This document is written based on FastGPT v4.17.0.
What the data for this product category looks like
Data sources for this product category include ERP/MES systems, component batch ledgers, tightening and refrigerant charging equipment, safety compliance testers, performance lab data, and after-sales fault records. Daily documents include configured BOMs, assembly work instructions, refrigerant charging records, electrical safety test reports, random performance inspection reports, and rework records. Data updates follow per-unit equipment, shift production batches, and random inspection schedules. Document fields include model number, component batch number, process parameter values, test results, and fault codes, among others.
What constraints these characteristics impose on the model access and configuration link
Mixed-model mixed-line assembly production scenarios require the model to accurately match the process documents of the corresponding model. A sufficient context window must be configured to carry process documents and equipment data for multiple models. Frequently updated real-time equipment parameters and batch-generated random inspection and rework records require the model to call rules adapted to real-time performance and batch data processing. Abnormal judgment related to safety compliance and energy efficiency must strictly match compliance requirements. The model's randomness parameters must be controlled within a reasonable range. The current situation of scattered multi-source data sources requires configuring a unified API access endpoint to integrate various documents and data from ERP, MES, and production equipment. This avoids format conflicts in cross-system calls.
How to set the configuration
| Configuration Item | Recommended Value Range | Rationale |
|---|---|---|
maxContext | 8000–12000 token | Must carry multi-model BOMs, assembly work instructions, and recent equipment data, avoid context truncation leading to missing process parameters |
maxTokens | 4000–6000 token | Limit the length of a single generated process document or abnormal report, avoid overly long output exceeding platform limits |
temperature | 0.1–0.3 | Safety compliance and process judgment require low randomness, ensure consistent expression of abnormal reports and process documents |
quoteMaxToken | 2000–3000 characters | Must reference sufficient equipment test records and rework data as the basis for process drafting |
| Workflow/API tool selection | auto | Automatically call the document retrieval tool to obtain assembly work instructions and safety compliance requirements for the corresponding model |
AIPROXY_API_ENDPOINT | Official access address of the corresponding model | Adapt to access scenarios of different models, ensure data transmission meets production environment requirements |
Three common errors
- Phenomenon: Returns 401 unauthorized error when calling Azure OpenAI models. Cause: The corresponding relationship between
AIPROXY_API_ENDPOINTandAIPROXY_API_TOKENis not configured correctly, or the key permission does not cover the access scope of process documents and equipment data. - Phenomenon: Cross-model incorrect assembly parameters appear in generated process documents. Cause:
maxContextis configured too small, causing the model to fail to load the exclusive assembly work instructions of the corresponding model, confusing process requirements of different models. - Phenomenon: The signature time field in the workflow is empty or displays abnormally. Cause: The built-in system time variable is not bound correctly, and the platform-provided current date variable is not explicitly required to be called in the prompt.
How to confirm the configuration is complete
- Initiate a single-model process document drafting test, check whether the generated content includes exclusive assembly parameters and safety compliance requirements for the corresponding model. Adjust
maxContextandquoteMaxTokenuntil the content is complete. - Submit an abnormal report generation task, verify whether the randomness of the output result meets the process rigor requirements. Adjust the value range of
temperature. - Call the API access test tool, verify that the configuration of
AIPROXY_API_ENDPOINTandAIPROXY_API_TOKENcan normally pull BOM data from ERP, with no format conversion errors. - Trigger the signature time generation step in the workflow, confirm that the field automatically fills in the current date. Verify that the system time variable binding logic is normal.
The values provided are common starting points and should be measured against the reader's own samples.