Task processing method, code task processing method and task platform

Abstract

The embodiment of the invention provides a task processing method, a code task processing method and a task platform. The task processing method comprises the steps of obtaining task data and task description information of a target task; based on the task description information, task guide information of the target task is screened out from the multiple pieces of candidate guide information, and the task guide information is used for describing task processing logic of the target task; determining model prompt information and a task processing model of the target task according to the task description information and the task guide information; and inputting the task data and the model prompt information into a task processing model to obtain a task processing result of the target task. According to the method, the accurate model prompt information is automatically generated for the target task by utilizing the task guide information of the target task, and the model prompt information is automatically routed to the task processing model matched with the target task, so that the task processing process is more flexible, and the task processing efficiency is improved.

Description

Technical Field

[0001] The embodiments in this specification relate to the field of computer technology, and in particular to task processing methods, code task processing methods, and task platforms. Background Technology

[0002] With the development of computer technology, large-scale models have begun to shine, demonstrating extraordinary capabilities in language understanding, generation, interaction, and reasoning, and are widely used in processing fields such as dialogue, translation, and code processing. Taking code processing as an example, large-scale models can provide developers with intelligent capabilities such as code completion and code annotation, and therefore have gradually become a research focus.

[0003] Currently, due to the high complexity of code completion and commenting tasks, large models suitable for the task are typically selected and processed manually, and corresponding model hints are manually written to guide the task processing. However, manually selecting suitable large models and writing model hints requires a significant amount of manpower, resulting in extremely low efficiency. Therefore, a more efficient task processing solution is urgently needed. Summary of the Invention

[0004] In view of the above, embodiments of this specification provide a task processing method. One or more embodiments of this specification also relate to a code task processing method, a task platform, a task processing device, a code task processing apparatus, a computing device, a computer-readable storage medium, and a computer program product, to address the technical deficiencies existing in the prior art.

[0005] According to a first aspect of the embodiments of this specification, a task processing method is provided, comprising: Obtain the task data and task description information for the target task; Based on the task description information, the task guidance information of the target task is selected from multiple candidate guidance information. The task guidance information is used to describe the task processing logic of the target task. Based on the task description information and task guidance information, determine the model prompt information and task processing model for the target task; Input the task data and model prompts into the task processing model to obtain the task processing results for the target task.

[0006] According to a second aspect of the embodiments of this specification, a code task processing method is provided, comprising: Obtain the task data and task description information of the target code task; Based on the task description information, the task guidance information of the target code task is selected from multiple candidate guidance information. The task guidance information is used to describe the task processing logic of the target code task. Based on the task description information and task guidance information, determine the model prompt information and task processing model of the target code task; Input the task data and model hints into the task processing model to obtain the task processing results of the target code task.

[0007] According to a third aspect of the embodiments of this specification, a task processing apparatus is provided, comprising: The first acquisition module is configured to acquire the task data and task description information of the target task. The first filtering module is configured to filter the task guidance information of the target task from multiple candidate guidance information based on the task description information. The task guidance information is used to describe the task processing logic of the target task. The first determining module is configured to determine the model prompt information and task processing model of the target task based on the task description information and task guidance information. The first input module is configured to input task data and model prompts into the task processing model to obtain the task processing result of the target task.

[0008] According to a fourth aspect of the embodiments of this specification, a code task processing apparatus is provided, comprising: The second acquisition module is configured to acquire the task data and task description information of the target code task. The second filtering module is configured to filter the task guidance information of the target code task from multiple candidate guidance information based on the task description information. The task guidance information is used to describe the task processing logic of the target code task. The second determining module is configured to determine the model prompt information and task processing model of the target code task based on the task description information and task guidance information. The second input module is configured to input task data and model prompts into the task processing model to obtain the task processing results of the target code task.

[0009] According to a fifth aspect of the embodiments of this specification, a task platform is provided, including a request interface and a response unit; The request interface is used to receive task processing requests for a target task. The response unit is used to respond to a task processing request, obtain the task data and task description information of the target task; based on the task description information, filter out the task guidance information of the target task from multiple candidate guidance information, wherein the task guidance information is used to describe the task processing logic of the target task; determine the model prompt information and task processing model of the target task according to the task description information and task guidance information; input the task data and model prompt information into the task processing model to obtain the task processing result of the target task.

[0010] According to a sixth aspect of the embodiments of this specification, a computing device is provided, comprising: Memory and processor; The memory is used to store computer programs / instructions, and the processor is used to execute the computer programs / instructions, which, when executed by the processor, implement the steps of the methods provided in the first or second aspect above.

[0011] According to a seventh aspect of the embodiments of this specification, a computer-readable storage medium is provided that stores a computer program / instructions that, when executed by a processor, implement the steps of the method provided in the first or second aspect described above.

[0012] According to an eighth aspect of the embodiments of this specification, a computer program product is provided, including a computer program / instructions that, when executed by a processor, implement the steps of the method provided in the first or second aspect described above.

[0013] This specification provides a task processing method according to one embodiment, comprising: acquiring task data and task description information of a target task; filtering task guidance information of the target task from multiple candidate guidance information based on the task description information, wherein the task guidance information is used to describe the task processing logic of the target task; determining model prompt information and a task processing model for the target task based on the task description information and the task guidance information; and inputting the task data and model prompt information into the task processing model to obtain the task processing result of the target task. By utilizing the task guidance information of the target task, accurate model prompt information is automatically generated for the target task, and the task is automatically routed to a task processing model adapted to the target task, thereby making the task processing process more flexible and improving task processing efficiency. Attached Figure Description

[0014] Figure 1 This is an architecture diagram of a task processing system provided in one embodiment of this specification; Figure 2 This is a flowchart illustrating a task processing method provided in one embodiment of this specification; Figure 3 This is a schematic diagram of a task processing logic diagram provided in one embodiment of this specification; Figure 4 This is a schematic diagram illustrating the generation process of model prompt information according to one embodiment of this specification; Figure 5 This is a schematic diagram of the screening process of a task processing model provided in one embodiment of this specification; Figure 6This is a schematic diagram of the test process of a task processing model provided in one embodiment of this specification; Figure 7 This is a flowchart illustrating the processing procedure of a code task processing method provided in one embodiment of this specification; Figure 8 This is a schematic diagram of the structure of a task platform provided in one embodiment of this specification; Figure 9 This is a schematic diagram of the structure of a task processing device provided in one embodiment of this specification; Figure 10 This is a schematic diagram of the structure of a code task processing device provided in one embodiment of this specification; Figure 11 This is a structural block diagram of a computing device provided in one embodiment of this specification. Detailed Implementation

[0015] Many specific details are set forth in the following description to provide a full understanding of this specification. However, this specification can be implemented in many other ways than those described herein, and those skilled in the art can make similar extensions without departing from the spirit of this specification. Therefore, this specification is not limited to the specific implementations disclosed below.

