{
  "schema_version": 1,
  "family": "audiosr",
  "display_name": "AudioSR",
  "description": "AudioSR performs latent-diffusion audio super-resolution from an input waveform.",
  "category": "audio_tools",
  "status": "experimental",
  "tasks": [
    "s2s"
  ],
  "modes": [
    "offline"
  ],
  "languages": [
    "auto"
  ],
  "runtime": {
    "tags": [
      "gguf",
      "cuda"
    ]
  },
  "capabilities": {
    "s2s": [
      "audio_enhancement"
    ]
  },
  "options": {
    "request": [
      {
        "name": "num_inference_steps",
        "type": "int",
        "description": "DDIM inference steps. The uniform DDIM schedule must produce exactly this many timesteps; not every count is valid (e.g. 51 fails with 1000 diffusion timesteps).",
        "required": false,
        "min": 1,
        "max": 1000,
        "default": 50
      },
      {
        "name": "guidance_scale",
        "type": "float",
        "description": "Classifier-free guidance scale. Any finite value is accepted.",
        "required": false,
        "default": 3.5
      },
      {
        "name": "ddim_eta",
        "type": "float",
        "description": "DDIM eta; 1.0 matches the official AudioSR default. Must be nonnegative; there is no runtime upper bound.",
        "required": false,
        "min": 0,
        "default": 1
      },
      {
        "name": "audio_chunk_duration_sec",
        "type": "float",
        "description": "Long-audio chunk duration.",
        "required": false,
        "min": 0.001,
        "default": 15
      },
      {
        "name": "audio_chunk_overlap_sec",
        "type": "float",
        "description": "Long-audio chunk overlap; must be smaller than audio_chunk_duration_sec, including after rounding to samples.",
        "required": false,
        "min": 0,
        "default": 2
      },
      {
        "name": "seed",
        "type": "int",
        "description": "Generation seed; omitted or negative values use the fixed seed 42, not a random seed. Zero is a fixed seed.",
        "required": false,
        "max": 2147483647,
        "default": 42
      }
    ],
    "session": [
      {
        "name": "weight_type",
        "type": "enum",
        "description": "AudioSR weight storage type.",
        "preset": "weight_type_full",
        "required": false,
        "default": "native"
      }
    ],
    "load": []
  },
  "package_defaults": {
    "download": {
      "kind": "huggingface_snapshot",
      "repo": "audio-cpp/audio.cpp-gguf",
      "revision": "main",
      "gated": false
    }
  },
  "packages": [
    {
      "id": "audiosr_basic_f32",
      "display_name": "AudioSR Basic F32 GGUF",
      "default": true,
      "format": "gguf",
      "precision": "f32",
      "target_directory": "AudioSR-GGUF",
      "files": [
        "AudioSR-GGUF/audiosr-basic-f32.gguf"
      ],
      "strip_prefix": "AudioSR-GGUF"
    }
  ],
  "dependencies": [],
  "ui": {
    "recommended_package": "audiosr_basic_f32",
    "tags": [
      "GGUF"
    ],
    "docs": [
      "docs/models/audiosr.md",
      "docs/gguf.md"
    ]
  },
  "sources": [
    {
      "format": "gguf",
      "roots": {
        "model": ".",
        "weights": "$gguf"
      },
      "tensors": {
        "weights": {
          "source": "weights:"
        },
        "frontend": {
          "source": "weights:",
          "prefix": "frontend"
        }
      }
    }
  ]
}