config.json

{
  "model_type": "personaplex",
  "version": "personaplex-7b-v1",
  "base_model": "kyutai/moshiko-pytorch-bf16",
  "temporal": {
    "dim": 4096,
    "num_layers": 32,
    "num_heads": 32,
    "hidden_scale": 4.125,
    "n_q": 8,
    "card": 2048,
    "text_card": 32000,
    "context": 3000,
    "max_period": 10000
  },
  "depformer": {
    "dim": 1024,
    "num_layers": 6,
    "num_heads": 16,
    "dim_feedforward": 2816,
    "num_steps": 16,
    "card": 2048,
    "text_card": 32000,
    "context": 8,
    "weights_per_step": true,
    "multi_linear": true
  },
  "mimi": {
    "sample_rate": 24000,
    "frame_rate": 12.5,
    "num_codebooks": 16,
    "codebook_size": 2048,
    "codebook_dim": 256,
    "dimension": 512,
    "seanet_ratios": [
      8,
      6,
      5,
      4
    ],
    "transformer_layers": 8
  },
  "sampling": {
    "audio_temp": 0.8,
    "audio_top_k": 250,
    "text_temp": 0.7,
    "text_top_k": 25
  },
  "delays": [
    0,
    0,
    1,
    1,
    1,
    1,
    1,
    1,
    1,
    0,
    1,
    1,
    1,
    1,
    1,
    1,
    1
  ],
  "quantization": {
    "bits": 8,
    "group_size": 64,
    "quantized_components": [
      "temporal",
      "depformer"
    ]
  },
  "voices": [
    "NATF0",
    "NATF1",
    "NATF2",
    "NATF3",
    "NATM0",
    "NATM1",
    "NATM2",
    "NATM3",
    "VARF0",
    "VARF1",
    "VARF2",
    "VARF3",
    "VARF4",
    "VARM0",
    "VARM1",
    "VARM2",
    "VARM3",
    "VARM4"
  ]
}