[0016] The terminology used in one or more embodiments of this specification is for the purpose of describing particular embodiments only and is not intended to be limiting of the one or more embodiments of this specification. The singular forms “a,” “described,” and “the” as used in one or more embodiments of this specification and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in one or more embodiments of this specification refers to and includes any or all possible combinations of one or more associated listed items.

[0017] It should be understood that although the terms first, second, etc., may be used to describe various information in one or more embodiments of this specification, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, first may also be referred to as second without departing from the scope of one or more embodiments of this specification, and similarly, second may also be referred to as first. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to a determination."

[0018] Furthermore, it should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in one or more embodiments of this specification are all information and data authorized by the user or fully authorized by all parties. Moreover, the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0019] In one or more embodiments of this specification, a large model refers to a deep learning model with a large number of model parameters, typically containing hundreds of millions, tens of billions, hundreds of billions, trillions, or even tens of trillions of model parameters. A large model can also be called a foundation model. It is pre-trained using large-scale unlabeled corpora to produce a pre-trained model with hundreds of millions of parameters. Such models can adapt to a wide range of downstream tasks and have good generalization ability. Examples include Large Language Models (LLMs) and multi-modal pre-training models.

[0020] In practical applications, large models only require a small number of samples to fine-tune the pre-trained model before they can be applied to different tasks. Large models can be widely used in fields such as Natural Language Processing (NLP) and Computer Vision. Specifically, they can be applied to computer vision tasks such as Visual Question Answering (VQA), Image Captioning (IC), and Image Generation, as well as natural language processing tasks such as text-based sentiment classification, text summarization, and machine translation. The main application scenarios of large models include digital assistants, intelligent robots, search, online education, office software, e-commerce, and intelligent design.

[0021] First, the terms and concepts used in one or more embodiments of this specification will be explained.

[0022] A / B testing is a statistical method commonly used to conduct gray-scale testing of different algorithm models to determine which model performs better. Traffic splitting or user splitting are two common methods in A / B testing.

[0023] Traffic splitting refers to an A / B testing approach that divides traffic based on a certain percentage of request volume. For example, if it's decided to conduct a 5% canary rollout test on a new feature, the system can be configured so that only 5% of total traffic is redirected to the new version, while the remaining 95% of traffic continues to use the old version. This approach is suitable for situations requiring rapid response to traffic changes.

[0024] User-based routing refers to an A / B testing method that assigns users to a specific version based on specific rules (such as user ID hashes). User-based routing ensures that the same user sees the same version throughout the entire test, which helps to collect more consistent and reliable data. In addition, it allows for more granular grouping based on user characteristics (such as geographic location, device type, etc.).

[0025] Tracking: This refers to the process of inserting code snippets at key locations in a model to record user behavior data.

[0026] Hypertext Transfer Protocol (HTTP) is an application-layer protocol used to transmit hypermedia information (such as web pages). It is the primary protocol on the Internet for distributed, collaborative, and hypermedia information systems, mainly used for communication between clients (such as browsers) and servers.

[0027] Code projects are typically complex. For example, code completion and Q&A tasks often involve multiple models. The user's task type and behavior influence the selection of prompts and models, and these prompts and models are not static but constantly updated. Furthermore, besides differences in models and prompts, the same model may be deployed in different regions or on clusters of varying sizes. Different resource allocations and project requirements will also lead to variations in the clusters and models used. During online operation, due to limited model resources, there is a risk of insufficient resources preventing service provision. Additionally, user feedback on model effectiveness or performance may necessitate switching traffic allocations and model configurations in emergency situations to ensure project stability.

[0028] Currently, model hints are typically written manually, and suitable large models are selected manually. Manually written model hints can be stored in the following locations: First, placing the model hints, task execution logic, and model configuration within the code. This method requires updating the model hints via server-side deployment, which is inefficient and cannot handle online emergencies in seconds. Second, placing the model hints on the client-side. This method requires updating the client version to update the effect, and different client versions cannot control the effect. If problematic hints are encountered, a full rollback is not possible. Third, placing the model hints in a configuration service. This method can only configure limited content, such as simple hints and model names. However, these configurations cannot be strongly bound to the code logic, and the amount of configuration required is large, making configuration service management difficult and dependent on external product services. Fourth, placing the model hints in a storage service and retrieving them periodically. This method is not timely in configuration updates, depends on external products, and can only configure large amounts of text such as hints. It is difficult to handle small, scattered configurations and cannot be strongly bound to the code logic.

[0029] To address the aforementioned issues, this specification proposes a model prompting information and model orchestration scheme based on task guidance information. The scheme involves acquiring task data and task description information for the target task; filtering task guidance information from multiple candidate guidance information based on the task description information, whereby the task guidance information describes the task processing logic of the target task; determining the model prompting information and task processing model for the target task based on the task description information and task guidance information; and inputting the task data and model prompting information into the task processing model to obtain the task processing result for the target task. By utilizing the task guidance information of the target task, accurate model prompting information is automatically generated for the target task, and the task is automatically routed to a task processing model adapted to the target task. Furthermore, the capabilities of the task processing model are leveraged to provide users with intelligent task processing capabilities, improving the efficiency of the task processing process and the quality of the task processing results.

[0030] This specification provides a task processing method, and also relates to a code task processing method, a task platform, a task processing device, a code task processing device, a computing device, a computer-readable storage medium, and a computer program product, which will be described in detail in the following embodiments.

[0031] Considering the large number of model parameters in the task processing model and the limited computing resources of the client, the task processing method proposed in the embodiments of this specification can be applied to, for example, Figure 1 The task processing system shown is not limited to this. See also Figure 1 , Figure 1This specification illustrates an architecture diagram of a task processing system provided in one embodiment of the specification. The task processing system may include a client 100 and a server 200. Client 100 is used to send the target task's task data and task description information to server 200; Server 200 is used to filter the task guidance information of the target task from multiple candidate guidance information based on the task description information. The task guidance information describes the task processing logic of the target task. Based on the task description information and the task guidance information, the server determines the model prompt information and task processing model of the target task. The server inputs the task data and model prompt information into the task processing model to obtain the task processing result of the target task. The server sends the task processing result of the target task to the client 100. Client 100 is also used to receive the task processing results of the target task sent by server 200.

[0032] like Figure 1 As shown, the task processing model is deployed in server 200. Server 200 can connect to one or more clients 100 via a local area network (LAN), wide area network (WAN), internet connection, or other types of data network. Client 100 may include, but is not limited to, smartphones, tablets, laptops, PDAs, personal computers, smart home devices, and in-vehicle devices. Client 100 can also interact with users through a graphical user interface to invoke the task processing model, thereby implementing the task processing method provided in the embodiments of this specification.

