GPT-Summarizer

Lightweight library to summarize long texts using GPT models

GPT-Summarizer provides the capability of summarizing very large text corpus exceeding the token limit of normal GPT models. Of course, you can use it to summarize shorter texts as well.

Installation

• Clone this repo
• python setup.py install inside the cloned folder
• Done !

Usage

>>> from gptsummarizer import summarizer
>>> generator = summarizer.Summarizer(key="put_your_openai_key_here")
>>> summary = generator.getSummary(text="Hello! How are you?")
>>> summary
Two people are exchanging greetings and inquiring about each others wellbeing.

💪 Power Usage

Setting the GPT model to use for summarization. Currently supports two GPT engines, text-davinci-003 and gpt-3.5-turbo. If no engine is specified, it defaults to text-davinci-003

>>> summary = generator.getSummary(text="Hello! How are you?", engine="gpt-3.5-turbo")

Setting other model parameters like temperature, max_tokens are optional, and can be done similarly.

>>> generator.getSummary(text="Hello! How are you?", 
                         engine="text-davinci-003", 
                         temperature=0.3, 
                         max_tokens=600, 
                         top_p=1, 
                         frequency_penalty=0, 
                         presence_penalty=1)

For more information on how to fine-tune these parameters, follow OpenAI documentation.

License

MIT