Documentation
¶
Index ¶
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
Types ¶
type LLMConfig ¶
type LLMConfig struct {
Models []types.ModelConfig // Models lists optional per model metadata.
DefaultModel string // DefaultModel is the model used for chat/generation when none is specified.
UserPromptTmpl string // UserPromptTmpl is a go template used to build the user query + context.
EmbeddingModel string // EmbeddingModel is the model used to produce embeddings.
RetrievalTopK int // RetrievalTopK is the number of results to fetch from the vector DB for RAG. Use 0 to disable retrieval.
DefaultContext int // DefaultContext is the fallback maximum context length (in tokens).
DefaultTemperature *float64 // DefaultTemperature is the fallback sampling temperature.
}
LLMConfig contains high-level LLM settings for the RAG pipeline.
Click to show internal directories.
Click to hide internal directories.