AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

AudioGen: Textually Guided Audio Generation

Felix Kreuk, Gabriel Synnaeve, Adam Polyak, Uriel Singer, Alexandre Défossez, Jade Copet, Devi Parikh, Yaniv Taigman, Yossi Adi

ICLR 2023 · Image generation › Autoregressive and latent generation · approach: Diffusion audio synthesis · 52 citations (OpenAlex)

Read the paper

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add mixing augmentation (fusing audio pairs and concatenating captions) (ablation)AudioGen-base · AudioCaps testREL (text relevance): 63.23 -> 66.01Improved
add mixing augmentation (ablation)AudioGen-base · AudioCaps testFAD: 2.84 -> 3.13Hurt
add mixing augmentation (ablation)AudioGen-base · AudioCaps testKL: 2.14 -> 2.09Improved
add mixing augmentation (ablation)AudioGen-base · AudioCaps testOVL: 70.85 -> 71.68 (within CI)No clear effect
scale up ALM to 1B parameters and T5 to large (mix augmentation)AudioGen · AudioCaps testFAD: 3.13 -> 1.82 (vs DiffSound 7.39)Improved
use AudioGen-base instead of DiffSound (diffusion over VQ-VAE codes)AudioGen · AudioCaps testKL: 2.57 -> 2.09 (base); 1.69 (large)Improved
multi-stream modeling with 2 streams (x64 DSF) (ablation)AudioGen-base · AudioCaps testFAD: 3.13 -> 10.35; speed-up x2.0Hurt
multi-stream modeling with 4 streams (x128 DSF) (ablation)AudioGen-base · AudioCaps testKL: 2.09 -> 2.36; speed-up x5.1Hurt
multi-stream modeling with 2 streams (x64 DSF) (ablation)AudioGen-large · AudioCaps testFAD: 1.82 -> 6.89; speed-up x2.3Hurt
use T5-large instead of T5-base text encoder (base ALM) (ablation)AudioGen · AudioCaps testKL: 2.09 -> 1.91Improved
use ALM-large instead of ALM-base (T5-base) (ablation)AudioGen · AudioCaps testFAD: 3.13 -> 2.27Improved
apply classifier-free guidance (remove it, gamma=1 vs 3.0) (ablation)AudioGen · AudioCaps testFAD: FAD minimum at gamma=3.0; removing CFG gives poor performanceImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy