https://github.com/troyanovsky/llm_summarizer

Use LLM and Whisper to summarize long text and audio/video
https://github.com/troyanovsky/llm_summarizer

gpt summarization whisper

Last synced: about 1 month ago
JSON representation

Use LLM and Whisper to summarize long text and audio/video

• Host: GitHub
• URL: https://github.com/troyanovsky/llm_summarizer
• Owner: Troyanovsky
• Created: 2023-06-25T07:37:37.000Z (over 1 year ago)
• Default Branch: main
• Last Pushed: 2023-06-27T06:49:41.000Z (over 1 year ago)
• Last Synced: 2023-12-17T11:01:10.495Z (about 1 year ago)
• Topics: gpt, summarization, whisper
• Language: HTML
• Homepage: https://troyanovsky.github.io/LLM_Summarizer/
• Size: 17.6 KB
• Stars: 1
• Watchers: 3
• Forks: 2
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# LLM_Summarizer - Too Lazy to View

## Too Lazy To View

Too Lazy To View is a web application that allows users to summarize content they are too lazy to read or view. The application supports both text and audio/video input, with the ability to transcribe audio files using Whisper. Summaries are generated using OpenAI's GPT-3.5 model. For very long text, summaries will be generated for each segment, and the overall summary will be generated from the concatenation of the segment summaries.

## Getting Started

To use the application, simply open the index.html file in your browser or visit https://troyanovsky.github.io/LLM_Summarizer/

## Prerequisites

The following resources are required to run the application:

A modern web browser (tested on Chrome)  

An internet connection (for loading external libraries and Whisper model)  

An OpenAI API key (to generate summaries)  

## Usage

Upon opening the index page, you can select a file to upload for transcription by clicking the "Choose file" button. If the file is a text-based file, the content will be loaded to the text area; if the file is an audio file, you can click the "Transcribe audio" button to initiate the transcription process. Transcription may take some time depending on the length of the audio file. For the first time transcribing the audio, the application will use the transformer.js library to download the Whisper-tiny model to transcribe your audio. Subsequent transcription will be much faster. All transcription is done locally.

Once the transcription process is complete, the raw text will be displayed in the text area below the buttons. Users can then enter their OpenAI API key in the input field provided and click the "Summarize" button to generate a summary of the text.

## Built With

Vue.js - JavaScript framework used for building user interface  

Tailwind CSS - CSS framework used for styling  

transformers.js - JavaScript library used for machine learning-based speech-to-text transcription  

OpenAI GPT-3.5 model - LLM for generating summaries  

## Author

Troy (Guodong) Zhao - https://github.com/Troyanovsky