GitHub - llSourcell/Music_Generation: This is the code for "Song Generator&...

 6 years ago
source link: https://github.com/llSourcell/Music_Generation
Go to the source link to view the article. You can view the picture content, updated content and better typesetting reading experience. If the link is broken, please click the button below to view the snapshot at that time.


Keras implementation of deepmind's wavenet paper

This is the code for this video on Youtube by Siraj Raval on music generation ft Taryn Southern.

Link to paper

Dataset used

I have used Librispeech corpus. I have concatenated all audio files in dev-clean to create train.wav and all files in test-clean to create validate.wav. I have resampled the audio files to 8000 Hz. Here is how you can create train.wav & validate.wav using vlc on linux:

cvlc -vvv --sout-keep --sout-all --sout "#gather:transcode{acodec=s16l,channels=1,samplerate=8000}:std{access=file,mux=wav,dst=validate.wav}" `find LibriSpeech/test-clean/ -name "*.flac"` vlc://quit
cvlc -vvv --sout-keep --sout-all --sout "#gather:transcode{acodec=s16l,channels=1,samplerate=8000}:std{access=file,mux=wav,dst=train.wav}" `find LibriSpeech/dev-clean/ -name "*.flac"` vlc://quit
  • The basic generative model
  • Conditioning logic (speaker)
  • Conditioning logic (TTS)

Credits for this code go to usernammee. I've merely created a wrapper to get people started.

About Joyk

Aggregate valuable and interesting links.
Joyk means Joy of geeK