This project is a part of Chord Guitar AI Product. The purpose of this project is to build a model that can predict the chord of a guitar from an audio file.
Dataset can be found here. Save audio files in audio folder.
- audio/
- A.wav
- A#.wav
- etc
- helper/
- models/
- Install dependencies
if your computer is using GPU, run this
pip install -r requirements.txt
if your computer is using CPU, run this
pip install -r requirements.nogpu.txt
- Run the program
open main.ipynb
and run all cells for re training the model or open predict.ipynb
and run all cells for predicting chord from audio file
- Fork this repository
- Clone your forked repository
- Create new branch
- Make changes
- Commit and push your changes
- Create pull request
- Wait for review
- Done
- Record your guitar sound with standar tuning for 60s each chord
- You must record with 2 different beat (2/4 and 4/4). 30s for each beat
- Trim your audio file to 50s only. So the audio file doesn't contain noise from beginning and end of the audio
- Save your recorded file with .wav format
- Send your recorded file to email with subject "Guitar Chord AI Dataset"