory, the execution entity uses the large model to determine an alternative rule chain and employs it to regenerate a new processing result for the information being processed. This alternative rule chain is a re-formulated decision rule based on the knowledge and logic of the large model, replacing the original processing logic to ensure the accuracy and reliability of the processing results.

[0056] In this optional implementation, if the validation results indicate that the small model's processing results are unqualified, and if some of the small model's processing results are detected to be outside the acceptable range for the large model's processing results, the execution entity will limit the range of these out-of-range small model processing results. By adjusting these results to meet the range requirements of the large model, a new processing result for the information to be processed is obtained, and a range-limiting strategy is generated for the corresponding target small model to optimize the subsequent performance of the small model.

[0057] In this optional implementation, when processing new variety information, if the information being processed is characterized as new (e.g., a new crop variety or a new agricultural scenario) and the verification result fails, the execution entity directly obtains the processing results of the information being processed from the large model processing result set. Simultaneously, the large model extracts similar regional features corresponding to the information being processed to generate a virtual training dataset. This virtual training dataset is then used to retrain the target small model in the target small model set to improve the small model's ability to process new variety information.

[0058] The model collaborative processing method provided in this embodiment enables the executing entity to flexibly adjust the processing logic according to different verification failure situations, optimize the performance of the small model, and generate more accurate processing results, thereby improving the overall efficiency and reliability of agricultural intelligence.

[0059] In some optional implementations of the present disclosure, the above-mentioned sending of the information to be processed and the decision information prompt words to the big model to obtain the small model information and decision parameter information output by the big model includes: sending the information to be processed to the big model to obtain the standardized rule chain set and the initial small model set output by the big model; based on the standardized rule chain set and the initial small model set, determining the target small model set and the small model information of the target small model set; based on the target small model set, determining the target standardized rule chain; sending the target standardized rule chain and the decision information prompt words to the big model to obtain the decision parameter information output by the big model.

[0060] In this optional implementation, the information to be processed is sent to the large model. Based on this information, the large model outputs a standardized rule chain set and an initial small model set. The standardized rule chain set is a set of standardized rule chains, which includes at least two standardized rule chains, each standardized rule chain includes at least two standardized rules, each standardized rule corresponds to an execution step, and the execution step is used to standardize the processing process of the small model. The initial small model set is the small model that may be applicable that is preliminarily screened out by the large model based on the information to be processed. Based on the standardized rule chain set and the initial small model set, the execution subject further screens out the target small model set and extracts the relevant information of these target small models (i.e., small model information). The target small model set refers to those small models that are most suitable for processing the current information to be processed. The execution subject determines the target standardized rule chain from the target small model set. The target standardized rule chain is a rule chain selected or customized from the standardized rule chain set based on the characteristics of the target small model set, and is used to guide the subsequent processing logic.

[0061] In this optional implementation, the execution entity sends the target standardized rule chain and decision information prompts (such as specific decision requirements or problem descriptions) to the large model. Based on the target standardized rule chain and decision information prompts, the large model generates decision parameter information for the current information to be processed, such as specific irrigation amounts, fertilizer application rates, and pest and disease control measures.

[0062] This optional implementation provides a method for obtaining small model information and decision parameter information. First, the large model determines the standardized rule chain set and the initial small model set through the information to be processed; based on the standardized rule chain set and the initial small model set, the target small model set and small model information are determined; based on the target small model set, the target standardized rule chain is determined; the target standardized rule chain and decision information prompt words are sent to the large model to obtain the decision parameter information output by the large model. The execution subject can obtain the small model information and decision parameter information suitable for the current agricultural scenario from the large model, thereby providing support for subsequent processing and decision-making.

[0063] Optionally, the step of sending the information to be processed and the decision information prompt word to the large model to obtain the small model information and decision parameter information output by the large model includes: sending the information to be processed to the large model to obtain a target small model set output by the large model; determining a target standardized rule chain based on the target small model set; and sending the target standardized rule chain and the decision information prompt word to the large model to obtain the decision parameter information. The target standardized rule chain is a set of standardized rule chains applicable to the target small model set.

