Text-to-Music Generation on MusicCaps (evaluation set)
2.18FADGT-AudioMAE
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| GT-AudioMAEType=Directly applying Ground Truth 'Language of Audio'2023.08 | 2.18 | 0.27 | 0.257 | 3.59 | 3.92 | — | — | — | |
| AudioLDM 2Variant=Full2023.08 | 3.13 | 1.2 | 0.301 | 3.34 | 3.54 | — | — | — | |
| AudioLDM-MSource=Reproduced implementation2023.08 | 3.2 | 1.29 | 0.36 | 3.03 | 3.25 | — | — | — | |
| FlashAudioNFE=4, Refinement/Distillation=false2024.10 | 3.25 | 1.2 | 0.44 | — | — | 0.014 | 83.96 | 85.83 | |
| MusicGen-MediumSource=Original Report2023.08 | 3.4 | 1.23 | 0.32 | — | — | — | — | — | |
| FlashAudioNFE=1, Refinement/Distillation=true2024.10 | 3.48 | 1.22 | 0.42 | — | — | 0.0025 | 83.14 | 85.06 | |
| AudioLDM 2NFE=2002024.10 | 3.81 | 1.22 | 0.43 | — | — | 2.2 | 82.42 | 85.64 | |
| AudioLCMNFE=22024.10 | 3.92 | 1.24 | 0.4 | — | — | 0.003 | 82.56 | 85.71 | |
| MusicLM2023.08 | 4 | — | — | — | — | — | — | — | |
| MusicLM2024.10 | 4 | — | — | — | — | — | — | — | |
| AudioLDM 2Variant=MSD2023.08 | 4.47 | 1.32 | 0.294 | 3.41 | 3.3 | — | — | — | |
| MusicGen2024.10 | 4.5 | 1.41 | 0.42 | — | — | 1.28 | 81.15 | 84.21 | |
| MusicGen-MediumSource=Reproduced implementation2023.08 | 4.89 | 1.35 | 0.291 | 3.37 | 3.38 | — | — | — | |
| MusicLDMNFE=2002024.10 | 5.2 | 1.47 | 0.4 | — | — | 1.4 | 79.34 | 82.1 | |
| MeLoDy2023.08 | 5.41 | — | — | — | — | — | — | — | |
| Melody2024.10 | 5.41 | — | — | — | — | — | — | — | |
| Mousai2023.08 | 7.5 | 1.59 | — | — | — | — | — | — | |
| Mousai2024.10 | 7.5 | — | — | — | — | — | — | — | |
| Riffusion2024.10 | 13.31 | 2.1 | 0.19 | — | — | 0.4 | 75.23 | 76.12 | |
| Riffusion2023.08 | 14.8 | 2.06 | 0.19 | — | — | — | — | — | |
| GroundTruth2023.08 | — | — | 0.253 | 3.82 | 4.26 | — | — | — | |
| GroundTruth2024.10 | — | — | 0.46 | — | — | — | 89.32 | 91.04 |