voice_audio_multi/model_specs/moss_tts_local.json

{
  "family": "moss_tts_local",
  "display_name": "MOSS-TTS-Local",
  "description": "Flagship MOSS-TTS model for high-fidelity 31-language and code-switched speech, zero-shot voice cloning, long-form generation, and fine-grained Pinyin, phoneme, and duration control.",
  "category": "tts",
  "status": "supported",
  "tasks": [
    "tts",
    "clone"
  ],
  "modes": [
    "offline"
  ],
  "languages": [
    "ar",
    "cs",
    "da",
    "de",
    "el",
    "en",
    "es",
    "fa",
    "fi",
    "fr",
    "he",
    "hi",
    "hu",
    "it",
    "ja",
    "ko",
    "mk",
    "ms",
    "nl",
    "pl",
    "pt",
    "ro",
    "ru",
    "sv",
    "sw",
    "th",
    "tl",
    "tr",
    "vi",
    "yue",
    "zh"
  ],
  "options": {
    "request": [
      {
        "name": "max_tokens",
        "type": "int",
        "description": "Maximum generated audio frames.",
        "required": false,
        "min": 1,
        "default": 4096
      },
      {
        "name": "do_sample",
        "type": "bool",
        "description": "Enable sampling.",
        "required": false,
        "default": true
      },
      {
        "name": "temperature",
        "type": "float",
        "description": "Audio sampling temperature. Must be finite and positive when sampling.",
        "required": false,
        "default": 1.7
      },
      {
        "name": "top_p",
        "type": "float",
        "description": "Audio nucleus sampling probability.",
        "required": false,
        "default": 0.8
      },
      {
        "name": "top_k",
        "type": "int",
        "description": "Audio top-k sampling limit.",
        "required": false,
        "default": 25
      },
      {
        "name": "repetition_penalty",
        "type": "float",
        "description": "Audio repetition penalty. Must be finite and positive when token history is nonempty.",
        "required": false,
        "default": 1
      },
      {
        "name": "text_temperature",
        "type": "float",
        "description": "Text sampling temperature. Must be finite and positive when sampling.",
        "required": false,
        "default": 1
      },
      {
        "name": "text_top_p",
        "type": "float",
        "description": "Text nucleus sampling probability.",
        "required": false,
        "default": 1
      },
      {
        "name": "text_top_k",
        "type": "int",
        "description": "Text top-k sampling limit.",
        "required": false,
        "default": 50
      },
      {
        "name": "seed",
        "type": "int",
        "description": "Sampling seed; omitted chooses a random seed.",
        "required": false,
        "min": 0,
        "max": 4294967295
      },
      {
        "name": "text_chunk_size",
        "type": "int",
        "description": "Framework long-form text chunk size.",
        "required": false,
        "default": 2048,
        "min": 1
      },
      {
        "name": "text_chunk_mode",
        "type": "enum",
        "description": "Framework text chunking mode.",
        "required": false,
        "values": [
          "default",
          "tag_aware",
          "japanese",
          "endline"
        ],
        "default": "default"
      }
    ],
    "session": [
      {
        "name": "weight_type",
        "type": "enum",
        "description": "Backbone weight storage; auto uses F32 on CPU, BF16 on CUDA, native on other backends.",
        "required": false,
        "values": [
          "auto",
          "native",
          "f32",
          "f16",
          "bf16",
          "q4_0",
          "q4_1",
          "q5_0",
          "q5_1",
          "q2_k",
          "q3_k",
          "q4_k",
          "q5_k",
          "q6_k",
          "q8_0",
          "nvfp4"
        ],
        "default": "auto"
      },
      {
        "name": "reference_cache_slots",
        "type": "int",
        "description": "Encoded reference cache slots; zero disables reuse.",
        "required": false,
        "min": 0,
        "default": 1
      }
    ],
    "load": []
  },
  "capabilities": {
    "clone": [
      "speaker_reference"
    ]
  },
  "runtime": {
    "tags": [
      "gguf"
    ]
  },
  "ui": {
    "recommended_package": "moss_tts_local_v1_5_q8_0",
    "tags": [
      "TTS",
      "Clone",
      "GGUF"
    ],
    "docs": [
      "docs/models/moss_tts_local.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": "moss_tts_local_v1_5_q8_0",
      "display_name": "MOSS-TTS-Local v1.5 Q8_0 GGUF",
      "default": true,
      "format": "gguf",
      "precision": "q8_0",
      "target_directory": "MOSS-TTS-Local-v1.5-GGUF",
      "files": [
        "MOSS-TTS-Local-v1.5-GGUF/moss-tts-local-v1.5-q8_0.gguf"
      ],
      "strip_prefix": "MOSS-TTS-Local-v1.5-GGUF"
    },
    {
      "id": "moss_tts_local_v1_5_bf16",
      "display_name": "MOSS-TTS-Local v1.5 BF16 GGUF",
      "format": "gguf",
      "precision": "bf16",
      "target_directory": "MOSS-TTS-Local-v1.5-GGUF",
      "files": [
        "MOSS-TTS-Local-v1.5-GGUF/moss-tts-local-v1.5-bf16.gguf"
      ],
      "strip_prefix": "MOSS-TTS-Local-v1.5-GGUF"
    }
  ],
  "sources": [
    {
      "format": "gguf",
      "roots": {
        "model": ".",
        "weights": "$gguf"
      },
      "files": {
        "config": "model:config.json",
        "tokenizer_json": "model:tokenizer.json",
        "tokenizer_config": "model:tokenizer_config.json",
        "tokenizer_vocab": "model:vocab.json",
        "tokenizer_merges": "model:merges.txt",
        "audio_tokenizer_config": "model:audio_tokenizer/config.json"
      },
      "tensors": {
        "model_weights": {
          "source": "weights:",
          "prefix": "model_weights"
        },
        "audio_tokenizer_weights": {
          "source": "weights:",
          "prefix": "audio_tokenizer_weights"
        }
      }
    },
    {
      "format": "safetensors",
      "roots": {
        "model": ".",
        "audio_tokenizer": "audio_tokenizer"
      },
      "files": {
        "config": "model:config.json",
        "tokenizer_json": "model:tokenizer.json",
        "tokenizer_config": "model:tokenizer_config.json",
        "tokenizer_vocab": "model:vocab.json",
        "tokenizer_merges": "model:merges.txt",
        "audio_tokenizer_config": "audio_tokenizer:config.json"
      },
      "tensors": {
        "model_weights": "model:model.safetensors",
        "audio_tokenizer_weights": "audio_tokenizer:model.safetensors.index.json"
      }
    }
  ]
}