[0064] In some optional implementations of the present disclosure, the above-mentioned sending of the information to be processed to the target small model set includes: in response to detecting that the target small models in the target small model set have an execution order, determining the arrangement order of each target small model in the target small model set to obtain a sequence of target small models arranged in sequence; sending the information to be processed to the first target small model in the target small model sequence, and controlling each target small model in the target small model sequence to process data according to their respective execution orders.

[0065] In this optional implementation, when the execution entity needs to send information to be processed to a set of target small models, it first checks whether these small models have a specific execution order. If an execution order is detected, the execution entity determines the order of each target small model in the target small model set based on preset rules or logic, thereby generating a sequence of target small models arranged in sequence. For example, some small models may need to process basic data first, while other small models rely on the output results of the previous model for further analysis or optimization.

[0066] In this optional implementation, the execution entity sends the information to be processed to the first target model in the target model sequence. This model processes the information and generates a preliminary result. The execution entity then passes this preliminary result to the next target model in the sequence, which then continues processing. This process continues until all models in the target model sequence have completed their processing tasks according to their respective execution order.

[0067] This optional implementation method provides a method for sending information to be processed to a target small model set. When the target small models in the target small model set have an execution order, the execution subject sends the information to be processed to the first target small model in the target small model sequence, and controls each target small model in the target small model sequence to process data according to their respective execution order. It can ensure that each target small model processes data in the correct order, thereby realizing step-by-step processing and collaborative work of complex tasks, and improving overall processing efficiency and accuracy of results.

[0068] In some optional implementations of the present disclosure, the above-mentioned information to be processed includes: pest and disease diagnosis scenario information; the processing results of the information to be processed include: the identified pest and disease type and the pesticide corresponding to the pest and disease type, and the above-mentioned model collaborative processing method also includes: sending the pest and disease type and the pesticide to the big model, so as to use the big model to detect whether there is a pesticide of the same type as the pesticide in the local inventory; in response to not detecting the same type of pesticide as the pesticide in the local inventory, searching through the Internet to see whether there is a pesticide of the same type as the pesticide and for treating pests of the same type as the pesticide; in response to retrieving the same type of pesticide as the pesticide and for treating pests of the same type as the pesticide through the Internet, giving the source of the pesticide on the Internet.

[0069] In this embodiment, the standardized rule chain for pest and disease diagnosis and control scenarios is as follows: calling the crop classification model tool to determine the crop type; calling the crop pest and disease identification model to determine what kind of pest and disease the crop in the image has; recommending the use of pesticides based on the identified pest and disease type; finding pesticides of the same type based on the pesticides in the local inventory; and searching for suitable pesticides on the market through the Internet if there is no local inventory.

[0070] In this embodiment, the execution entity on which the model collaborative processing method runs can perform steps 1 to 4: Step 1: Maintain a pest and disease database and a local pesticide database.

[0071] Step 2: Receive the image uploaded by the user and generate information to be processed based on the image. The decision information prompt includes: Based on the image, identify the crop in the image and then determine the presence of pests and diseases on the crop. The information to be processed and the decision information prompt are sent to the large model. The large model determines the target set of small models, including a visual model and a pest and disease recognition model. The system calls the visual model (such as YOLO disease detection) to identify the crop in the image and calls the pest and disease model to identify the crop pest as "large leaf spot disease." The system already has two visual models for crop identification and pest and disease identification. The calling method of these two models is included in the rules written in the description of the standardized rule chain. In this case, large leaf spot disease is a prerequisite for pesticide selection.

[0072] Step 3: The large model searches the plant protection knowledge base and combines it with the processing results of the small models provided by the target small model set to generate a processing result. This processing result includes the control plan ("spray 25% azoxystrobin, dosage 50ml / mu") and verifies compliance with pesticide safety regulations (the small model verifies the dosage threshold). The large model uses the corresponding tools or databases according to the scenario rules to gradually achieve the expected results. The control plan is based on the identified disease, such as "big leaf spot", and then queries the pesticides and usage plans required to treat big leaf spot.

