[
  {
    "metadataOutputVersion" : "3.0",
    "userDefinedMetadata" : {
      "com.github.apple.coremltools.source" : "torch==2.5.0",
      "com.github.apple.coremltools.source_dialect" : "TorchScript",
      "com.github.apple.coremltools.version" : "9.0",
      "com.anemll.context_length" : "4096",
      "com.anemll.lut_bits" : "4",
      "com.anemll.num_chunks" : "1",
      "com.anemll.batch_size" : "64",
      "com.anemll.info" : "Converted with Anemll v0.3.5",
      "com.anemll.chunk_no" : "1"
    },
    "availability" : {
      "macOS" : "15.0",
      "tvOS" : "18.0",
      "visionOS" : "2.0",
      "watchOS" : "11.0",
      "iOS" : "18.0",
      "macCatalyst" : "18.0"
    },
    "inputSchema" : [
      {
        "hasShapeFlexibility" : "0",
        "isOptional" : "0",
        "dataType" : "Float16",
        "formattedType" : "MultiArray (Float16 1 × 1 × 2048)",
        "shortDescription" : "",
        "shape" : "[1, 1, 2048]",
        "name" : "hidden_states",
        "type" : "MultiArray"
      },
      {
        "hasShapeFlexibility" : "0",
        "isOptional" : "0",
        "dataType" : "Int32",
        "formattedType" : "MultiArray (Int32 1)",
        "shortDescription" : "",
        "shape" : "[1]",
        "name" : "position_ids",
        "type" : "MultiArray"
      },
      {
        "hasShapeFlexibility" : "0",
        "isOptional" : "0",
        "dataType" : "Float16",
        "formattedType" : "MultiArray (Float16 1 × 1 × 1 × 4096)",
        "shortDescription" : "",
        "shape" : "[1, 1, 1, 4096]",
        "name" : "causal_mask",
        "type" : "MultiArray"
      },
      {
        "hasShapeFlexibility" : "0",
        "isOptional" : "0",
        "dataType" : "Int32",
        "formattedType" : "MultiArray (Int32 1)",
        "shortDescription" : "",
        "shape" : "[1]",
        "name" : "current_pos",
        "type" : "MultiArray"
      }
    ],
    "outputSchema" : [
      {
        "hasShapeFlexibility" : "0",
        "isOptional" : "0",
        "dataType" : "Float16",
        "formattedType" : "MultiArray (Float16 1 × 1 × 2048)",
        "shortDescription" : "",
        "shape" : "[1, 1, 2048]",
        "name" : "output_hidden_states",
        "type" : "MultiArray"
      }
    ],
    "modelParameters" : [

    ],
    "storagePrecision" : "Mixed (Float16, Palettized (10 bits), Palettized (12 bits), Palettized (14 bits), UInt4)",
    "method" : "predict",
    "functions" : [
      {
        "inputSchema" : [
          {
            "hasShapeFlexibility" : "0",
            "isOptional" : "0",
            "dataType" : "Float16",
            "formattedType" : "MultiArray (Float16 1 × 1 × 2048)",
            "shortDescription" : "",
            "shape" : "[1, 1, 2048]",
            "name" : "hidden_states",
            "type" : "MultiArray"
          },
          {
            "hasShapeFlexibility" : "0",
            "isOptional" : "0",
            "dataType" : "Int32",
            "formattedType" : "MultiArray (Int32 1)",
            "shortDescription" : "",
            "shape" : "[1]",
            "name" : "position_ids",
            "type" : "MultiArray"
          },
          {
            "hasShapeFlexibility" : "0",
            "isOptional" : "0",
            "dataType" : "Float16",
            "formattedType" : "MultiArray (Float16 1 × 1 × 1 × 4096)",
            "shortDescription" : "",
            "shape" : "[1, 1, 1, 4096]",
            "name" : "causal_mask",
            "type" : "MultiArray"
          },
          {
