audiocraft-audio-generation

To generate music and sound effects from text, use the AudioCraft library which converts natural language prompts into audio content via MusicGen and AudioGen models, producing original tracks or realistic sound clips instantly.

Yes, you can generate audio that follows a specific provided melody using the melody conditioning feature of AudioCraft, which guides the MusicGen model to produce music matching your input melodic structure.

Yes, you need PyTorch and HuggingFace Transformers installed in your environment to run AudioCraft audio generation, as the Skill depends on these frameworks to load and execute the MusicGen and AudioGen models.

Yes, AudioCraft supports stereo audio output for text-to-music generation, allowing you to produce music tracks with stereo panning and width directly from your natural language text prompts.

In text-to-audio generation, MusicGen is used for creating original music tracks from descriptive prompts, while AudioGen is designed to generate realistic sound effects like a dog barking or environmental audio from text.

The metadata does not specify exact duration limitations for AudioCraft, but the Skill is demonstrated generating 10-second clips, and users should anticipate computational constraints with longer text-to-audio generation tasks.