[
  {
    "metadataOutputVersion" : "3.0",
    "storagePrecision" : "Mixed (Float32, Palettized (8 bits))",
    "outputSchema" : [
      {
        "hasShapeFlexibility" : "0",
        "isOptional" : "0",
        "dataType" : "Float16",
        "formattedType" : "MultiArray (Float16 1 × 15 × 2048)",
        "shortDescription" : "",
        "shape" : "[1, 15, 2048]",
        "name" : "all_logits",
        "type" : "MultiArray"
      },
      {
        "hasShapeFlexibility" : "0",
        "isOptional" : "0",
        "dataType" : "Float16",
        "formattedType" : "MultiArray (Float16 1 × 1024 × 1 × 1)",
        "shortDescription" : "",
        "shape" : "[1, 1024, 1, 1]",
        "name" : "hidden_states",
        "type" : "MultiArray"
      },
      {
        "hasShapeFlexibility" : "0",
        "isOptional" : "0",
        "dataType" : "Float16",
        "formattedType" : "MultiArray (Float16 1 × 5120 × 1 × 16)",
        "shortDescription" : "",
        "shape" : "[1, 5120, 1, 16]",
        "name" : "new_key_cache",
        "type" : "MultiArray"
      },
      {
        "hasShapeFlexibility" : "0",
        "isOptional" : "0",
        "dataType" : "Float16",
        "formattedType" : "MultiArray (Float16 1 × 5120 × 1 × 16)",
        "shortDescription" : "",
        "shape" : "[1, 5120, 1, 16]",
        "name" : "new_value_cache",
        "type" : "MultiArray"
      }
    ],
    "modelParameters" : [

    ],
    "specificationVersion" : 8,
    "mlProgramOperationTypeHistogram" : {
      "Ios17.cos" : 1,
      "Ios17.reshape" : 45,
      "Ios16.reduceMean" : 21,
      "Ios16.softmax" : 5,
      "Ios17.matmul" : 11,
      "Ios17.transpose" : 11,
      "Ios17.sin" : 1,
      "Ios17.expandDims" : 51,
      "Ios17.add" : 58,
      "Tile" : 10,
      "Ios17.sliceByIndex" : 30,
      "Ios17.squeeze" : 68,
      "Ios17.cast" : 10,
      "Ios17.sub" : 1,
      "Ios16.constexprLutToDense" : 50,
      "Ios17.pow" : 21,
      "Ios16.silu" : 5,
      "Ios17.linear" : 50,
      "Ios17.concat" : 13,
      "Stack" : 1,
      "Ios17.rsqrt" : 21,
      "Ios17.mul" : 106
    },
    "computePrecision" : "Mixed (Float16, Float32, Int32)",
    "isUpdatable" : "0",
    "stateSchema" : [

    ],
    "availability" : {
      "macOS" : "14.0",
      "tvOS" : "17.0",
      "visionOS" : "1.0",
      "watchOS" : "10.0",
      "iOS" : "17.0",
      "macCatalyst" : "17.0"
    },
    "modelType" : {
      "name" : "MLModelType_mlProgram"
    },
    "userDefinedMetadata" : {
      "com.github.apple.coremltools.conversion_date" : "2026-04-03",
      "com.github.apple.coremltools.source" : "torch==2.10.0",
      "com.github.apple.coremltools.version" : "9.0",
      "com.github.apple.coremltools.source_dialect" : "TorchScript"
    },
    "inputSchema" : [
      {
        "hasShapeFlexibility" : "0",
        "isOptional" : "0",
        "dataType" : "Float16",
        "formattedType" : "MultiArray (Float16 1 × 1024 × 1 × 1)",
        "shortDescription" : "",
        "shape" : "[1, 1024, 1, 1]",
        "name" : "input_embeds",
        "type" : "MultiArray"
      },
      {
        "hasShapeFlexibility" : "0",
        "isOptional" : "0",
        "dataType" : "Int32",
        "formattedType" : "MultiArray (Int32 1)",
        "shortDescription" : "",
        "shape" : "[1]",
        "name" : "cache_length",
        "type" : "MultiArray"
      },
      {
        "hasShapeFlexibility" : "0",
        "isOptional" : "0",
        "dataType" : "Float16",
        "formattedType" : "MultiArray (Float16 1 × 5120 × 1 × 16)",
        "shortDescription" : "",
        "shape" : "[1, 5120, 1, 16]",
        "name" : "key_cache",
        "type" : "MultiArray"
      },
      {
        "hasShapeFlexibility" : "0",
        "isOptional" : "0",
        "dataType" : "Float16",
        "formattedType" : "MultiArray (Float16 1 × 16)",
        "shortDescription" : "",
        "shape" : "[1, 16]",
        "name" : "key_padding_mask",
        "type" : "MultiArray"
      },
      {
        "hasShapeFlexibility" : "0",
        "isOptional" : "0",
        "dataType" : "Float16",
        "formattedType" : "MultiArray (Float16 1 × 16)",
        "shortDescription" : "",
        "shape" : "[1, 16]",
        "name" : "kv_cache_update_mask",
        "type" : "MultiArray"
      },
      {
        "hasShapeFlexibility" : "0",
        "isOptional" : "0",
        "dataType" : "Float16",
        "formattedType" : "MultiArray (Float16 1 × 5120 × 1 × 16)",
        "shortDescription" : "",
        "shape" : "[1, 5120, 1, 16]",
        "name" : "value_cache",
        "type" : "MultiArray"
      }
    ],
    "generatedClassName" : "MultiCodeDecoder_fp32",
    "method" : "predict"
  }
]