[ { "metadataOutputVersion" : "3.0", "storagePrecision" : "Float16", "outputSchema" : [ { "hasShapeFlexibility" : "0", "isOptional" : "0", "dataType" : "Float16", "formattedType" : "MultiArray (Float16 1 × 2 × 3840)", "shortDescription" : "", "shape" : "[1, 2, 3840]", "name" : "audio", "type" : "MultiArray" } ], "modelParameters" : [ ], "specificationVersion" : 9, "mlProgramOperationTypeHistogram" : { "Ios18.linear" : 56, "Ios18.readState" : 24, "Ios18.sub" : 24, "Ios18.matmul" : 24, "Ios18.gelu" : 12, "Stack" : 24, "Ios18.concat" : 24, "Ios18.add" : 60, "Ios18.sliceUpdate" : 24, "Ios18.sliceByIndex" : 108, "Ios18.layerNorm" : 24, "Ios18.transpose" : 38, "Ios18.softmax" : 12, "Ios18.writeState" : 24, "Ios18.reshape" : 82, "Ios18.mul" : 132 }, "computePrecision" : "Mixed (Float16, Int32)", "isUpdatable" : "0", "stateSchema" : [ { "dataType" : "Float16", "isOptional" : "0", "formattedType" : "State (Float16 1 × 4 × 500 × 64)", "shortDescription" : "", "shape" : "[1, 4, 500, 64]", "name" : "s0_l0_k", "type" : "State" }, { "dataType" : "Float16", "isOptional" : "0", "formattedType" : "State (Float16 1 × 4 × 500 × 64)", "shortDescription" : "", "shape" : "[1, 4, 500, 64]", "name" : "s0_l0_v", "type" : "State" }, { "dataType" : "Float16", "isOptional" : "0", "formattedType" : "State (Float16 1 × 4 × 500 × 64)", "shortDescription" : "", "shape" : "[1, 4, 500, 64]", "name" : "s0_l1_k", "type" : "State" }, { "dataType" : "Float16", "isOptional" : "0", "formattedType" : "State (Float16 1 × 4 × 500 × 64)", "shortDescription" : "", "shape" : "[1, 4, 500, 64]", "name" : "s0_l1_v", "type" : "State" }, { "dataType" : "Float16", "isOptional" : "0", "formattedType" : "State (Float16 1 × 4 × 500 × 64)", "shortDescription" : "", "shape" : "[1, 4, 500, 64]", "name" : "s0_l2_k", "type" : "State" }, { "dataType" : "Float16", "isOptional" : "0", "formattedType" : "State (Float16 1 × 4 × 500 × 64)", "shortDescription" : "", "shape" : "[1, 4, 500, 64]", "name" : "s0_l2_v", "type" : "State" }, { "dataType" : "Float16", "isOptional" : "0", "formattedType" : "State (Float16 1 × 4 × 500 × 64)", "shortDescription" : "", "shape" : "[1, 4, 500, 64]", "name" : "s0_l3_k", "type" : "State" }, { "dataType" : "Float16", "isOptional" : "0", "formattedType" : "State (Float16 1 × 4 × 500 × 64)", "shortDescription" : "", "shape" : "[1, 4, 500, 64]", "name" : "s0_l3_v", "type" : "State" }, { "dataType" : "Float16", "isOptional" : "0", "formattedType" : "State (Float16 1 × 4 × 800 × 64)", "shortDescription" : "", "shape" : "[1, 4, 800, 64]", "name" : "s1_l0_k", "type" : "State" }, { "dataType" : "Float16", "isOptional" : "0", "formattedType" : "State (Float16 1 × 4 × 800 × 64)", "shortDescription" : "", "shape" : "[1, 4, 800, 64]", "name" : "s1_l0_v", "type" : "State" }, { "dataType" : "Float16", "isOptional" : "0", "formattedType" : "State (Float16 1 × 4 × 800 × 64)", "shortDescription" : "", "shape" : "[1, 4, 800, 