[0073] Step 4: If there is no local stock of azoxystrobin, recommend alternatives based on the substitution rule chain (such as "similar fungicide → active ingredient matching") and supplement procurement channel information.

[0074] In this embodiment, the information to be processed involves pest and disease diagnosis scenarios, specifically including characteristic information about the pests and diseases (such as symptoms and site of disease). The processing results are the identified pest and disease type and the corresponding recommended pesticide. After obtaining these processing results, the execution entity sends the pest and disease type and pesticide information to the large model. The large model first checks the local inventory to determine whether there is a pesticide of the same type as the recommended pesticide. If no matching pesticide is found in the local inventory, the execution entity searches the internet for a pesticide of the same type that can treat the pest and disease. If a matching pesticide is found, the execution entity provides the pesticide's internet source information to help the user obtain the required pesticide.

[0075] The model collaborative processing method provided in this embodiment can quickly and effectively determine the pesticides for treating pests of the same type for pest diagnosis scenarios, ensuring that when local resources are insufficient, users can quickly find alternative pesticides and improve the efficiency of pest control.

[0076] Further references Figure 3 As an implementation of the methods shown in the above figures, the present disclosure provides an embodiment of a model collaborative processing device. Figure 1 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0077] like Figure 3 As shown, the model collaborative processing device 300 provided in this embodiment includes: an acquisition unit 301, a decision unit 302, a determination unit 303, a sending unit 304, a receiving unit 305, and a obtaining unit 306. The acquisition unit 301 can be configured to acquire information to be processed in an agricultural scenario. The decision unit 302 can be configured to send the information to be processed and the decision information prompt to the large model, and obtain small model information and decision parameter information output by the large model. The determination unit 303 can be configured to determine a target small model set based on the small model information and send the decision parameter information to the relevant target small models in the target small model set. The sending unit 304 can be configured to send the information to be processed to the target small model set. The receiving unit 305 can be configured to receive a small model processing result set output by the target small model set. The obtaining unit 306 can be configured to obtain a processing result of the information to be processed based on the small model processing result set, the information to be processed, the small model information, the decision parameter information, and the large model.

[0078] In this embodiment, the specific processing and technical effects of the acquisition unit 301, the decision unit 302, the determination unit 303, the sending unit 304, the receiving unit 305, and the obtaining unit 306 in the model collaborative processing device 300 can be referred to respectively. Figure 1 The relevant descriptions of step 101, step 102, step 103, step 104, step 105, and step 106 in the corresponding embodiment are not repeated here.

[0079] In some embodiments of the present disclosure, the above-mentioned model collaborative processing device 300 also includes: the above-mentioned obtaining unit 306 is configured to: based on the information to be processed, the small model information and the decision parameter information, verify the small model processing result set through the large model to obtain the verification result; in response to detecting that the verification result represents that the verification is qualified, extract the processing result of the information to be processed in the small model processing result set.

[0080] In some embodiments of the present disclosure, the above-mentioned obtaining unit 306 is also configured to: determine the standardized rule chain of the target small model based on the small model information; fill in the standardized rule chain based on the decision parameter information to obtain the processing logic chain; control the large model to use the processing logic chain to process the information to be processed to obtain the large model processing result set; detect whether the small model processing result set and the large model processing result set are consistent; in response to detecting that the small model processing result set and the large model processing result set are consistent, obtain a verification result that is qualified; in response to detecting that the small model processing result set and the large model processing result set are inconsistent, obtain a verification result that is unqualified.

