{
  "family": "irodori_tts",
  "display_name": "Irodori-TTS",
  "description": "Japanese TTS model based on RF-DiT continuous audio latents, supporting zero-shot voice cloning, automatic duration prediction, multimodal voice design, and emoji-style control.",
  "category": "tts",
  "status": "supported",
  "tasks": [
    "tts",
    "clone",
    "design"
  ],
  "modes": [
    "offline"
  ],
  "languages": [
    "ja"
  ],
  "capabilities": {
    "clone": [
      "speaker_reference"
    ],
    "design": [
      "voice_design"
    ]
  },
  "runtime": {
    "tags": [
      "gguf"
    ]
  },
  "ui": {
    "recommended_package": "irodori_tts_600m_v3_voicedesign_q8_0",
    "tags": [
      "TTS",
      "Clone",
      "Design",
      "GGUF"
    ],
    "docs": [
      "docs/tts.md",
      "docs/gguf.md"
    ]
  },
  "package_defaults": {
    "download": {
      "kind": "huggingface_snapshot",
      "repo": "audio-cpp/audio.cpp-gguf",
      "revision": "main",
      "gated": false
    }
  },
  "packages": [
    {
      "id": "irodori_tts_600m_v3_voicedesign_q8_0",
      "display_name": "Irodori-TTS 600M v3 VoiceDesign Q8_0 GGUF",
      "default": true,
      "format": "gguf",
      "precision": "q8_0",
      "target_directory": "Irodori-TTS-600M-v3-VoiceDesign-GGUF",
      "files": [
        "Irodori-TTS-600M-v3-VoiceDesign-GGUF/irodori-tts-600m-v3-voicedesign-q8_0.gguf"
      ],
      "strip_prefix": "Irodori-TTS-600M-v3-VoiceDesign-GGUF"
    },
    {
      "id": "irodori_tts_600m_v3_voicedesign_f16",
      "display_name": "Irodori-TTS 600M v3 VoiceDesign F16 GGUF",
      "format": "gguf",
      "precision": "f16",
      "target_directory": "Irodori-TTS-600M-v3-VoiceDesign-GGUF",
      "files": [
        "Irodori-TTS-600M-v3-VoiceDesign-GGUF/irodori-tts-600m-v3-voicedesign-f16.gguf"
      ],
      "strip_prefix": "Irodori-TTS-600M-v3-VoiceDesign-GGUF"
    },
    {
      "id": "irodori_tts_500m_v3_q8_0",
      "display_name": "Irodori-TTS 500M v3 Q8_0 GGUF",
      "format": "gguf",
      "precision": "q8_0",
      "target_directory": "Irodori-TTS-500M-v3-GGUF",
      "files": [
        "Irodori-TTS-500M-v3-GGUF/irodori-tts-500m-v3-q8_0.gguf"
      ],
      "strip_prefix": "Irodori-TTS-500M-v3-GGUF"
    },
    {
      "id": "irodori_tts_500m_v3_f16",
      "display_name": "Irodori-TTS 500M v3 F16 GGUF",
      "format": "gguf",
      "precision": "f16",
      "target_directory": "Irodori-TTS-500M-v3-GGUF",
      "files": [
        "Irodori-TTS-500M-v3-GGUF/irodori-tts-500m-v3-f16.gguf"
      ],
      "strip_prefix": "Irodori-TTS-500M-v3-GGUF"
    }
  ],
  "sources": [
    {
      "format": "gguf",
      "roots": {
        "model": ".",
        "weights": "$gguf"
      },
      "files": {
        "model_config": "model:model_config.json",
        "tokenizer_json": "model:tokenizer.json"
      },
      "tensors": {
        "model_weights": {
          "source": "weights:",
          "prefix": "model_weights"
        },
        "codec_weights": {
          "source": "weights:",
          "prefix": "codec_weights"
        }
      }
    },
    {
      "format": "safetensors",
      "roots": {
        "model": ".",
        "tokenizer": "../llm-jp-3-150m",
        "codec": "../Semantic-DACVAE-Japanese-32dim"
      },
      "files": {
        "model_config": "model:model_config.json",
        "tokenizer_json": "tokenizer:tokenizer.json"
      },
      "tensors": {
        "model_weights": "model:model.safetensors",
        "codec_weights": "codec:weights.safetensors"
      }
    }
  ]
}