🎶 Overview

This demo showcases a simple single-turn audio–language interface built for integration with the WeaveMuse framework.
It uses the open-source NVIDIA Audio Flamingo 3 model for audio understanding, transcription, and sound reasoning.

You can upload an audio file and ask natural-language questions such as:

• “What kind of sound is this?”
• “Describe the scene.”
• “Transcribe any speech.”

Acknowledgment:
Model and research credit to NVIDIA, for the development of the open Audio Flamingo 3 model and datasets used for training.
This interface is a simplified demonstration of how such models can be integrated into broader creative AI systems like WeaveMuse.

Tech stack: Gradio + PyTorch + llava + WeaveMuse Integration