[0081] In some embodiments of the present disclosure, the model collaborative processing device 300 further includes: a generation unit (not shown in the figure), which is configured to: in response to detecting that the verification result indicates that the verification is unqualified, determine a replacement rule chain through the large model, and use the replacement rule chain to generate a new processing result for the information to be processed; or, in response to detecting that the verification result indicates that the verification is unqualified, detect whether there is a small model processing result in the small model processing result set that exceeds the range of the corresponding large model processing result, and in response to detecting that a small model processing result exceeds the range of the corresponding large model processing result, limit the range of the small model processing result to obtain a new processing result for the information to be processed and a range limitation strategy for the target small model corresponding to the small model processing result; or, in response to detecting that the verification result indicates that the verification is unqualified and the information to be processed is characterized as new variety information, obtain the processing result of the information to be processed through the large model processing result set, extract similar area features corresponding to the information to be processed through the large model, generate a virtual training data set, and use the virtual training data set to train the target small model in the target small model set.

[0082] In some embodiments of the present disclosure, the above-mentioned decision unit 302 is configured to: send the information to be processed to the big model to obtain the standardized rule chain set and the initial small model set output by the big model; determine the target small model set and the small model information of the target small model set based on the standardized rule chain set and the initial small model set; determine the target standardized rule chain based on the target small model set; send the target standardized rule chain and the decision information prompt word to the big model to obtain the decision parameter information output by the big model.

[0083] In some embodiments of the present disclosure, the above-mentioned sending unit 304 is configured to: in response to detecting that the target small models in the target small model set have an execution order, determine the arrangement order of each target small model in the target small model set to obtain a sequence of target small models arranged in sequence; send the information to be processed to the first target small model in the target small model sequence, and control each target small model in the target small model sequence to process data according to their respective execution order.

[0084] In some embodiments of the present disclosure, the above-mentioned information to be processed includes: pest and disease diagnosis scenario information, and the processing results of the above-mentioned information to be processed include: the identified pest and disease type and the pesticide corresponding to the pest and disease type. The above-mentioned device also includes an extraction unit (not shown in the figure), and the above-mentioned extraction unit is configured to: send the pest and disease type and the pesticide to the big model, so as to use the big model to detect whether there is a pesticide of the same type as the pesticide in the local inventory; in response to not detecting the pesticide of the same type as the pesticide in the local inventory, search through the Internet whether there is a pesticide of the same type as the pesticide and for treating pests of the same type as the pesticide; in response to retrieving the pesticide of the same type as the pesticide and for treating pests of the same type as the pesticide through the Internet, give the source of the pesticide on the Internet.

[0085] The model collaborative processing device provided by the embodiment of the present disclosure is as follows: first, the acquisition unit 301 acquires the information to be processed in the agricultural scene; secondly, the decision unit 302 sends the information to be processed and the decision information prompt word to the large model, and obtains the small model information and decision parameter information output by the large model; thirdly, the determination unit 303 determines the target small model set based on the small model information, and sends the decision parameter information to the relevant target small model in the target small model set; from then on, the sending unit 304 sends the information to be processed to the target small model set; then, the receiving unit 305 receives the small model processing result set output by the target small model set; finally, the acquisition unit 306 obtains the processing result of the information to be processed based on the small model processing result set, the information to be processed, the small model information, the decision parameter information and the large model. Thus, by combining the large model with the target small model to collaboratively process the information to be processed, this collaborative processing method fully utilizes the global decision-making ability of the large model and the refined processing ability of the small model, realizes the efficiency, accuracy and flexibility of information processing in the agricultural scene, can effectively improve the decision-making efficiency and management level of agricultural production, and provides strong support for the development of agricultural intelligence.

[0086] 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.

[0087] Figure 4 A schematic block diagram of an electronic device 400 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 assistants, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their modes are merely examples and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0088] like Figure 4 As shown, electronic device 400 includes a computing unit 401, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 402 or a computer program loaded from a storage unit 408 into a random access memory (RAM) 403. Various programs and data required for the operation of electronic device 400 may also be stored in RAM 403. Computing unit 401, ROM 402, and RAM 403 are interconnected via a bus 404. An input / output (I / O) interface 405 is also connected to bus 404.