[0033] It is worth noting that the task processing methods provided in the embodiments of this specification are generally executed by the server. However, in other embodiments of this specification, if the client's runtime resources can meet the deployment and runtime conditions of the task processing model, the client can also have similar functions to the server, thereby executing the task processing methods provided in the embodiments of this specification. In other embodiments, the task processing methods provided in the embodiments of this specification can also be executed jointly by the client and the server. Next, taking the server executing the task processing methods provided in the embodiments of this specification as an example, the task processing methods will be described in detail.

[0034] See Figure 2 , Figure 2 This specification shows a flowchart of a task processing method according to an embodiment, which specifically includes the following steps: Step 202: Obtain the task data and task description information of the target task.

[0035] It should be noted that the target task can be a task in different scenarios, such as a code task in a code processing scenario, a text processing task in a text processing scenario, and so on. Code tasks include, but are not limited to, code question-and-answer tasks, code generation tasks, code completion tasks, code explanation generation tasks, and code comment generation tasks. Text processing tasks include, but are not limited to, text question-and-answer tasks, text summarization extraction tasks, and translation tasks. The task data of the target task refers to information related to completing the target task. Taking a code comment generation task in a code processing scenario as an example, the task data can be a piece of code. Taking a code generation task as an example, the task data can be a text description of the code to be generated. The task description information is a detailed description of the target task, used to make the server understand the requirements of the target task. The task description information includes at least one of the following: task type, request object behavior, request resource information, and task processing requirements. The specific selection depends on the actual situation, and this specification does not impose any limitations on this. The task type refers to the type of task to which the target task belongs. Taking a code task as an example, the task type of the target task may be a free-response question-and-answer type, a unit test generation type, a comment generation type, a code explanation type, etc. The request object refers to the behavioral information of the requesting object that sends the task processing request, such as whether the user's behavior is an initial question-and-answer session or a retry question-and-answer session. Request resource information refers to the resource information required by the task processing request, such as resource location, resource access permissions, etc. Task processing requirements refer to the conditions or specifications required to complete the target task, such as the required hardware and software environment, time constraints, quality standards, etc.

[0036] In practical applications, there are various ways to obtain the task data and task description information of the target task, and the specific method should be selected according to the actual situation. This specification does not impose any limitations on these methods in the embodiments. In one possible implementation of this specification, the target task's task data and task description information sent by the user through a client can be received.

[0037] In another possible implementation of this specification, the task data of the target task can be obtained, the task data can be parsed, and the task description information of the target task can be obtained. That is, the above-mentioned acquisition of the task data and task description information of the target task may include the following steps: In response to a task processing request for a target task, retrieve the task data for the target task; Key information is extracted from the task data to obtain the task description information of the target task. The task description information includes at least one of the following: task type, request object behavior, request resource information, and task processing requirements.

[0038] It should be noted that a task processing request for a target task refers to a request made to the server in order to complete the target task. A task processing request typically includes the instructions, data, and resource requirements needed to complete the target task. Therefore, after receiving a task processing request for a target task, the server can parse the task data of the target task from the task processing request.

[0039] In practical applications, there are various ways to extract key information from task data to obtain the task description information of the target task. The specific method chosen depends on the actual situation, and the embodiments in this specification do not impose any limitations on this. One possible implementation of this specification involves obtaining a key information template and matching the task data with the key information template to obtain the task description information of the target task. Another possible implementation involves inputting the task data into a key information extraction model to obtain the task description information of the target task output by the key information extraction model. The key information extraction model can be a pre-trained large model or a deep learning model trained based on sample task data and the corresponding sample description information.

[0040] By applying the solutions in the embodiments of this specification, task data and task description information of the target task are obtained by extracting task data of the target task from the task processing request and extracting key information from the task data. This allows for efficient and accurate acquisition of task data and task description information.

[0041] Step 204: Based on the task description information, filter out the task guidance information of the target task from multiple candidate guidance information, wherein the task guidance information is used to describe the task processing logic of the target task.

[0042] It should be noted that candidate guidance information refers to guidance information set for different candidate tasks. Each candidate guidance information corresponds one-to-one with a candidate task, and the candidate guidance information describes the task processing logic of the corresponding candidate task. The target task's task guidance information describes the task processing logic to be followed when completing the target task. Through the task guidance information, the server can clearly understand how to correctly and efficiently complete the target task. Task guidance information can be in text form; for example, it could be something like, "First, use the model hint template to determine the model hint information for the target task; then, based on the target task's task description information, route the task processing model used to process the target task; finally, use the task processing model to process the target task's task data based on the model hint information to obtain the task processing result." Task guidance information can also be in graph form; the choice depends on the specific situation, and this specification does not impose any limitations on this aspect in the embodiments.

[0043] In practical applications, there are various ways to select the task guidance information for the target task from multiple candidate guidance information based on task description information. The specific method should be selected according to the actual situation, and the embodiments in this specification do not impose any limitations on this. In one possible implementation of this specification, the text similarity between the task description information and each candidate guidance information can be directly calculated, and the candidate guidance information with the higher text similarity can be determined as the task guidance information for the target task.

[0044] In another possible implementation of this specification, the candidate guidance information includes a candidate processing logic diagram, and the task guidance information includes a task processing logic diagram; the above-mentioned selection of the target task's task guidance information from multiple candidate guidance information based on task description information may include the following steps: Obtain multiple candidate processing logic diagrams, each carrying a candidate task identifier; The target task is matched with the candidate task identifiers to obtain a second matching result, which is then used in conjunction with the candidate processing logic. Figure 1 One-to-one correspondence; Based on the second matching result, the task processing logic diagram of the target task is selected from multiple candidate processing logic diagrams.

[0045] It should be noted that the task processing logic graph can also be called the task logic computation graph. The task processing logic graph is a directed acyclic graph (DAG) formed by connecting graph nodes and edges. Candidate task identifiers are used to uniquely identify candidate tasks, such as unique task numbers. The second matching result describes the degree of matching between the target task and the candidate tasks. The second matching result can be a matching value, such as 80%, or a matching level, such as very good match or no match.

[0046] For example, see Figure 3 , Figure 3 This diagram illustrates a task processing logic diagram according to one embodiment of this specification. Figure 3 The task processing logic diagram shown includes multiple graph nodes, namely the start node, model hint assembly node, task type judgment node, model A node, model B node, and end node. These graph nodes are categorized into three types: ordinary nodes, routing nodes (which branch routes to different child nodes based on conditions), and merging nodes (which merge multiple links into one node). Ordinary nodes, such as... Figure 3 The starting node, model hint assembly node, model A node, and model B node are listed. Routing nodes are as follows: Figure 3 The task type determination node in the middle. Merging nodes, such as Figure 3 The end node in.

