Large language models (LLMs) are transforming the way developers build applications by allowing them to combine LLMs with other sources of computation or knowledge.
LangChain is a library designed to assist in the development of such applications, providing support for prompt management, chains, data augmented generation, agents, memory, and evaluation.
Key terms:
LangChain: A library designed to assist in the development of applications using LLMs, providing support for various development aspects.
Chains: Sequences of calls that go beyond just a single LLM call, including integrations with other tools and end-to-end chains for common applications.
Data Augmented Generation: Specific types of chains that interact with external data sources to fetch data to use in the generation step, such as summarization or question/answering.
Agents: LLMs making decisions about actions, taking actions, and observing, with a standard interface and examples of end-to-end agents provided by LangChain.
Memory: The concept of persisting state between calls of a chain or agent, with a standard interface, memory implementations, and examples provided by LangChain.