config.json

{
  "model_type": "hibiki",
  "lm_gen_config": {
    "temp": 0.8,
    "temp_text": 0.8,
    "top_k": 250,
    "top_k_text": 50
  },
  "dim": 2048,
  "text_card": 48000,
  "existing_text_padding_id": 3,
  "n_q": 16,
  "dep_q": 8,
  "card": 2048,
  "num_heads": 16,
  "num_layers": 16,
  "hidden_scale": 4.125,
  "causal": true,
  "layer_scale": null,
  "context": 500,
  "max_period": 100000,
  "gating": "silu",
  "norm": "rms_norm_f32",
  "positional_embedding": "rope",
  "depformer_dim": 1024,
  "depformer_dim_feedforward": 4224,
  "depformer_num_heads": 16,
  "depformer_num_layers": 6,
  "depformer_causal": true,
  "depformer_layer_scale": null,
  "depformer_multi_linear": true,
  "depformer_context": 16,
  "depformer_max_period": 10000,
  "depformer_gating": "silu",
  "depformer_pos_emb": "none",
  "depformer_weights_per_step": true,
  "delays": [
    0,
    0,
    2,
    2,
    2,
    2,
    2,
    2,
    2,
    0,
    2,
    2,
    2,
    2,
    2,
    2,
    2
  ],
  "conditioners": {
    "description": {
      "type": "lut",
      "lut": {
        "n_bins": 31,
        "dim": 16,
        "tokenizer": "noop",
        "possible_values": [
          "very_bad",
          "bad",
          "neutral",
          "good",
          "very_good"
        ]
      }
    }
  },
  "fuser": {
    "sum": [
      "description"
    ]
  },
  "mimi_name": "mimi-pytorch-e351c8d8@125.safetensors",
  "moshi_name": "hibikim-pytorch-dc2cf5a5@80.safetensors",
  "tokenizer_name": "tokenizer_spm_48k_multi6_2.model"
}