Audio Flamingo 2 Demo

Audio Flamingo 2 is NVIDIA's latest Large Audio-Language Model that is capable of understanding audio inputs and answer any open-ended question about it.
Audio Flamingo 2 is not an ASR model and has limited ability to recognize the speech content. It primarily focuses on perception and understanding of non-speech sounds and music.
The demo is hosted on the Stage 2 checkpoints and supports upto 90 seconds of audios. Stage 3 checkpoints that support upto 5 minutes will be released at a later point.