CLI-tool for Stable Audio Open 🎵🎵🎵

Features

• Interactive prompt shell
• Batching
• Python CLI with arguments
• Output to custom folder

Setup

This has been tested under Python 3.8.10, as specified in the Stable Audio Open Repo and Conda.

Once you are in your Python 3.8.10 environment:

• Install your version of PyTorch (>2.0)
• pip install stable-audio-tools

It might take a long time to resolve dependencies, be patient!

Then, just run python stable-audio-cli.py to launch the interactive prompt shell, or provide arguments as shown below.

Help

usage: Stable Audio Open CLI tool [-h] [--prompt PROMPT] [--start START] [--total TOTAL] [--num NUM] [--out-dir OUT_DIR]

Generate audio from prompt using Stable Audio Open model

optional arguments:
  -h, --help         show this help message and exit
  --prompt PROMPT    Prompt for the audio generation
  --start START      Start time in seconds
  --total TOTAL      Total time in seconds (max. 47)
  --num NUM          Number of sounds to generate
  --out-dir OUT_DIR  Output directory for generated audio. Default is current directory.

Run this script without any arguments to enter interactive mode, or provide, at minimum, --prompt to generate audio directly.

Example: python stable-audio-cli.py --prompt "lofi synth tune" --out-dir "/home/cat" --num 2

Let me know if you have any feedback or suggestions!