{
  "family": "miotts",
  "display_name": "MioTTS",
  "description": "Lightweight LLM-based English and Japanese TTS family built on MioCodec, supporting low-latency speech generation and zero-shot voice cloning from short reference audio.",
  "category": "tts",
  "status": "supported",
  "tasks": [
    "tts",
    "clone"
  ],
  "modes": [
    "offline"
  ],
  "languages": [
    "en",
    "ja"
  ],
  "options": {
    "request": [
      {
        "name": "max_tokens",
        "type": "int",
        "description": "Maximum generated tokens; default 700.",
        "required": false,
        "min": 1
      },
      {
        "name": "top_k",
        "type": "int",
        "description": "Generation control; defaults to the checkpoint config.",
        "required": false
      },
      {
        "name": "top_p",
        "type": "float",
        "description": "Generation control; defaults to the checkpoint config.",
        "required": false,
        "min": 0,
        "max": 1
      },
      {
        "name": "temperature",
        "type": "float",
        "description": "Generation control; defaults to the checkpoint config.",
        "required": false,
        "min": 0,
        "max": 2
      },
      {
        "name": "repetition_penalty",
        "type": "float",
        "description": "Generation control; defaults to the checkpoint config.",
        "required": false,
        "min": 1,
        "max": 1.5
      },
      {
        "name": "presence_penalty",
        "type": "float",
        "description": "Presence penalty; default 0.",
        "required": false,
        "min": 0,
        "max": 1
      },
      {
        "name": "frequency_penalty",
        "type": "float",
        "description": "Frequency penalty; default 0.",
        "required": false,
        "min": 0,
        "max": 1
      },
      {
        "name": "do_sample",
        "type": "bool",
        "description": "Generation control; defaults to the checkpoint config.",
        "required": false
      },
      {
        "name": "seed",
        "type": "int",
        "description": "Sampling seed; omitted chooses a random seed.",
        "required": false,
        "min": 0,
        "max": 4294967295
      },
      {
        "name": "best_of_n",
        "type": "int",
        "description": "Candidate count, capped by the session maximum; greater than one enables best-of-N.",
        "required": false,
        "min": 1
      },
      {
        "name": "best_of_n_enabled",
        "type": "bool",
        "description": "Override the session best-of-N setting.",
        "required": false
      },
      {
        "name": "best_of_n_language",
        "type": "enum",
        "description": "ASR language used to rank candidates.",
        "required": false,
        "values": [
          "auto",
          "en",
          "ja"
        ]
      },
      {
        "name": "text_chunk_size",
        "type": "int",
        "description": "Framework long-form text chunk size; inherits the session setting.",
        "required": false,
        "min": 1
      }
    ],
    "session": [
      {
        "name": "weight_type",
        "type": "enum",
        "description": "Weight storage type; defaults to native.",
        "required": false,
        "preset": "weight_type_full"
      },
      {
        "name": "codec_model_path",
        "type": "path",
        "description": "MioCodec model path; defaults to sibling MioCodec-25Hz-44.1kHz-v2.",
        "required": false
      },
      {
        "name": "best_of_n_asr_model_path",
        "type": "path",
        "description": "ASR model path for candidate ranking; defaults to sibling Qwen3-ASR-0.6B.",
        "required": false
      },
      {
        "name": "text_chunk_size",
        "type": "int",
        "description": "Default framework text chunk size; 120 when hanzi, kana and Hangul outnumber other letters and digits, otherwise 180.",
        "required": false
      },
      {
        "name": "best_of_n_enabled",
        "type": "bool",
        "description": "Enable candidate generation and ASR ranking.",
        "required": false,
        "default": false
      },
      {
        "name": "best_of_n_default",
        "type": "int",
        "description": "Default candidate count.",
        "required": false,
        "min": 1,
        "default": 1
      },
      {
        "name": "best_of_n_max",
        "type": "int",
