PipelineStage

Enums · AiDotNet.Enums

Defines which stage of an AI pipeline a component operates in.

For Beginners

AI systems often process data through multiple stages, like an assembly line. This tells you where in the pipeline a component fits. Components in the same stage can be swapped; components in different stages are composed sequentially.

How It Works

A typical RAG pipeline: DataIngestion → Indexing → Retrieval → PostRetrieval → Generation. A typical ML pipeline: Preprocessing → Training → Evaluation.

Fields

Field Summary
PipelineStage DataIngestion Data ingestion stage: parsing, chunking, and preparing raw data.
PipelineStage Evaluation Evaluation stage: measuring model/pipeline quality.
PipelineStage Generation Generation stage: producing final output from retrieved context.
PipelineStage Indexing Indexing stage: embedding and storing processed data for retrieval.
PipelineStage PostRetrieval Post-retrieval stage: refining retrieved results before generation.
PipelineStage Preprocessing Preprocessing stage: transforming raw features before model training/inference.
PipelineStage QueryProcessing Query processing stage: transforming user queries before retrieval.
PipelineStage Retrieval Retrieval stage: searching and filtering stored data given a query.
PipelineStage Training Training stage: optimizing model parameters from data.
int value__