            "hasShapeFlexibility" : "0",
            "isOptional" : "0",
            "dataType" : "Int32",
            "formattedType" : "MultiArray (Int32 1)",
            "shortDescription" : "",
            "shape" : "[1]",
            "name" : "current_pos",
            "type" : "MultiArray"
          }
        ],
        "computePrecision" : "Mixed (Float16, Int16, Int32, UInt16)",
        "storagePrecision" : "Mixed (Float16, Palettized (10 bits), Palettized (12 bits), Palettized (14 bits), UInt4)",
        "stateSchema" : [
          {
            "dataType" : "Float16",
            "isOptional" : "0",
            "formattedType" : "State (Float16 32 × 8 × 4096 × 64)",
            "shortDescription" : "",
            "shape" : "[32, 8, 4096, 64]",
            "name" : "model_model_kv_cache_0",
            "type" : "State"
          }
        ],
        "outputSchema" : [
          {
            "hasShapeFlexibility" : "0",
            "isOptional" : "0",
            "dataType" : "Float16",
            "formattedType" : "MultiArray (Float16 1 × 1 × 2048)",
            "shortDescription" : "",
            "shape" : "[1, 1, 2048]",
            "name" : "output_hidden_states",
            "type" : "MultiArray"
          }
        ],
        "name" : "infer",
        "mlProgramOperationTypeHistogram" : {
          "Ios18.expandDims" : 64,
          "Ios18.mul" : 226,
          "Ios18.matmul" : 32,
          "Identity" : 1,
          "Ios18.exp" : 16,
          "Ios18.realDiv" : 16,
          "Ios18.greaterEqual" : 2,
          "Select" : 2,
          "Ios18.readState" : 33,
          "Tile" : 32,
          "Ios18.gather" : 2,
          "Ios18.add" : 83,
          "Ios18.layerNorm" : 33,
          "Ios18.sliceUpdate" : 32,
          "Ios18.writeState" : 32,
          "Ios18.reshape" : 98,
          "Ios16.reduceMax" : 16,
          "Ios16.reduceSum" : 16,
          "Ios18.constexprLutToDense" : 112,
          "Ios18.conv" : 96,
          "Ios18.concat" : 129,
          "Ios18.transpose" : 64,
          "Ios18.sub" : 48,
          "Ios18.cast" : 5,
          "Ios18.linear" : 16,
          "Ios18.silu" : 16,
          "Ios18.sliceByIndex" : 131,
          "Ios18.squeeze" : 48
        }
      },
      {
        "inputSchema" : [
          {
            "hasShapeFlexibility" : "0",
            "isOptional" : "0",
            "dataType" : "Float16",
            "formattedType" : "MultiArray (Float16 1 × 64 × 2048)",
            "shortDescription" : "",
            "shape" : "[1, 64, 2048]",
            "name" : "hidden_states",
            "type" : "MultiArray"
          },
          {
            "hasShapeFlexibility" : "0",
            "isOptional" : "0",
            "dataType" : "Int32",
            "formattedType" : "MultiArray (Int32 64)",
            "shortDescription" : "",
            "shape" : "[64]",
            "name" : "position_ids",
            "type" : "MultiArray"
          },
          {
            "hasShapeFlexibility" : "0",
            "isOptional" : "0",
            "dataType" : "Float16",
            "formattedType" : "MultiArray (Float16 1 × 1 × 64 × 4096)",
            "shortDescription" : "",
            "shape" : "[1, 1, 64, 4096]",
            "name" : "causal_mask",
            "type" : "MultiArray"
          },
          {
            "hasShapeFlexibility" : "0",
            "isOptional" : "0",
            "dataType" : "Int32",
            "formattedType" : "MultiArray (Int32 1)",
            "shortDescription" : "",
            "shape" : "[1]",
