voice_audio_multi/model_specs/index_echo.json

{
  "schema_version": 1,
  "family": "index_echo",
  "display_name": "Index-Echo-S2TT",
  "description": "Index-Echo speech-to-text translation with bilingual timestamped subtitles.",
  "category": "asr",
  "status": "experimental",
  "tasks": [
    "asr"
  ],
  "modes": [
    "offline"
  ],
  "languages": [
    "zh",
    "en",
    "es",
    "ja"
  ],
  "runtime": {
    "tags": [
      "gguf"
    ]
  },
  "capabilities": {},
  "dependencies": [],
  "options": {
    "request": [
      {
        "name": "target_language",
        "type": "enum",
        "values": ["en", "es", "ja"],
        "description": "Translation language for Chinese input audio.",
        "required": false,
        "default": "en"
      },
      {
        "name": "max_tokens",
        "type": "int",
        "description": "Maximum translated text tokens per window.",
        "required": false,
        "min": 1,
        "default": 2000
      },
      {
        "name": "temperature",
        "type": "float",
        "description": "Sampling temperature; zero selects greedy decoding.",
        "required": false,
        "min": 0,
        "max": 2,
        "default": 0
      },
      {
        "name": "seed",
        "type": "int",
        "description": "Sampling seed when temperature is positive.",
        "required": false,
        "min": 0,
        "max": 2147483647,
        "default": 0
      },
      {
        "name": "glossary",
        "type": "string",
        "description": "Translation glossary entries.",
        "required": false
      },
      {
        "name": "audio_chunk_mode",
        "type": "enum",
        "values": ["auto", "quiet_energy", "fixed", "none"],
        "description": "Long-form audio windowing mode.",
        "required": false,
        "default": "auto"
      },
      {
        "name": "audio_chunk_duration_sec",
        "type": "float",
        "description": "Maximum audio-window duration.",
        "required": false,
        "min": 1,
        "max": 60,
        "default": 60
      }
    ],
    "session": [
      {
        "name": "weight_type",
        "type": "enum",
        "preset": "weight_type_full",
        "description": "Weight storage type for runtime matmul operations.",
        "required": false,
        "default": "native"
      }
    ],
    "load": []
  },
  "packages": [
    {
      "id": "index_echo_s2tt_2b_orig",
      "display_name": "Index-Echo-S2TT 2B Original",
      "default": true,
      "format": "gguf",
      "precision": "orig",
      "target_directory": "Index-Echo-S2TT-GGUF",
      "files": ["index-echo-s2tt-2b-orig.gguf"],
      "download": {
        "kind": "huggingface_snapshot",
        "repo": "audio-cpp/Index-Echo-S2TT-GGUF",
        "revision": "main",
        "gated": false
      }
    },
    {
      "id": "index_echo_s2tt_2b_q8_0",
      "display_name": "Index-Echo-S2TT 2B Q8_0",
      "format": "gguf",
      "precision": "q8_0",
      "target_directory": "Index-Echo-S2TT-GGUF",
      "files": ["index-echo-s2tt-2b-q8_0.gguf"],
      "download": {
        "kind": "huggingface_snapshot",
        "repo": "audio-cpp/Index-Echo-S2TT-GGUF",
        "revision": "main",
        "gated": false
      }
    },
    {
      "id": "index_echo_s2tt_2b_q4_k",
      "display_name": "Index-Echo-S2TT 2B Q4_K (experimental)",
      "format": "gguf",
      "precision": "q4_k",
      "target_directory": "Index-Echo-S2TT-GGUF",
      "files": ["index-echo-s2tt-2b-q4_k.gguf"],
      "download": {
        "kind": "huggingface_snapshot",
        "repo": "audio-cpp/Index-Echo-S2TT-GGUF",
        "revision": "main",
        "gated": false
      }
    },
    {
      "id": "index_echo_s2tt_9b_orig",
      "display_name": "Index-Echo-S2TT 9B Original",
      "format": "gguf",
      "precision": "orig",
      "target_directory": "Index-Echo-S2TT-GGUF",
      "files": ["index-echo-s2tt-9b-orig.gguf"],
      "download": {
        "kind": "huggingface_snapshot",
        "repo": "audio-cpp/Index-Echo-S2TT-GGUF",
        "revision": "main",
        "gated": false
      }
    },
    {
      "id": "index_echo_s2tt_9b_q8_0",
      "display_name": "Index-Echo-S2TT 9B Q8_0",
      "format": "gguf",
      "precision": "q8_0",
      "target_directory": "Index-Echo-S2TT-GGUF",
      "files": ["index-echo-s2tt-9b-q8_0.gguf"],
      "download": {
        "kind": "huggingface_snapshot",
        "repo": "audio-cpp/Index-Echo-S2TT-GGUF",
        "revision": "main",
        "gated": false
      }
    }
  ],
  "ui": {
    "recommended_package": "index_echo_s2tt_2b_orig",
    "tags": ["ASR", "GGUF"],
    "docs": ["docs/models/index_echo.md"]
  },
  "sources": [
    {
      "format": "gguf",
      "roots": {
        "model": ".",
        "weights": "$gguf"
      },
      "files": {
        "llm_config": "model:llm/config.json",
        "tokenizer": "model:llm/tokenizer.json",
        "tokenizer_config": "model:llm/tokenizer_config.json",
        "audio_config": "model:audio_config.json"
      },
      "optional_files": {
        "mapper_config": "model:bridge/mapper_config.json"
      },
      "tensors": {
        "stlm_weights": {"source": "weights:", "prefix": "stlm"},
        "audio_weights": {"source": "weights:", "prefix": "audio"},
        "connector_weights": {"source": "weights:", "prefix": "connector"}
      },
      "optional_tensors": {
        "mapper_weights": {"source": "weights:", "prefix": "mapper"}
      }
    }
  ]
}