AudioGen Medium v2

model·generative audio model·active

Meta's released 1.5B text-to-sound checkpoint, identified as AudioGen model version 2.

Recorded facts

Official site	https://huggingface.co/facebook/audiogen-medium ↗
Geography	unknown
parameter count	1.5B
huggingface repo	facebook/audiogen-medium
api	unknown
inputs	English text description
license	CC-BY-NC-4.0 model weights; MIT code
outputs	16 kHz mono audio
weights	public
modalities	16 kHz mono audio; English text description
dependencies	unknown
model family	AudioGen
access method	downloadable weights; AudioCraft API
exact version	v2 / facebook/audiogen-medium
repository links	https://huggingface.co/facebook/audiogen-medium; https://github.com/facebookresearch/audiocraft
evaluation results	result: FAD: 1.77 · KLD: 1.58 · text consistency: 0.3 · source: https://huggingface.co/spaces/facebook/MusicGen/blob/main/model_cards/AUDIOGEN_MODEL_CARD.md · benchmark: AudioCaps
evidence separation	developer claims: · measured results: Released checkpoint reports FAD 1.77, KLD 1.58 and text consistency 0.30 on AudioCaps. · independent evaluations:
developer maintainer	Meta AI
material us relevance	Publicly accessible in or materially offered to the United States market as of snapshot.
documented capabilities	text-to-sound generation; 10-second audio
products exposing model	unknown
commercial use conditions	Model weights are non-commercial under CC-BY-NC-4.0.
known geographic or policy restrictions	Not able to generate realistic vocals; English descriptions; Non-commercial weights
safety watermarking disclosure provenance	unknown