Summarizer

A tool that summarizes audio into text, with focus on academic use, annotate key points of a class, summarize really long lectures, or record yourself reading the material to then have it summarized.

Updates I'd like to make

• Have it to ask questions to you about the material
• direct mic transcription
• A nice GUI
• phone compatibility

Requirements

• whisper (for managing audio)
• Ollama (for managing text)
• Torch

Examples

• English

This example was made by feeding a video explaining ohm's law (https://youtu.be/_rSHqvjDksg?si=K-6AEUJeF_q6pKgO)

• Español

Este ejemplo fue hecho utilizando un video corto explicando la ley de ohm (https://youtu.be/izDKIyCPRWE)

Installation

Ollama

• linux

curl -fsSL https://ollama.com/install.sh | sh

• Windows & MacOS visit Ollama's website (https://ollama.com/download/windows) / (https://ollama.com/download/mac)

Whisper

sudo pacman -S python-openai-whisper

Running summarizer

git clone https://github.com/aguscsc/Summarizer.git
cd code
python summarizer.py

Usage

Run summarizer.py and you'll be prompted to choose a whisper model, you can visit openai's whisper repo for more information about the models (https://github.com/openai/whisper.git).

Choose a Whisper model (e.g. tiny, base, small, medium, large):
"YOUR CHOICE HERE"
Insert audio file path:
"YOUR AUDIO PATH HERE"