[0047] In practical applications, there are various ways to obtain multiple candidate processing logic diagrams, and the specific method should be selected according to the actual situation. This specification does not impose any limitations on the embodiments described herein. In one possible implementation, multiple candidate processing logic diagrams can be read from other data acquisition devices or databases. In another possible implementation, multiple candidate task processing logic diagrams can be obtained by constructing configurations based on the candidate processing logic of multiple candidate tasks.

[0048] It's worth noting that multiple candidate processing logic graphs can be configured within a graph service. The server can use a unified HTTP interface to call the graph service, thereby selecting the target task's processing logic graph from multiple candidate graphs based on task description information. When configuring a graph service, the service name can be defined according to specific requirements. The inputs and outputs of the graph service, as well as the intermediate processes, can all use parameter variable sets, which are stored in a dictionary. Each graph service can contain several graph service versions, with at most one enabled version. Each graph service version corresponds to a specific candidate processing logic graph.

[0049] By applying the solution of the embodiments in this specification, the task processing logic diagram of the target task is selected from multiple candidate processing logic diagrams based on the second matching result, thereby improving the filtering efficiency of task guidance information of the target task.

[0050] In one optional embodiment of this specification, obtaining multiple candidate processing logic diagrams as described above may include the following steps: Obtain the candidate processing logic for multiple candidate tasks respectively; For the first candidate task, multiple processing nodes are extracted from the candidate processing logic of the first candidate task. These multiple processing nodes include prompt nodes and multiple model nodes. The first candidate task is any one of the multiple candidate tasks. An initial processing logic graph is constructed using multiple processing nodes as graph nodes and the logical relationships between multiple processing nodes as edges. The initial processing logic graph is used to describe the candidate processing logic of the first candidate task. Configure the model prompt template for the first candidate task on the prompt node, and configure the candidate processing model for the first candidate task on multiple model nodes to obtain the candidate processing logic diagram of the first candidate task.

[0051] It's important to note that a prompt node is a graph node configured with a model prompt template, while a model node is a graph node configured with a candidate processing model. Each graph node corresponds to a segment of task processing logic; for example, each graph node corresponds to a segment of Java code logic in a code processing task. The code in the graph node is stateless, with inputs and outputs being global parameter dictionaries. Each graph node is configured with editable JSON text, allowing you to abstract variables from the task processing logic to form keys in the JSON, making them configurable fields. The task processing logic of each graph node is connected by directed connections to form a complete directed acyclic graph.

[0052] In practical applications, there are various ways to extract multiple processing nodes from the task processing logic of the first candidate task. The specific method should be selected according to the actual situation, and the embodiments in this specification do not impose any limitations on this. In one possible implementation of this specification, key information can be extracted from the task processing logic of the first candidate task, and then the processing nodes can be divided based on the extracted key information. In another possible implementation of this specification, the task processing logic of the first candidate task can be matched with preset processing nodes, and then multiple processing nodes in the task processing logic can be determined based on the matching results.

[0053] The solution implemented in this specification constructs a directed acyclic graph on the server side, allowing for highly flexible assembly of task processing logic. This adapts to rapidly evolving domain requirements, supports complex and ever-changing task processing needs, provides configurable model hint information, and enables the modification of model hint templates at any time to optimize model output. Furthermore, integrating the construction of model hint information and the arrangement and selection of models into the task processing logic graph can improve the iteration efficiency of model hint information and models.

[0054] Step 206: Based on the task description information and task guidance information, determine the model prompt information and task processing model for the target task.

[0055] It's important to note that model prompts (Prompts) are guidance provided to machine learning models to help them understand and generate specific types of output. Model prompts can be words or statements, and they can be continuously modified and improved as the model evolves and requirements change. A task-processing model refers to an algorithm or machine learning model used to handle a target task.

[0056] In practical applications, there are various ways to determine the model prompts and task processing model for the target task based on task description information and task guidance information. The specific method chosen depends on the actual situation, and this specification does not impose any limitations on these methods. In one possible implementation, the task description information and task guidance information can be input into a text processing model to obtain the model prompts and model identification information output by the text processing model. The text processing model is trained based on sample description information, sample guidance information, sample prompts, and the model identification information of the sample models.

[0057] In one possible implementation of this specification, determining the model prompt information and task processing model of the target task based on the task description information and task guidance information may include the following steps: Analyze the task guidance information to determine the model prompt template for the target task and multiple candidate processing models; By integrating task description information and model hint templates, model hint information for the target task can be obtained. Based on the task description information, the task processing model for the target task is selected from multiple candidate processing models.

[0058] It's important to note that model suggestion templates are pre-designed text structures used to guide the model in generating more accurate, relevant, and useful responses. Model suggestion templates can be updated and replaced as needed. Taking code processing tasks as an example, model suggestion templates can be pre-configured or automatically generated using text template code tools (such as FreeMarker, Thymeleaf, Enjoy, and Velocity). This allows model suggestion templates to support statements like logical loops, making the generation of model suggestion information more flexible. Candidate processing models are models configured for candidate tasks; these models are used to process candidate tasks. Multiple candidate processing models configured on different model nodes all have the ability to process candidate tasks, but these models may have different model parameters or model sizes.

[0059] In practical applications, after extracting the model hint template and multiple candidate processing models from the task guidance information, the task description information can be filled into the model hint template to obtain the model hint information for the target task. Since different candidate processing models correspond to different candidate description information, the candidate processing model that matches the candidate description information with the task description information can be selected from the multiple candidate processing models based on the task description information and the candidate description information of multiple candidate processing models, and this model can be used as the task processing model for the target task.

[0060] See Figure 4 , Figure 4This diagram illustrates a process for generating model hint information according to an embodiment of this specification. Assuming the client's task processing request is a code explanation request, the task description information for the code explanation request is determined to be "Task Requirements: Use English, detailed explanation, task type is code explanation type," and the task data is "select code." The server processes the request, executes the graph service, obtains the task processing logic graph corresponding to the task processing request, reads the model hint template from the hint node in the task processing logic graph, fills the model hint template with the task description information, and generates the model hint information "Explain the following code, following additional requirements: Use English, detailed explanation," and determines the task processing model to be a code explanation model based on the task type.

[0061] By applying the solutions in the embodiments of this specification, through the task processing logic diagram, the client-side model can encapsulate the processing requests using different model prompts based on task description information such as task type and request object behavior, and route them to the appropriate task processing model, making the task processing logic orchestration more flexible and visual.