64]", "name" : "s1_l1_k", "type" : "State" }, { "dataType" : "Float16", "isOptional" : "0", "formattedType" : "State (Float16 1 × 4 × 800 × 64)", "shortDescription" : "", "shape" : "[1, 4, 800, 64]", "name" : "s1_l1_v", "type" : "State" }, { "dataType" : "Float16", "isOptional" : "0", "formattedType" : "State (Float16 1 × 4 × 1200 × 64)", "shortDescription" : "", "shape" : "[1, 4, 1200, 64]", "name" : "s2_l0_k", "type" : "State" }, { "dataType" : "Float16", "isOptional" : "0", "formattedType" : "State (Float16 1 × 4 × 1200 × 64)", "shortDescription" : "", "shape" : "[1, 4, 1200, 64]", "name" : "s2_l0_v", "type" : "State" }, { "dataType" : "Float16", "isOptional" : "0", "formattedType" : "State (Float16 1 × 4 × 1200 × 64)", "shortDescription" : "", "shape" : "[1, 4, 1200, 64]", "name" : "s2_l1_k", "type" : "State" }, { "dataType" : "Float16", "isOptional" : "0", "formattedType" : "State (Float16 1 × 4 × 1200 × 64)", "shortDescription" : "", "shape" : "[1, 4, 1200, 64]", "name" : "s2_l1_v", "type" : "State" }, { "dataType" : "Float16", "isOptional" : "0", "formattedType" : "State (Float16 1 × 4 × 1600 × 64)", "shortDescription" : "", "shape" : "[1, 4, 1600, 64]", "name" : "s3_l0_k", "type" : "State" }, { "dataType" : "Float16", "isOptional" : "0", "formattedType" : "State (Float16 1 × 4 × 1600 × 64)", "shortDescription" : "", "shape" : "[1, 4, 1600, 64]", "name" : "s3_l0_v", "type" : "State" }, { "dataType" : "Float16", "isOptional" : "0", "formattedType" : "State (Float16 1 × 4 × 1600 × 64)", "shortDescription" : "", "shape" : "[1, 4, 1600, 64]", "name" : "s3_l1_k", "type" : "State" }, { "dataType" : "Float16", "isOptional" : "0", "formattedType" : "State (Float16 1 × 4 × 1600 × 64)", "shortDescription" : "", "shape" : "[1, 4, 1600, 64]", "name" : "s3_l1_v", "type" : "State" }, { "dataType" : "Float16", "isOptional" : "0", "formattedType" : "State (Float16 1 × 4 × 1600 × 64)", "shortDescription" : "", "shape" : "[1, 4, 1600, 64]", "name" : "s3_l2_k", "type" : "State" }, { "dataType" : "Float16", "isOptional" : "0", "formattedType" : "State (Float16 1 × 4 × 1600 × 64)", "shortDescription" : "", "shape" : "[1, 4, 1600, 64]", "name" : "s3_l2_v", "type" : "State" }, { "dataType" : "Float16", "isOptional" : "0", "formattedType" : "State (Float16 1 × 4 × 1600 × 64)", "shortDescription" : "", "shape" : "[1, 4, 1600, 64]", "name" : "s3_l3_k", "type" : "State" }, { "dataType" : "Float16", "isOptional" : "0", "formattedType" : "State (Float16 1 × 4 × 1600 × 64)", "shortDescription" : "", "shape" : "[1, 4, 1600, 64]", "name" : "s3_l3_v", "type" : "State" } ], "availability" : { "macOS" : "15.0", "tvOS" : "18.0", "visionOS" : "2.0", "watchOS" : "11.0", "iOS" : "18.0", "macCatalyst" : "18.0" }, "modelType" : { "name" : "MLModelType_mlProgram" }, "userDefinedMetadata" : { "com.github.apple.coremltools.conversion_date" : "2026-08-03", "com.github.apple.coremltools.source" : "torch==2.7.0", "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 × 768 × 1)", "shortDescription" : "", "shape" : "[1, 768, 1]", "name" : "latent", "type" : "MultiArray" }, { "hasShapeFlexibility" : "0", "isOptional" : "0", "dataType" : "Float16", "formattedType" : "MultiArray (Float16 1 × 1 × 4 × 32)", "shortDescription" : "", "shape" : "[1, 1, 4, 32]", "name" : "s0_cos", "type" : "MultiArray" }, { "hasShapeFlexibility" : "0", "isOptional" : "0", "dataType" : "Float16", "formattedType" : "MultiArray (Float16 1 × 1 × 4 × 32)", "shortDescription" : "", "shape" : "[1, 1, 4, 32]", "name" : "s0_sin", "type" : "MultiArray" }, { "hasShapeFlexibility" : "0", "isOptional" : "0", "dataType" : "Float16", "formattedType" : "MultiArray (Float16 1 × 1 × 4 × 500)", "shortDescription" : "", "shape" : "[1, 1, 4, 500]", "name" : "s0_mask", "type" : "MultiArray" }, { "hasShapeFlexibility" : "0", "isOptional" : "0", "dataType" : "Float16", "formattedType" : "MultiArray (Float16 1 × 1 × 8 × 32)", "shortDescription" : "", "shape" : "[1, 1, 8, 32]", "name" : "s1_cos", "type" : "MultiArray" }, { "hasShapeFlexibility" : "0", "isOptional" : "0", "dataType" : "Float16", "formattedType" : "MultiArray (Float16 1 × 1 × 8 × 32)", "shortDescription" : "", "shape" : "[1, 1, 8, 32]", "name" : "s1_sin", "type" : "MultiArray" }, { "hasShapeFlexibility" : "0", "isOptional" : "0", "dataType" : "Float16", "formattedType" : "MultiArray (Float16 1 × 1 × 8 × 800)", "shortDescription" : "", "shape" : "[1, 1, 8, 800]", "name" : "s1_mask", "type" : "MultiArray" }, { "hasShapeFlexibility" : "0", "isOptional" : "0", "dataType" : "Float16", "formattedType" : "MultiArray (Float16 1 × 1 × 16 × 32)", "shortDescription" : "", "shape" : "[1, 1, 16, 32]", "name" : "s2_cos", "type" : "MultiArray" }, { "hasShapeFlexibility" : "0", "isOptional" : "0", "dataType" : "Float16", "formattedType" : "MultiArray (Float16 1 × 1 × 16 × 32)", "shortDescription" : "", "shape" : "[1, 1, 16, 32]", "name" : "s2_sin", "type" : "MultiArray" }, { "hasShapeFlexibility" : "0", "isOptional" : "0", "dataType" : "Float16", "formattedType" : "MultiArray (Float16 1 × 1 × 16 × 1200)", "shortDescription" : "", "shape" : "[1, 1, 16, 1200]", "name" : "s2_mask", "type" : "MultiArray" }, { "hasShapeFlexibility" : "0", "isOptional" : "0", "dataType" : "Float16", "formattedType" : "MultiArray (Float16 1 × 1 × 32 × 32)", "shortDescription" : "", "shape" : "[1, 1, 32, 32]", "name" : "s3_cos", "type" : "MultiArray" }, { "hasShapeFlexibility" : "0", "isOptional" : "0", "dataType" : "Float16", "formattedType" : "MultiArray (Float16 1 × 1 × 32 × 32)", "shortDescription" : "", "shape" : "[1, 1, 32, 32]", "name" : "s3_sin", "type" : "MultiArray" }, { "hasShapeFlexibility" : "0", "isOptional" : "0", "dataType" : "Float16", "formattedType" : "MultiArray (Float16 1 × 1 × 32 × 1600)", "shortDescription" : "", "shape" : "[1, 1, 32, 1600]", "name" : "s3_mask", "type" : "MultiArray" } ], "generatedClassName" : "MossCodecDecoder", "method" : "predict" } ]