[0089] Multiple components in the electronic device 400 are connected to the I / O interface 405, including an input unit 406, such as a keyboard, a mouse, etc.; an output unit 407, such as various types of displays, speakers, etc.; a storage unit 408, such as a magnetic disk, an optical disk, etc.; and a communication unit 409, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 409 allows the electronic device 400 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0090] Computing unit 401 can be any general-purpose and / or specialized processing component with processing and computing capabilities. Some examples of computing unit 401 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various specialized artificial intelligence (AI) computing chips, various computing units that run machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Computing unit 401 performs the various methods and processes described above, such as the model collaborative processing method. For example, in some embodiments, the model collaborative processing method can be implemented as a computer software program tangibly embodied in a machine-readable medium, such as storage unit 408. In some embodiments, part or all of the computer program can be loaded and / or installed onto electronic device 400 via ROM 402 and / or communication unit 409. When the computer program is loaded into RAM 403 and executed by computing unit 401, one or more steps of the model collaborative processing method described above can be performed. Alternatively, in other embodiments, computing unit 401 can be configured to perform the model collaborative processing method via any other suitable means (e.g., via firmware).

[0091] Various embodiments of the systems and techniques described above 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), system-on-chip systems (SOCs), programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include being implemented in one or more computer programs that are executable and / or interpreted on a programmable system that includes 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.

[0092] The program code for implementing the method of the present disclosure can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable model collaborative processing device, so that when the program code is executed by the processor or controller, the modes / operations specified in the flow chart and / or block diagram are implemented. The program code can be executed entirely on the machine, partially on the machine, as a stand-alone software package, partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0093] 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, apparatus, or device. 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, apparatus, or device, or any suitable combination of the foregoing. More specific examples of machine-readable storage media may include an electrical connection based on one or more wires, 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), optical fibers, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0094] 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).

[0095] The systems and techniques described herein can be implemented in a computing system that includes back-end 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 front-end 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 back-end components, middleware components, or front-end 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), and the Internet.

[0096] It should be understood that the various forms of the processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this disclosure can be performed in parallel, sequentially, or in a different order, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved. This is not limited herein.

[0097] The foregoing descriptions of specific exemplary embodiments of the present disclosure are for purposes of illustration and description. These descriptions are not intended to limit the present disclosure to the precise forms disclosed, and it is apparent that many variations and modifications are possible in light of the foregoing teachings. The exemplary embodiments have been selected and described for the purpose of explaining the specific principles of the present disclosure and their practical application, thereby enabling those skilled in the art to realize and utilize a variety of exemplary embodiments of the present disclosure and various options and modifications. The scope of the present disclosure is intended to be defined by the claims and their equivalents.

Claims

1. A model collaborative processing method, characterized in that: The method comprises: Obtain information to be processed in agricultural scenarios; Sending the information to be processed and the decision information prompt words to the large model to obtain the small model information and decision parameter information output by the large model; Determine a target small model set based on the small model information, and send the decision parameter information to the relevant target small models in the target small model set; Sending the information to be processed to the target small model set; Receive a set of small model processing results output by the target small model set; Based on the small model processing result set, the information to be processed, the small model information, the decision parameter information and the large model, a processing result of the information to be processed is obtained.

2. The method according to claim 1, characterized in that Obtaining a processing result of the information to be processed based on the small model processing result set, the information to be processed, the small model information, the decision parameter information, and the large model includes: Based on the information to be processed, the small model information and the decision parameter information, verifying the small model processing result set through the large model to obtain a verification result; In response to detecting that the verification result indicates that the verification is qualified, the processing result of the information to be processed in the small model processing result set is extracted.

