Multi-stage LLM Worfklow to Summarize and Translate an Article using LangChain

Author(s): Steve George

Originally published on Towards AI.

“An LLMChain is a simple chain that adds some functionality around language models. It is used widely throughout LangChain, including in other chains and agents.

An LLMChain consists of a PromptTemplate and a language model (either an LLM or chat model). It formats the prompt template using the input key values provided (and also memory key values, if available), passes the formatted string to LLM and returns the LLM output.”

In this article, we will create two LLMChain which perform summarization and translation, respectively. And later create a sequence using these two chains.

Below is the overall framework of the approach.

Source: Image by the author

Using google-t5/t5-small, we are building a summarization pipeline. For creating the pipeline, the first thing we require is setting a prompt template. Using the prompt_template package available in langchain, we can define the template.

As per the template defined below, we will summarize the input text. The user can enhance the prompt template by using various methods like including delimiters.

from langchain import PromptTemplatesummary_template = """Write a summarization of the below article in two sentencesArticle:" {article}"Summary: """summary_prompt_template = PromptTemplate(input_variables=['article'], template = summary_template)#input variable value should match the value mentioned in the template. #Multiple variables can be passed based on the use-casearticle="As… Read the full blog for free on Medium.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI