1.0 KiB
1.0 KiB
| layout | title |
|---|---|
| default | Home |
Mini LLM Flow
A 100-line minimalist LLM framework for agents, task decomposition, RAG, etc.
Core Abstraction
We model the LLM workflow as a Nested Flow:
-
Each Node handles a simple LLM task (e.g., text summarization, structure extraction, or question answering).
-
Nodes are chained together to form a Flow for more complex tasks. One Node can be chained to multiple Nodes based on Actions, e.g., for agentic steps.
-
A Flow can be treated as a Node for Nested Flows.
-
Both Nodes and Flows can be Batched for data-intensive tasks (e.g., processing one file at a time in a directory of files).
-
Nodes and Flows can be Async, e.g., for user feedback before proceeding.
Paradigm Implementation
- Task Decomposition
- Agent
- Map Reduce
- RAG
- Structured Output
Example Use Cases
TODO
