diodiogod-TTS-Audio-Suite/utils/audio_cpp/model_specs/fun_asr_nano.json

{
  "schema_version": 1,
  "family": "fun_asr_nano",
  "display_name": "Fun-ASR-Nano",
  "description": "Offline multilingual speech recognition with the FunAudioLLM Fun-ASR-Nano-2512 model.",
  "category": "asr",
  "status": "wip",
  "tasks": [
    "asr"
  ],
  "modes": [
    "offline"
  ],
  "languages": [
    "auto",
    "zh",
    "en",
    "ja"
  ],
  "capabilities": {},
  "options": {
    "request": [
      {
        "name": "language",
        "type": "string",
        "description": "Recognition language, or auto to let the model infer it.",
        "required": false,
        "default": "auto"
      },
      {
        "name": "enable_itn",
        "type": "bool",
        "description": "Enable inverse text normalization in the transcription prompt.",
        "required": false,
        "default": true
      },
      {
        "name": "max_tokens",
        "type": "int",
        "description": "Maximum number of generated transcript tokens.",
        "required": false,
        "min": 1,
        "default": 512
      },
      {
        "name": "audio_chunk_mode",
        "type": "enum",
        "description": "Audio chunking mode: auto, fixed, or none.",
        "values": [
          "auto",
          "fixed",
          "none"
        ],
        "required": false,
        "default": "auto"
      },
      {
        "name": "audio_chunk_seconds",
        "type": "float",
        "description": "Fixed chunk duration in seconds.",
        "required": false,
        "min": 0.001,
        "default": 30
      }
    ],
    "session": [
      {
        "name": "weight_type",
        "type": "enum",
        "description": "Shared model weight storage type.",
        "preset": "weight_type_full",
        "required": false,
        "default": "native"
      }
    ],
    "load": []
  },
  "runtime": {
    "tags": [
      "gguf",
      "server",
      "cuda",
      "metal",
      "cpu"
    ]
  },
  "packages": [
    {
      "id": "fun_asr_nano_2512_q8_0",
      "display_name": "Fun-ASR-Nano-2512 Q8_0 GGUF",
      "description": "Standalone audio.cpp GGUF built from the pinned official checkpoint; governed by the FunASR Model Open Source License Agreement v1.1.",
      "default": true,
      "format": "gguf",
      "precision": "q8_0",
      "target_directory": "Fun-ASR-Nano-2512-GGUF",
      "files": [
        "fun-asr-nano-2512-q8_0.gguf"
      ],
      "download": {
        "kind": "huggingface_snapshot",
        "repo": "FunAudioLLM/Fun-ASR-Nano-2512-GGUF",
        "revision": "ce72677f84900f0dc57f498ace253bfb3c9155b6",
        "gated": false
      }
    },
    {
      "id": "fun_asr_nano_2512_f16",
      "display_name": "Fun-ASR-Nano-2512 F16 GGUF",
      "description": "Standalone audio.cpp GGUF built from the pinned official checkpoint; governed by the FunASR Model Open Source License Agreement v1.1.",
      "format": "gguf",
      "precision": "f16",
      "target_directory": "Fun-ASR-Nano-2512-GGUF",
      "files": [
        "fun-asr-nano-2512-f16.gguf"
      ],
      "download": {
        "kind": "huggingface_snapshot",
        "repo": "FunAudioLLM/Fun-ASR-Nano-2512-GGUF",
        "revision": "ce72677f84900f0dc57f498ace253bfb3c9155b6",
        "gated": false
      }
    },
    {
      "id": "fun_asr_nano_2512_safetensors",
      "display_name": "Fun-ASR-Nano-2512 HF Safetensors",
      "description": "Official checkpoint governed by the FunASR Model Open Source License Agreement v1.1.",
      "format": "safetensors",
      "precision": "native",
      "target_directory": "Fun-ASR-Nano-2512-hf",
      "files": [
        "chat_template.jinja",
        "config.json",
        "generation_config.json",
        "model.safetensors",
        "processor_config.json",
        "tokenizer.json",
        "tokenizer_config.json"
      ],
      "download": {
        "kind": "huggingface_snapshot",
        "repo": "FunAudioLLM/Fun-ASR-Nano-2512-hf",
        "revision": "854d88f94205cd17d2afdb24332130d86fbe654a",
        "gated": false
      }
    }
  ],
  "dependencies": [],
  "ui": {
    "recommended_package": "fun_asr_nano_2512_q8_0",
    "tags": [
      "ASR",
      "GGUF"
    ],
    "docs": [
      "docs/asr.md",
      "docs/gguf.md"
    ],
    "summary": "Offline Fun-ASR-Nano transcription from official safetensors or audio.cpp GGUF."
  },
  "sources": [
    {
      "format": "gguf",
      "roots": {
        "model": ".",
        "weights": "$gguf"
      },
      "files": {
        "config": "model:config.json",
        "generation_config": "model:generation_config.json",
        "processor_config": "model:processor_config.json",
        "tokenizer_json": "model:tokenizer.json"
      },
      "optional_files": {
        "chat_template_jinja": "model:chat_template.jinja",
        "tokenizer_config": "model:tokenizer_config.json"
      },
      "tensors": {
        "weights": "weights:"
      }
    },
    {
      "format": "safetensors",
      "roots": {
        "model": "."
      },
      "files": {
        "config": "model:config.json",
        "generation_config": "model:generation_config.json",
        "processor_config": "model:processor_config.json",
        "tokenizer_json": "model:tokenizer.json"
      },
      "optional_files": {
        "chat_template_jinja": "model:chat_template.jinja",
        "tokenizer_config": "model:tokenizer_config.json"
      },
      "tensors": {
        "weights": "model:model.safetensors"
      }
    }
  ]
}