[0062] In one optional embodiment of this specification, the candidate processing model carries candidate description information, which includes at least one of task type, request object behavior, request resource information, and task processing requirements. The process of selecting the target task's task processing model from multiple candidate processing models based on the task description information may include the following steps: The task description information is matched with the candidate description information to obtain the first matching result, wherein the first matching result corresponds one-to-one with the candidate processing model; Based on the first matching result, the task processing model for the target task is selected from multiple candidate processing models.

[0063] It should be noted that the first matching result is used to describe the degree of matching between the task description information and the candidate description information. The first matching result can be a matching value, such as 90%, or a matching level, such as very good match or no match.

[0064] In practical applications, there are multiple ways to match task description information with candidate description information to obtain the first matching result. The specific method chosen depends on the actual situation, and this specification does not limit this approach. In one possible implementation, the similarity between the task description information and the candidate description information can be calculated to obtain the first matching result. The similarity calculation methods include, but are not limited to, cosine similarity calculation and Euclidean distance calculation. In another possible implementation, task description keywords can be extracted from the task description information, candidate description keywords can be extracted from the candidate description information, the similarity between the task description keywords and the candidate description keywords can be calculated, and this similarity can be determined as the first matching result.

[0065] See Figure 5 , Figure 5 This diagram illustrates a task processing model selection process according to an embodiment of this specification. The task processing logic diagram includes the following nodes: a start node, a model hint assembly node, a task type determination node, a model A node, a model B node, a model C node, and an end node. Assume that model A in the model A node is used to process code interpretation type tasks, model B in the model B node is used to process code unit testing type tasks, and model C in the model C node is used to process code optimization type tasks. At the task type determination node, the task type of the target task is determined to be code unit testing type; therefore, model B is selected as the task processing model for the target task.

[0066] The solution implemented in the embodiments of this specification selects the task processing model of the target task from multiple candidate processing models based on the first matching result between the task description information and the candidate description information. This achieves flexible routing logic configuration and supports assigning the target task to different models based on different task description information, thus efficiently and accurately determining the task processing model.

[0067] Step 208: Input the task data and model prompts into the task processing model to obtain the task processing result of the target task.

[0068] It should be noted that the task processing result is related to the task type of the target task. If the target task's task type is code interpretation, the task processing result will be the code interpretation information. If the target task's task type is code optimization, the task processing result will be the optimized code.

[0069] The solution implemented in the embodiments of this specification utilizes the task guidance information of the target task to automatically generate accurate model prompts for the target task and automatically routes to the task processing model adapted to the target task, thereby making the task processing process more flexible and improving task processing efficiency.

[0070] In one optional embodiment of this specification, the task description information includes model configuration parameters; before inputting the task data and model prompt information into the task processing model to obtain the task processing result of the target task, the following steps may also be included: Configure the parameters of the task processing model according to the model configuration parameters to obtain the updated task processing model; Inputting task data and model hints into the task processing model to obtain the task processing result for the target task may include the following steps: Input the task data and model hints into the updated task processing model to obtain the task processing result of the target task.

[0071] It's important to note that the task description information may include parameters configured by the user for the task processing model used to handle the target task. These parameters, called model configuration parameters, can affect the model's performance, output quality, and generation process. Model configuration parameters include Temperature, Top-K, and Top-P. The Temperature parameter controls the randomness of the model's generated results, determining the probability distribution when generating the next word. The Top-K parameter specifies that the model only considers the top K words with the highest probabilities when generating the next word. The Top-P parameter specifies that only words with a cumulative probability sum of P are considered when generating the next word. Therefore, the original model parameters of the task processing model can be modified to match the model configuration parameters in the task description information to obtain an updated task processing model that matches the user's needs.

[0072] By applying the solutions in the embodiments of this specification, the task processing model is configured with parameters according to the model configuration parameters in the task description information of the target task, so that the updated task processing model is more compatible with the target task and the accuracy of the task processing results of the target task is improved.

[0073] In one optional embodiment of this specification, before inputting task data and model prompt information into the task processing model to obtain the task processing result of the target task, the following steps may be included: If the task processing model does not meet the task processing conditions, multiple candidate processing models for the target task are obtained by parsing from the task guidance information. Select the target switching model for the target task from the candidate processing models other than the task processing model; Inputting task data and model hints into the task processing model to obtain the task processing result for the target task may include the following steps: Input the task data and model prompts into the target to switch models and obtain the task processing results for the target task.

[0074] It should be noted that task processing conditions refer to a set of requirements or standards that a task processing model must meet to successfully complete the target task. Task processing conditions can be technical, logical, environmental, or other types of constraints. Only when the task processing model meets the task processing conditions can it effectively execute the target task. If the task processing model does not meet the task processing conditions, measures can be taken to adjust or replace the model that does not meet the conditions. This can be achieved by using the routing function of the task processing logic graph to switch the graph node A of requesting model A to the graph node B of requesting model B, i.e., "model switching". Task processing conditions include, but are not limited to, the following: Model capability: Whether the model's capabilities are sufficient to handle the current task, such as whether the model's training data covers the knowledge or skills required for the current task. Security requirements: The task may require meeting specific security standards or compliance requirements. Performance metrics: The performance metrics that the model needs to achieve when processing the task, such as response time and accuracy. The specific task processing conditions are selected based on the actual situation, and this specification does not impose any limitations on this. The target switching model refers to a model other than the task processing model that meets the task processing conditions among multiple candidate processing models.

[0075] By applying the solutions in the embodiments of this specification, emergency flow switching is performed when the task processing model does not meet the task processing conditions, thereby improving the stability and security of the task processing service.

[0076] In one optional embodiment of this specification, after inputting the task data and model prompt information into the task processing model to obtain the task processing result of the target task, the following steps may be further included: Receive feedback information from the client regarding the task processing results; Based on the feedback information, generate model test results for the task processing model.

[0077] It's important to note that result feedback information refers to the information provided by the client to the server regarding the validity and satisfaction of the task processing results. This information helps the server evaluate the model's performance and make improvements accordingly. Model test results refer to the conclusions drawn after evaluating the task processing model based on the received result feedback information. These conclusions typically include the model's performance on a specific task, existing problems, and directions for improvement.

[0078] In practical applications, after receiving the result feedback information corresponding to the task processing model, the server can perform ABTEST to experiment with the effects of different versions of the task processing model, and further divide the traffic into grayscale phases by traffic or by user, and evaluate the model effect with data tracking and reports, and then decide whether to change the model version.