            "name" : "current_pos",
            "type" : "MultiArray"
          }
        ],
        "computePrecision" : "Mixed (Float16, Int16, Int32, UInt16)",
        "storagePrecision" : "Mixed (Float16, Palettized (10 bits), Palettized (12 bits), Palettized (14 bits), UInt4)",
        "stateSchema" : [
          {
            "dataType" : "Float16",
            "isOptional" : "0",
            "formattedType" : "State (Float16 32 × 8 × 4096 × 64)",
            "shortDescription" : "",
            "shape" : "[32, 8, 4096, 64]",
            "name" : "model_model_kv_cache_0",
            "type" : "State"
          }
        ],
        "outputSchema" : [
          {
            "hasShapeFlexibility" : "0",
            "isOptional" : "0",
            "dataType" : "Float16",
            "formattedType" : "MultiArray (Float16 1 × 1 × 2048)",
            "shortDescription" : "",
            "shape" : "[1, 1, 2048]",
            "name" : "output_hidden_states",
            "type" : "MultiArray"
          }
        ],
        "name" : "prefill",
        "mlProgramOperationTypeHistogram" : {
          "Ios18.expandDims" : 63,
          "Ios18.mul" : 221,
          "Ios18.matmul" : 32,
          "Ios18.exp" : 16,
          "Ios18.realDiv" : 16,
          "Ios18.greaterEqual" : 2,
          "Select" : 2,
          "Ios18.readState" : 33,
          "Tile" : 32,
          "Ios18.gather" : 2,
          "Ios18.add" : 82,
          "Ios18.layerNorm" : 31,
          "Ios18.sliceUpdate" : 32,
          "Ios18.writeState" : 32,
          "Ios18.reshape" : 130,
          "Ios16.reduceMax" : 16,
          "Ios16.reduceSum" : 16,
          "Ios18.constexprLutToDense" : 109,
          "Ios18.conv" : 93,
          "Ios18.concat" : 127,
          "Ios18.transpose" : 112,
          "Ios18.sub" : 48,
          "Ios18.cast" : 5,
          "Ios18.linear" : 16,
          "Ios18.silu" : 15,
          "Ios18.sliceByIndex" : 130,
          "Ios18.squeeze" : 47
        }
      }
    ],
    "version" : "0.3.5",
    "isUpdatable" : "0",
    "defaultFunctionName" : "infer",
    "specificationVersion" : 9,
    "stateSchema" : [
      {
        "dataType" : "Float16",
        "isOptional" : "0",
        "formattedType" : "State (Float16 32 × 8 × 4096 × 64)",
        "shortDescription" : "",
        "shape" : "[32, 8, 4096, 64]",
        "name" : "model_model_kv_cache_0",
        "type" : "State"
      }
    ],
    "computePrecision" : "Mixed (Float16, Int16, Int32, UInt16)",
    "mlProgramOperationTypeHistogram" : {
      "Ios18.expandDims" : 64,
      "Ios18.mul" : 226,
      "Ios18.matmul" : 32,
      "Identity" : 1,
      "Ios18.exp" : 16,
      "Ios18.realDiv" : 16,
      "Ios18.greaterEqual" : 2,
      "Select" : 2,
      "Ios18.readState" : 33,
      "Tile" : 32,
      "Ios18.gather" : 2,
      "Ios18.add" : 83,
      "Ios18.layerNorm" : 33,
      "Ios18.sliceUpdate" : 32,
      "Ios18.writeState" : 32,
      "Ios18.reshape" : 98,
      "Ios16.reduceMax" : 16,
      "Ios16.reduceSum" : 16,
      "Ios18.constexprLutToDense" : 112,
      "Ios18.conv" : 96,
      "Ios18.concat" : 129,
      "Ios18.transpose" : 64,
      "Ios18.sub" : 48,
      "Ios18.cast" : 5,
      "Ios18.linear" : 16,
      "Ios18.silu" : 16,
      "Ios18.sliceByIndex" : 131,
      "Ios18.squeeze" : 48
    },
    "shortDescription" : "Anemll Model: Multifunction FFN+Prefill",
    "generatedClassName" : "llama_FFN_PF_lut4_chunk_01of01",
    "author" : "Converted with Anemll v0.3.5",
    "modelType" : {
      "name" : "MLModelType_mlProgram"
    }
  }
]