Audio Flamingo: A Novel Audio Language Model with Few-Shot Learning and Dialogue Abilities

Benchmark Model Rank Results
audio-captioning-on-clothoAudio Flamingo (Pengi trainset)#3SPIDEr: 0.312CIDEr: 0.489SPICE: 0.134BLEU-4: 17.4METEOR: 18.7…
audio-understanding-reasoning-on-mmauAudio Flamingo Chat (1B)#14Accuracy (Avg, Test-mini): 16.60
retrieval-augmented-few-shot-in-context-audio-captioning-on-audiocapsAudio Flamingo (4-shot)#1CIDEr: 0.518
zero-shot-audio-captioning-on-audiocapsAudio Flamingo#1METEOR: 20.5BLEU-4: 14.3CIDEr: 50.2ROUGE-L: 40.8SPICE: 15.1…