DIYKaraoke :Karaoke Track Genearion, Song Recognition, Song Recommendation and Lyrics Synchronization

Steps to run Django server

cd Backend

python3 manage.py makemigrations

python3 manage.py migrate

python3 manage.py runserver

Steps to run React Native client

cd Frontend

npm install expo-cli

expo start

Steps to train Karaoke Generation Module

cd Machine Learning Modules/ Karaoke Generation

Place your dataset

path/to/dataset/

+- instruments/

| +- 01_foo_inst.wav

| +- 02_bar_inst.mp3

| +- ...

+- mixtures/

+- 01_foo_mix.wav

+- 02_bar_mix.mp3

+- ...

Train a model

python train.py --dataset path/to/dataset --reduction_rate 0.5 --mixup_rate 0.5 --gpu 0

Steps to generate Karaoke Track

python inference.py --input path/to/an/audio/file

Steps to recognize song

Recognize existing mp3 file:

python recognize-from-file.py path/to/an/audio/file

Recognize song playing on microphone:

python recognize-from-microphone.py

Steps to generate recommendations

Open the Song_Recommender.ipynb on Jupyter notebook and run