[0079] For example, see Figure 6 , Figure 6 This diagram illustrates a test process for a task processing model provided in one embodiment of this specification. Traffic can be allocated to different models or Prompt processing logic based on user identifiers or random proportions. For example, an A / B test traffic allocator can allocate 40% of the traffic to model A, 30% to model B, and 30% to model C, finally ending the model test. If model A malfunctions, 40% of the traffic for model A can be urgently switched to models B and C. By combining traffic allocation routing functionality with model node configuration capabilities (placing the model name and a series of configurations in the JSON structure of the model node), online emergency traffic switching and effect repair are supported.

[0080] By applying the solutions in the embodiments of this specification, model test results of the task processing model are generated based on the result feedback information, thereby accurately determining the model processing capability of the task processing model.

[0081] In one optional embodiment of this specification, after generating the model test results of the task processing model based on the result feedback information, the following steps may be further included: If the model test results do not meet the model test conditions, adjust the task guidance information for the target task; or... If the model test results do not meet the model test conditions, adjust the model parameters of the task processing model.

[0082] It should be noted that model testing conditions refer to a series of standards or requirements set when evaluating model performance. These standards are used to determine whether the model has achieved the expected performance level. Model testing conditions typically include, but are not limited to, the following aspects: Performance Metrics: such as accuracy, recall, etc. Error Tolerance: such as maximum allowed error rate, fault tolerance. Response Time.

[0083] In practical applications, when the model test results do not meet the model test conditions, one possible approach is to adjust the task guidance information of the target task. For example, if the previous task guidance information identified model A as the task processing model for the target task, then model B can be adjusted to become the task processing model for the target task. Another possible approach is that if the model test results do not meet the model test conditions, it indicates that the task processing model's processing capability is insufficient. In this case, the task processing model can be trained to improve its processing capability.

[0084] By applying the solutions in the embodiments of this specification, when the model test results do not meet the model test conditions, the task guidance information of the target task and the model parameters of the task processing model are adjusted, thereby improving the accuracy of the task guidance information or the task processing model, and further improving the accuracy of the task processing results.

[0085] The following is in conjunction with the appendix Figure 7 Taking the application of the task processing method provided in this specification in the field of code processing as an example, the task processing method will be further explained. Figure 7 This specification illustrates a flowchart of a code task processing method according to an embodiment, which specifically includes the following steps: Step 702: Obtain the task data and task description information of the target code task.

[0086] Step 704: Based on the task description information, filter out the task guidance information of the target code task from multiple candidate guidance information, wherein the task guidance information is used to describe the task processing logic of the target code task.

[0087] Step 706: Based on the task description information and task guidance information, determine the model prompt information and task processing model of the target code task.

[0088] Step 708: Input the task data and model prompts into the task processing model to obtain the task processing results of the target code task.

[0089] It should be noted that the implementation methods of steps 702 to 708 are the same as those of steps 202 to 208 above, and will not be described again in the embodiments of this specification.

[0090] In practical applications, taking the target code task as a code interpretation task as an example, we can obtain the task data and task description information of the code interpretation task; call the graph service through a unified HTTP interface to filter out the task processing logic graph corresponding to the code interpretation task from multiple candidate processing logic graphs; parse the task processing logic graph to determine the model hint template and multiple candidate processing models for the code interpretation task; integrate the task description information and the model hint template to obtain the model hint information for the code interpretation task; based on the task description information, filter out the task processing model for the code interpretation task from multiple candidate processing models; input the task data and model hint information into the task processing model to obtain the task processing result of the code interpretation task.

[0091] The solution implemented in this specification utilizes the task guidance information of the target code task to automatically generate accurate model prompts for the target code task and automatically routes it to the task processing model adapted to the target code task. This makes the code task processing process more flexible. Through the capabilities of the task processing model, it provides developers with intelligent capabilities such as code completion and code Q&A, thereby improving coding efficiency and code quality.

[0092] See Figure 8 , Figure 8 This specification shows a schematic diagram of the structure of a task platform 800 provided in one embodiment of the specification. The task platform 800 includes a request interface 802 and a response unit 804. Request interface 802 is used to receive task processing requests for a target task; The response unit 804 is used to respond to a task processing request, obtain the task data and task description information of the target task; based on the task description information, filter out the task guidance information of the target task from multiple candidate guidance information, wherein the task guidance information is used to describe the task processing logic of the target task; determine the model prompt information and task processing model of the target task according to the task description information and task guidance information; input the task data and model prompt information into the task processing model to obtain the task processing result of the target task.

[0093] By applying the solutions in the embodiments of this specification, the task platform adapts to user needs for task processing, realizes personalized model services, provides users with an efficient, flexible and easy-to-use model service platform, and improves user experience.

[0094] The above is an illustrative scheme of a task platform according to this embodiment. It should be noted that the technical solution of this task platform and the technical solution of the task processing method described above belong to the same concept. For details not described in detail in the technical solution of the task platform, please refer to the description of the technical solution of the task processing method described above.

[0095] Corresponding to the above-described task processing method embodiments, this specification also provides embodiments of a task processing apparatus. Figure 9 A schematic diagram of a task processing apparatus according to one embodiment of this specification is shown. Figure 9 As shown, the device includes: The first acquisition module 902 is configured to acquire the task data and task description information of the target task. The first filtering module 904 is configured to filter the task guidance information of the target task from multiple candidate guidance information based on the task description information, wherein the task guidance information is used to describe the task processing logic of the target task. The first determining module 906 is configured to determine the model prompt information and task processing model of the target task based on the task description information and task guidance information. The first input module 908 is configured to input task data and model prompt information into the task processing model to obtain the task processing result of the target task.

[0096] Optionally, the first determining module 906 is further configured to parse the task guidance information, determine the model prompt template of the target task and multiple candidate processing models; integrate the task description information and the model prompt template to obtain the model prompt information of the target task; and select the task processing model of the target task from multiple candidate processing models based on the task description information.

[0097] Optionally, the candidate processing model carries candidate description information, which includes at least one of task type, request object behavior, request resource information, and task processing requirements; the first determining module 906 is further configured to match the task description information with the candidate description information to obtain a first matching result, wherein the first matching result corresponds one-to-one with the candidate processing model; and based on the first matching result, the task processing model of the target task is selected from multiple candidate processing models.

[0098] Optionally, the candidate guidance information includes a candidate processing logic diagram, and the task guidance information includes a task processing logic diagram; the first filtering module 904 is further configured to acquire multiple candidate processing logic diagrams, wherein the candidate processing logic diagram carries a candidate task identifier; and to match the target task with the candidate task identifier to obtain a second matching result, wherein the second matching result matches the candidate processing logic diagram. Figure 1 One-to-one correspondence; based on the second matching result, the task processing logic diagram of the target task is selected from multiple candidate processing logic diagrams.