3. The method according to claim 2, characterized in that The verification of the small model processing result set by the large model based on the information to be processed, the small model information and the decision parameter information includes: Determining a standardized rule chain of a target small model based on the small model information; Based on the decision parameter information, filling the standardized rule chain to obtain a processing logic chain; Controlling the large model to process the information to be processed using the processing logic chain to obtain a large model processing result set; Detecting whether the small model processing result set is consistent with the large model processing result set; In response to detecting that the small model processing result set and the large model processing result set are consistent, obtaining a verification result indicating that the verification is qualified; In response to detecting that the small model processing result set and the large model processing result set are inconsistent, a verification result indicating that the verification is unqualified is obtained.

4. The method according to claim 2 or 3, characterized in that The method further comprises: In response to detecting that the verification result indicates that the verification is unqualified, determining a replacement rule chain through the large model, and using the replacement rule chain to generate a new processing result for the information to be processed; Alternatively, in response to detecting that the verification result indicates that the verification is unqualified, detecting whether a small model processing result in the small model processing result set exceeds the range of the corresponding large model processing result; in response to detecting that a small model processing result exceeds the range of the corresponding large model processing result, performing a range restriction on the small model processing result to obtain a new processing result of the information to be processed and a range restriction strategy for the target small model corresponding to the small model processing result; Alternatively, in response to detecting that the verification result indicates that the verification is unqualified and the information to be processed is characterized as new variety information, the processing result of the information to be processed is obtained by processing the result set of the large model, and the similar area features corresponding to the information to be processed are extracted through the large model to generate a virtual training data set, and the virtual training data set is used to train the target small model in the target small model set.

5. The method according to any one of claims 1 to 3, wherein the sending of the information to be processed and the decision information prompt to the large model to obtain the small model information and decision parameter information output by the large model comprises: Sending the information to be processed to the large model to obtain a standardized rule chain set and an initial small model set output by the large model; Determine a target small model set and small model information of the target small model set based on the standardized rule chain set and the initial small model set; Determining a target normalization rule chain based on the target small model set; The target standardized rule chain and decision information prompt words are sent to the large model to obtain the decision parameter information output by the large model.

6. The method according to any one of claims 1 to 3, wherein sending the information to be processed to the target set of small models comprises: In response to detecting that the target small models in the target small model set have an execution order, determining an arrangement order of the target small models in the target small model set to obtain a sequence of target small models arranged in sequence; The information to be processed is sent to the first target small model in the target small model sequence, and each target small model in the target small model sequence is controlled to process data according to its own execution order.

7. The method according to any one of claims 1 to 3, wherein the information to be processed comprises: Pest and disease diagnosis scenario information; The processing result of the information to be processed includes: the identified pest and disease type and the pesticide corresponding to the pest and disease type, and the method further includes: Sending the pest type and the pesticide to the large model, so as to detect whether there is a pesticide of the same type as the pesticide in the local inventory using the large model; In response to not detecting a pesticide of the same type as the pesticide in the local inventory, searching via the Internet whether there is a pesticide of the same type as the pesticide that treats the pest of the type of pest; In response to searching through the Internet for a pesticide that is of the same type as the pesticide and treats the pests of the type, the source of the pesticide on the Internet is provided.

8. A model collaborative processing device, comprising: an acquisition unit, configured to acquire information to be processed in an agricultural scene; The decision unit is configured to send the information to be processed and the decision information prompt word to the large model, and obtain the small model information and decision parameter information output by the large model; a determining unit configured to determine a target small model set based on the small model information, and send the decision parameter information to relevant target small models in the target small model set; A sending unit, configured to send the information to be processed to the target small model set; A receiving unit, configured to receive a set of small model processing results output by the target small model set; The obtaining unit is configured to obtain a processing result of the information to be processed based on the small model processing result set, the information to be processed, the small model information, the decision parameter information and the large model.

9. An electronic device, characterized in that: include: at least one processor; and a memory communicatively coupled to the at least one processor; 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 7.

10. A non-transitory computer-readable storage medium storing computer instructions, characterized in that: The computer instructions are used to enable the computer to execute the method according to any one of claims 1 to 7.

Images