FL2VA/audio_vae/metadata.json
| 1 | { |
| 2 | "metadata": { |
| 3 | "kwargs": { |
| 4 | "attn_proj": true, |
| 5 | "decoder_dim": 1024, |
| 6 | "decoder_rates": [ |
| 7 | 5, |
| 8 | 5, |
| 9 | 2, |
| 10 | 2, |
| 11 | 2, |
| 12 | 2, |
| 13 | 2 |
| 14 | ], |
| 15 | "decoder_type": "bigvgan", |
| 16 | "encoder_dim": 64, |
| 17 | "encoder_rates": [ |
| 18 | 2, |
| 19 | 4, |
| 20 | 4, |
| 21 | 5, |
| 22 | 5 |
| 23 | ], |
| 24 | "latent_dim": 2048, |
| 25 | "sample_rate": 32000, |
| 26 | "vae_latent_channels": 32 |
| 27 | } |
| 28 | } |
| 29 | } |
| 30 | |