FastGPT v4.17.0
What This Category's Data Looks Like
Business data sources for metal cutting machining include product drawings and 3D models, CAPP/CAM systems, machine tool collection terminals, tool management systems, MES systems, and inspection and measuring tool data. Daily business documents include process routes, operation cards, tool lists, CNC programs, first-piece inspection records, patrol inspection forms, and measuring tool calibration records. Data updates are triggered by work order execution, single-piece machining, sampling inspection, or tool replacement. Document fields cover cutting parameters, dimensional tolerances, tool life, clamping benchmarks, and more. Units include rotational speed, feed rate, and surface roughness parameters, among others.
Constraints Imposed on Model Access and Configuration
Data sources are scattered and linked to multiple dimensions such as work orders, tools, and machining processes. When accessing data, ensure that the context association logic of multi-source data matches the production workflow. Frequently updated work order and tool data require configuration to support synchronization rules triggered by events such as work order execution or tool replacement. This prevents the knowledge base from being out of sync with actual production. There are diverse document types, including structured and semi-structured content such as CNC program code, tolerance values, and process parameters. Configuration must adapt to parsing rules for different documents, ensuring that key fields are accurately extracted. The process links associated with data directly affect the accuracy of Q&A responses. Restrict the context scope during model calls to avoid interference from irrelevant data on output.
How to Configure Parameters
| Configuration Item | Recommended Value Range | Rationale |
|---|---|---|
maxContext | 8000–12000 token | Adapts to multi-segment associated content in process documents, covers associated information from operation cards, CNC programs, and tool lists, prevents truncation of critical process parameters |
maxTokens | 2000–3000 token | Covers the Q&A output requirements for complete process specifications, ensures complete output of critical content such as clamping schemes and cutting steps |
temperature | 0.1–0.3 | Ensures output accuracy of structured content such as process parameters and tolerance values, reduces irrelevant divergent content |
AIPROXY_API_ENDPOINT | Fill in according to the deployment environment | Matches the FastGPT proxy access address, ensures normal model call link functionality |
AIPROXY_API_TOKEN | Fill in according to the key assigned by the platform | Completes identity verification for model calls, ensures access security |
| Workflow/API tool selection | auto | Enables tool calls as needed, adapts to multi-step query requirements such as process scheme verification and parameter validation |
Three Common Configuration Mistakes
- Phenomenon: Significant differences in return results between online conversations and API calls using the same knowledge base, prompt, and model. Cause: Failure to unify
maxContextandtemperatureparameter configurations, or failure to correctly pass context parameters during API calls. - Phenomenon: When using a specified model, only thinking content is output, with no main body answer. Cause: The
maxTokensconfiguration value is too small, truncating the main body output section, or the prompt fails to explicitly require output of main body content. - Phenomenon: An error prompt stating that no language model is configured when entering the workspace editing page, or the AI model option is empty in the workflow. Cause: Failure to correctly fill in
AIPROXY_API_ENDPOINTandAIPROXY_API_TOKEN, or the model has not been registered and enabled in the platform backend.
How to Confirm Configuration Is Complete
- Check the filled content of
AIPROXY_API_ENDPOINTandAIPROXY_API_TOKEN, ensure they match the information assigned by the platform. - Initiate a test call, verify that the model can correctly return Q&A content related to process parameters and tolerance requirements, adjust parameters to ranges that meet business needs.
- Compare return results from online conversations and API calls, confirm that
maxContextandmaxTokensconfigurations cover critical document content. - Check the workflow configuration page, confirm that the AI model option displays available deployed models.
The values provided are common starting points and should be measured against the reader's own samples.