[0099] Optionally, the first filtering module 904 is further configured to: obtain candidate processing logic for multiple candidate tasks respectively; for a first candidate task, extract multiple processing nodes from the candidate processing logic of the first candidate task, wherein the multiple processing nodes include prompt nodes and multiple model nodes, and the first candidate task is any one of the multiple candidate tasks; construct an initial processing logic graph using the multiple processing nodes as graph nodes and the logical relationships between the multiple processing nodes as edges, wherein the initial processing logic graph is used to describe the candidate processing logic of the first candidate task; configure the model prompt template of the first candidate task on the prompt node, and configure the candidate processing model of the first candidate task on the multiple model nodes to obtain the candidate processing logic graph of the first candidate task.

[0100] Optionally, the first acquisition module 902 is further configured to, in response to a task processing request for the target task, acquire task data of the target task; extract key information from the task data to obtain task description information of the target task, wherein the task description information includes at least one of task type, request object behavior, request resource information, and task processing requirements.

[0101] Optionally, the task description information includes model configuration parameters; the device further includes: a configuration module configured to configure the parameters of the task processing model according to the model configuration parameters to obtain an updated task processing model; and a first input module 908 further configured to input task data and model prompt information into the updated task processing model to obtain the task processing result of the target task.

[0102] Optionally, the device further includes: a parsing module, configured to parse multiple candidate processing models for the target task from the task guidance information when the task processing model does not meet the task processing conditions; and to select a target switching model for the target task from the candidate processing models other than the task processing model; and a first input module 908, further configured to input task data and model prompt information into the target switching model to obtain the task processing result of the target task.

[0103] Optionally, the device further includes: a generation module configured to receive result feedback information sent by the client regarding the task processing result; and to generate model test results of the task processing model based on the result feedback information.

[0104] Optionally, the device further includes: an adjustment module configured to adjust the task guidance information of the target task when the model test results do not meet the model test conditions; or, to adjust the model parameters of the task processing model when the model test results do not meet the model test conditions.

[0105] The solution implemented in the embodiments of this specification utilizes the task guidance information of the target task to automatically generate accurate model prompts for the target task and automatically routes to the task processing model adapted to the target task, thereby making the task processing process more flexible and improving task processing efficiency.

[0106] The above is an illustrative scheme of a task processing device according to this embodiment. It should be noted that the technical solution of this task processing device and the technical solution of the task processing method described above belong to the same concept. For details not described in detail in the technical solution of the task processing device, please refer to the description of the technical solution of the task processing method described above.

[0107] Corresponding to the above-described code task processing method embodiments, this specification also provides embodiments of code task processing apparatus. Figure 10A schematic diagram of a code task processing apparatus according to one embodiment of this specification is shown. Figure 10 As shown, the device includes: The second acquisition module 1002 is configured to acquire the task data and task description information of the target code task; The second filtering module 1004 is configured to filter the task guidance information of the target code task from multiple candidate guidance information based on the task description information, wherein the task guidance information is used to describe the task processing logic of the target code task. The second determining module 1006 is configured to determine the model prompt information and task processing model of the target code task based on the task description information and task guidance information. The second input module 1008 is configured to input task data and model prompt information into the task processing model to obtain the task processing result of the target code task.

[0108] The solution implemented in this specification utilizes the task guidance information of the target code task to automatically generate accurate model prompts for the target code task and automatically routes it to the task processing model adapted to the target code task. This makes the code task processing process more flexible. Through the capabilities of the task processing model, it provides developers with intelligent capabilities such as code completion and code Q&A, thereby improving coding efficiency and code quality.

[0109] The above is an illustrative scheme of a code task processing device according to this embodiment. It should be noted that the technical solution of this code task processing device and the technical solution of the above-described code task processing method belong to the same concept. For details not described in detail in the technical solution of the code task processing device, please refer to the description of the technical solution of the above-described code task processing method.

[0110] Figure 11 A structural block diagram of a computing device according to one embodiment of this specification is shown. The components of the computing device 1100 include, but are not limited to, a memory 1110 and a processor 1120. The processor 1120 is connected to the memory 1110 via a bus 1130, and a database 1150 is used to store data.

[0111] The computing device 1100 also includes an access device 1140, which enables the computing device 1100 to communicate via one or more networks 1160. Examples of these networks include Public Switched Telephone Network (PSTN), Local Area Network (LAN), Wide Area Network (WAN), Personal Area Network (PAN), or combinations of communication networks such as the Internet. The access device 1140 may include one or more of any type of wired or wireless network interface (e.g., Network Interface Card (NIC)), such as an IEEE 802.11 Wireless Local Area Networks (WLAN) interface, a Wi-MAX (World Interoperability for Microwave Access) interface, an Ethernet interface, a Universal Serial Bus (USB) interface, a cellular network interface, a Bluetooth interface, a Near Field Communication (NFC) interface, and so on.

[0112] In one embodiment of this specification, the aforementioned components of the computing device 1100 and Figure 11 Other components, not shown, can also be connected to each other, for example, via a bus. It should be understood that... Figure 11 The block diagram of the computing device shown is for illustrative purposes only and is not intended to limit the scope of this specification. Those skilled in the art can add or replace other components as needed.

[0113] The computing device 1100 can be any type of stationary or mobile computing device, including mobile computers or mobile computing devices (e.g., tablet computers, personal digital assistants, laptop computers, notebook computers, netbooks, etc.), mobile phones (e.g., smartphones), wearable computing devices (e.g., smartwatches, smart glasses, etc.) or other types of mobile devices, or stationary computing devices such as desktop computers or personal computers (PCs). The computing device 1100 can also be a mobile or stationary server.

[0114] The processor 1120 is used to execute computer programs / instructions, which, when executed by the processor, implement the steps of the above-mentioned task processing method or code task processing method.

[0115] The above is an illustrative scheme of a computing device according to this embodiment. It should be noted that the technical solution of this computing device belongs to the same concept as the technical solutions of the task processing method and the code task processing method described above. For details not described in detail in the technical solution of the computing device, please refer to the description of the technical solutions of the task processing method or the code task processing method described above.

[0116] An embodiment of this specification also provides a computer-readable storage medium storing a computer program / instructions that, when executed by a processor, implement the steps of the above-described task processing method or code task processing method.

[0117] The above is an illustrative scheme of a computer-readable storage medium according to this embodiment. It should be noted that the technical solution of this storage medium belongs to the same concept as the technical solutions of the task processing method and the code task processing method described above. For details not described in detail in the technical solution of the storage medium, please refer to the description of the technical solutions of the task processing method or the code task processing method described above.

[0118] An embodiment of this specification also provides a computer program product, including a computer program / instructions that, when executed by a processor, implement the steps of the above-described task processing method or code task processing method.

