DEV Community

Cover image for Run MusicGen on Google Colab/Windows11
0xkoji
0xkoji

Posted on

Run MusicGen on Google Colab/Windows11

What is MusicGen

Audiocraft is a library for audio processing and generation with deep learning. It features the state-of-the-art EnCodec audio compressor / tokenizer, along with MusicGen, a simple and controllable music generation LM with textual and melodic conditioning.

Meta's open source AI MusicGen turns text and melody into new songs
https://the-decoder.com/metas-open-source-ai-musicgen-turns-text-and-melody-into-new-songs/

Steps for Colab

  1. Go to Google Colab
  2. Change Runtime
  3. Run code
  4. Try to generate sound

step 1. Go to Google Colab

https://colab.research.google.com/

step 2. Change Runtime

Click Runtime and select GPU

Google Colab Runtime

step 3. Run code

!git clone -b dev https://github.com/camenduru/MusicGen-hf
cd MusicGen-hf/
!pip install -r requirements.txt
!python app.py
Enter fullscreen mode Exit fullscreen mode

Installing python packages will take some time.
If everything works properly, you will see something like this.

Google Colab output

step4. Try to generate sound

If you click the link Running on public URL:, you will see Gardio.

Gardio

Now time to create your own sound via prompt!
It would take around 1-2 minutes to generate a mp4 file.

Steps for Windows11

You need to install cuda11.7 on your Windows11.

  1. Create a venv
  2. Install pytorch
  3. Clone repo and install requirements
  4. Try to generate sound

step 1. Create a venv

python -m venv musicgen
# activate venv
./musicgen/scripts/activate
Enter fullscreen mode Exit fullscreen mode

step 2. Install pytorch

This step will take some time.

pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu117
Enter fullscreen mode Exit fullscreen mode

step 3. Clone repo and install requirements

git clone https://github.com/facebookresearch/audiocraft.git
cd audiocraft
pip install -e .
Enter fullscreen mode Exit fullscreen mode

step 4. Try to generate sound

As Google Colab, we will run web ui with the following command.

python app.py
Enter fullscreen mode Exit fullscreen mode

Top comments (2)

Collapse
 
mdabir1203 profile image
Info Comment hidden by post author - thread only accessible via permalink
uknowWho

Code is wrong. It would be %cd because % allows you go into the folder.

%cd MusicGen-hf/
!pip install -r requirements.txt
!python app.py

Collapse
 
0xkoji profile image
0xkoji

% is not needed

Some comments have been hidden by the post's author - find out more