        "description": "Maximum candidate count.",
        "required": false,
        "min": 1,
        "default": 8
      },
      {
        "name": "best_of_n_language",
        "type": "enum",
        "description": "ASR language used to rank candidates.",
        "required": false,
        "values": [
          "auto",
          "en",
          "ja"
        ],
        "default": "auto"
      },
      {
        "name": "prefill_graph_arena_mb",
        "type": "int",
        "description": "Context or graph arena size in MiB.",
        "required": false,
        "min": 1,
        "default": 1024
      },
      {
        "name": "decode_graph_arena_mb",
        "type": "int",
        "description": "Context or graph arena size in MiB.",
        "required": false,
        "min": 1,
        "default": 1024
      },
      {
        "name": "weight_context_mb",
        "type": "int",
        "description": "Context or graph arena size in MiB.",
        "required": false,
        "min": 1,
        "default": 512
      },
      {
        "name": "codec_weight_context_mb",
        "type": "int",
        "description": "Context or graph arena size in MiB.",
        "required": false,
        "min": 1,
        "default": 256
      },
      {
        "name": "codec_constant_context_mb",
        "type": "int",
        "description": "Context or graph arena size in MiB.",
        "required": false,
        "min": 1,
        "default": 256
      },
      {
        "name": "global_graph_arena_mb",
        "type": "int",
        "description": "Context or graph arena size in MiB.",
        "required": false,
        "min": 1,
        "default": 256
      },
      {
        "name": "wave_graph_arena_mb",
        "type": "int",
        "description": "Context or graph arena size in MiB.",
        "required": false,
        "min": 1,
        "default": 512
      }
    ],
    "load": []
  },
  "capabilities": {
    "clone": [
      "speaker_reference"
    ]
  },
  "runtime": {
    "tags": [
      "gguf"
    ]
  },
  "ui": {
    "recommended_package": "miotts_1_7b_q8_0",
    "tags": [
      "TTS",
      "Clone",
      "GGUF"
    ],
    "docs": [
      "docs/models/miotts.md",
      "docs/tts.md",
      "docs/gguf.md"
    ]
  },
  "package_defaults": {
    "download": {
      "kind": "huggingface_snapshot",
      "repo": "audio-cpp/audio.cpp-gguf",
      "revision": "main",
      "gated": false
    }
  },
  "packages": [
    {
      "id": "miotts_1_7b_q8_0",
      "display_name": "MioTTS 1.7B Q8_0 GGUF",
      "default": true,
      "format": "gguf",
      "precision": "q8_0",
      "target_directory": "MioTTS-1.7B-GGUF",
      "files": [
        "MioTTS-1.7B-GGUF/miotts-1.7b-q8_0.gguf"
      ],
      "strip_prefix": "MioTTS-1.7B-GGUF"
    },
    {
      "id": "miotts_1_7b_bf16",
      "display_name": "MioTTS 1.7B BF16 GGUF",
      "format": "gguf",
      "precision": "bf16",
      "target_directory": "MioTTS-1.7B-GGUF",
      "files": [
        "MioTTS-1.7B-GGUF/miotts-1.7b-bf16.gguf"
      ],
      "strip_prefix": "MioTTS-1.7B-GGUF"
    },
    {
      "id": "miotts_1_7b_orig",
      "display_name": "MioTTS 1.7B Original-Dtype GGUF",
      "format": "gguf",
      "precision": "orig",
      "target_directory": "MioTTS-1.7B-GGUF",
      "files": [
        "MioTTS-1.7B-GGUF/miotts-1.7b-orig.gguf"
      ],
      "strip_prefix": "MioTTS-1.7B-GGUF"
    }
  ],
  "sources": [
    {
      "format": "gguf",
      "roots": {
        "model": ".",
        "weights": "$gguf"
      },
      "files": {
        "config": "model:config.json",
        "generation_config": "model:generation_config.json",
        "tokenizer_config": "model:tokenizer_config.json",
        "vocab": "model:vocab.json",
        "merges": "model:merges.txt"
      },
      "optional_files": {
        "tokenizer_json": "model:tokenizer.json"
      },
      "tensors": {
        "weights": "weights:"
      }
    },
    {
      "format": "safetensors",
      "roots": {
        "model": "."
      },
      "files": {
        "config": "model:config.json",
        "generation_config": "model:generation_config.json",
        "tokenizer_config": "model:tokenizer_config.json",
        "vocab": "model:vocab.json",
        "merges": "model:merges.txt"
      },
      "optional_files": {
        "tokenizer_json": "model:tokenizer.json"
      },
      "tensors": {
        "weights": "model:model.safetensors"
      }
    }
  ]
}