Audiocraft Installation

gistfile1.txt

# must have miniconda installed

# installation for musicgen
conda create -n audiocraft python=3.9
conda activate audiocraft
brew install ffmpeg (for mac)
winget install ffmpeg (for windows)
git clone https://github.com/facebookresearch/audiocraft.git
cd audiocraft
python -m pip install -r requirements.txt
python -m demos.musicgen_app --share
# first time you run inference, it'll download the model

# installation for audiogen
# NOTE: after already installing music gen
# NOTE: after PR185 gets merged, you won't have to execute the next two commands
git fetch origin pull/185/head:PR185
git checkout PR185
python -m demos.audiogen_app --share
# first time you run inference, it'll download the model

Nice doc, but just a small hint:

I needed to install pytorch also by calling
conda install -c pytorch pytorch
before line installing the requirements in line 10.

sudo apt install ffmpeg for ubuntu

thankyou, after a lot of headbanging, this actually worked for me

with new versions, you need to type:

python demos/musicgen_app.py

@klein-artur thanks for calling out pytorch here. I didn't realize I didn't have it installed on my new machine- totally would have gotten stuck without your comment!

Had some issues with zsh.

To fix run:

conda init zsh
exit terminal/iterm then reopen it
conda activate audiocraft