[0119] The above is an illustrative scheme of a computer program product according to this embodiment. It should be noted that the technical solution of this computer program product belongs to the same concept as the technical solutions of the task processing method and the code task processing method described above. For details not described in detail in the technical solution of the computer program product, please refer to the description of the technical solutions of the task processing method or the code task processing method described above.

[0120] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0121] The computer instructions include computer program code, which may be in the form of source code, object code, executable file, or certain intermediate forms. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording media, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium may be appropriately added or removed according to the requirements of patent practice. For example, in some regions, according to patent practice, computer-readable media may not include electrical carrier signals and telecommunication signals.

[0122] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that the embodiments in this specification are not limited to the described order of actions, because according to the embodiments in this specification, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in this specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to the embodiments in this specification.

[0123] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0124] The preferred embodiments disclosed above are merely illustrative of this specification. The optional embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the embodiments described herein. These embodiments are selected and specifically described in this specification to better explain the principles and practical applications of the embodiments, thereby enabling those skilled in the art to better understand and utilize this specification. This specification is limited only by the claims and their full scope and equivalents.

Claims

1. A task processing method, comprising: Obtain the task data and task description information for the target task; Based on the task description information, the task guidance information of the target task is selected from multiple candidate guidance information, wherein the task guidance information is used to describe the task processing logic of the target task; Based on the task description information and the task guidance information, determine the model prompt information and task processing model for the target task; The task data and model prompts are input into the task processing model to obtain the task processing result of the target task.

2. The method according to claim 1, wherein determining the model prompt information and task processing model of the target task based on the task description information and the task guidance information comprises: The task guidance information is parsed to determine the model prompt template and multiple candidate processing models for the target task; The task description information and the model prompt template are integrated to obtain the model prompt information for the target task; Based on the task description information, the task processing model for the target task is selected from the plurality of candidate processing models.

3. The method according to claim 2, wherein the candidate processing model carries candidate description information, the candidate description information including at least one of task type, request object behavior, request resource information, and task processing requirements; The step of selecting the task processing model for the target task from the plurality of candidate processing models based on the task description information includes: The task description information is matched with the candidate description information to obtain a first matching result, wherein the first matching result corresponds one-to-one with the candidate processing model; Based on the first matching result, the task processing model for the target task is selected from the plurality of candidate processing models.

4. The method according to claim 1, wherein the candidate guidance information includes a candidate processing logic diagram, and the task guidance information includes a task processing logic diagram; The step of filtering the task guidance information for the target task from multiple candidate guidance information based on the task description information includes: Obtain multiple candidate processing logic diagrams, wherein the candidate processing logic diagrams carry candidate task identifiers; The target task is matched with the candidate task identifier to obtain a second matching result, wherein the second matching result corresponds one-to-one with the candidate processing logic diagram; Based on the second matching result, the task processing logic diagram of the target task is selected from the plurality of candidate processing logic diagrams.

5. The method according to claim 4, wherein obtaining multiple candidate processing logic diagrams includes: Obtain the candidate processing logic for multiple candidate tasks respectively; For a first candidate task, multiple processing nodes are extracted from the candidate processing logic of the first candidate task, wherein the multiple processing nodes include prompt nodes and multiple model nodes, and the first candidate task is any one of the multiple candidate tasks; An initial processing logic graph is constructed using the plurality of processing nodes as graph nodes and the logical relationships between the plurality of processing nodes as edges. The initial processing logic graph is used to describe the candidate processing logic of the first candidate task. Configure the model prompt template of the first candidate task on the prompt node, and configure the candidate processing model of the first candidate task on the plurality of model nodes to obtain the candidate processing logic diagram of the first candidate task.

6. The method according to claim 1, wherein obtaining the task data and task description information of the target task includes: In response to a task processing request for a target task, the task data of the target task is obtained; Key information is extracted from the task data to obtain the task description information of the target task, wherein the task description information includes at least one of the following: task type, request object behavior, request resource information, and task processing requirements.

7. The method according to any one of claims 1 to 6, wherein the task description information includes model configuration parameters; Before inputting the task data and the model prompt information into the task processing model to obtain the task processing result of the target task, the method further includes: The task processing model is configured according to the model configuration parameters to obtain an updated task processing model; The step of inputting the task data and the model prompt information into the task processing model to obtain the task processing result of the target task includes: The task data and model prompts are input into the updated task processing model to obtain the task processing result of the target task.

8. The method according to any one of claims 1 to 6, wherein before inputting the task data and the model prompt information into the task processing model to obtain the task processing result of the target task, the method further comprises: If the task processing model does not meet the task processing conditions, multiple candidate processing models for the target task are obtained by parsing from the task guidance information. From the candidate processing models other than the task processing model, select the target switching model for the target task; The step of inputting the task data and the model prompt information into the task processing model to obtain the task processing result of the target task includes: The task data and the model prompt information are input into the target switching model to obtain the task processing result of the target task.

9. The method according to any one of claims 1 to 6, wherein after inputting the task data and the model prompt information into the task processing model to obtain the task processing result of the target task, the method further comprises: Receive result feedback information sent by the client regarding the task processing result; Based on the feedback information, model test results for the task processing model are generated.

10. The method according to claim 9, further comprising, after generating the model test results of the task processing model based on the result feedback information: If the model test results do not meet the model test conditions, adjust the task guidance information of the target task. or, If the model test results do not meet the model test conditions, adjust the model parameters of the task processing model.

11. A code task processing method, comprising: Obtain the task data and task description information of the target code task; Based on the task description information, the task guidance information of the target code task is selected from multiple candidate guidance information, wherein the task guidance information is used to describe the task processing logic of the target code task; Based on the task description information and the task guidance information, determine the model prompt information and task processing model for the target code task; The task data and model prompts are input into the task processing model to obtain the task processing result of the target code task.

12. A task platform, comprising a request interface and a response unit; The request interface is used to receive task processing requests for the target task; The response unit is configured to, in response to the task processing request, acquire the task data and task description information of the target task; and, based on the task description information, filter out the task guidance information of the target task from multiple candidate guidance information, wherein... The task guidance information is used to describe the task processing logic of the target task; based on the task description information and the task guidance information, the model prompt information and task processing model of the target task are determined; the task data and the model prompt information are input into the task processing model to obtain the task processing result of the target task.

13. A computing device, comprising: Memory and processor; The memory is used to store computer programs / instructions, and the processor is used to execute the computer programs / instructions, which, when executed by the processor, implement the steps of the method according to any one of claims 1 to 11.

14. A computer-readable storage medium storing a computer program / instructions that, when executed by a processor, implement the steps of the method according to any one of claims 1 to 11.

15. A computer program product comprising a computer program / instructions that, when executed by a processor, implement the steps of the method according to any one of claims 1 to 11.

Images