Structure

Explanation of concepts

Node & Module

Node: Node is a higher-level concept that acts as a container for modules. It allows for the interchange of modules, meaning within a single node, multiple modules can be swapped or combined to find the optimal mix for processing data.

  • Node Parameters are common parameters that apply across all modules within a node. They are set at the node level and affect the operation of each module contained within that node.

Module: Module refers to the individual components that can be fitted into a node. A system can have multiple modules within a node, and the number of modules can significantly increase over time, while the number of nodes remains limited.

  • Module Parameters are specific to each module, allowing for the customization of individual module behavior within the broader context of the node they reside in.

Tip

If you want to know more about the node and module, please refer to the Nodes & Modules section.

Node Line

Node Line: A Collection of Nodes.

Example Node Lines

Node Lines

  • Purpose and Future Enhancements: Aims to support merging, splitting, and looping in node sequences. These functionalities are in development.

  • Modular RAG Integration: Essential for Modular Retrieval-Augmented Generation, enhancing large language models by integrating retrieval mechanisms for dynamic data processing. Refer to documentation for more details.

  • Current Temporary Configuration:Nodes are currently arranged temporarily to simulate Node Lines’ intended functionalities until full integration is achieved.

Tip

Node lines can be changed at any time in the YAML file

Strategy

Strategy: Strategy is what you decide to optimize ‘based on’ when optimizing. It is a crucial aspect of optimization that guides the selection and arrangement of optimization techniques.

  • Metrics: It is essential for all nodes to include metrics as part of their optimization strategy. Metrics serve as a quantitative basis for evaluating the effectiveness of optimization efforts and guiding strategic decisions.

    Summarize

    By default, every node must contain their own metric.

  • Speed Threshold: Optionally, the speed_threshold parameter can be added to all nodes. This parameter serves as a criterion for optimization, focusing efforts on enhancing the speed of operations up to a specified threshold.

    Summarize

    The speed threshold can be optionally applied to all nodes.

  • Node-Specific Strategies: The optimization strategy may vary from one node to another within the same system. This flexibility allows for tailored optimization that addresses the unique requirements or limitations of each node.

    Summarize

    1. Different metrics for different nodes.

    2. You can find more information in the Nodes & Modules section.