| { | |
| "model_type": "44khz", | |
| "sample_rate": 44100, | |
| "encoder_dim": 64, | |
| "encoder_rates": [2, 4, 8, 8], | |
| "decoder_dim": 1536, | |
| "decoder_rates": [8, 8, 4, 2], | |
| "latent_dim": 1024, | |
| "n_codebooks": 9, | |
| "codebook_size": 1024, | |
| "codebook_dim": 8, | |
| "hop_length": 512 | |
| } | |
| { | |
| "model_type": "44khz", | |
| "sample_rate": 44100, | |
| "encoder_dim": 64, | |
| "encoder_rates": [2, 4, 8, 8], | |
| "decoder_dim": 1536, | |
| "decoder_rates": [8, 8, 4, 2], | |
| "latent_dim": 1024, | |
| "n_codebooks": 9, | |
| "codebook_size": 1024, | |
| "codebook_dim": 8, | |
| "hop_length": 512 | |
| } | |