| [ | |
| { | |
| "metadataOutputVersion" : "3.0", | |
| "userDefinedMetadata" : { | |
| "com.github.apple.coremltools.source" : "torch==2.5.0", | |
| "com.github.apple.coremltools.version" : "8.2", | |
| "com.anemll.context_length" : "1546", | |
| "com.github.apple.coremltools.source_dialect" : "TorchScript", | |
| "com.anemll.chunk_no" : "1", | |
| "com.anemll.num_chunks" : "2", | |
| "com.anemll.info" : "Converted with Anemll v0.1.2", | |
| "com.anemll.batch_size" : "64", | |
| "com.anemll.lut_bits" : "4" | |
| }, | |
| "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 × 1546)", | |
| "shortDescription" : "", | |
| "shape" : "[1, 1, 1, 1546]", | |
| "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))", | |
| "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 × 1546)", | |
| "shortDescription" : "", | |
| "shape" : "[1, 1, 1, 1546]", | |
| "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, Int32)", | |
| "storagePrecision" : "Mixed (Float16, Palettized (10 bits), Palettized (12 bits), Palettized (14 bits))", | |
| "stateSchema" : [ | |
| { | |
| "dataType" : "Float16", | |
| "isOptional" : "0", | |
| "formattedType" : "State (Float16 32 × 8 × 1546 × 64)", | |
| "shortDescription" : "", | |
| "shape" : "[32, 8, 1546, 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" : 32, | |
| "Ios18.mul" : 80, | |
| "Ios18.matmul" : 16, | |
| "Identity" : 1, | |
| "Ios16.reduceMean" : 16, | |
| "Ios18.exp" : 8, | |
| "Ios18.realDiv" : 8, | |
| "Ios18.greaterEqual" : 1, | |
| "Select" : 1, | |
| "Ios18.readState" : 17, | |
| "Tile" : 16, | |
| "Ios18.gather" : 2, | |
| "Ios18.add" : 42, | |
| "Ios18.layerNorm" : 16, | |
| "Ios18.sliceUpdate" : 16, | |
| "Ios18.writeState" : 16, | |
| "Ios18.reshape" : 50, | |
| "Ios16.reduceMax" : 8, | |
| "Ios16.reduceSum" : 8, | |
| "Ios18.constexprLutToDense" : 56, | |
| "Ios18.conv" : 48, | |
| "Ios18.concat" : 48, | |
| "Ios18.transpose" : 32, | |
| "Ios18.sub" : 40, | |
| "Ios18.linear" : 8, | |
| "Ios18.silu" : 8, | |
| "Ios18.sliceByIndex" : 50, | |
| "Ios18.squeeze" : 24 | |
| } | |
| }, | |
| { | |
| "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 × 1546)", | |
| "shortDescription" : "", | |
| "shape" : "[1, 1, 64, 1546]", | |
| "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, Int32)", | |
| "storagePrecision" : "Mixed (Float16, Palettized (10 bits), Palettized (12 bits), Palettized (14 bits))", | |
| "stateSchema" : [ | |
| { | |
| "dataType" : "Float16", | |
| "isOptional" : "0", | |
| "formattedType" : "State (Float16 32 × 8 × 1546 × 64)", | |
| "shortDescription" : "", | |
| "shape" : "[32, 8, 1546, 64]", | |
| "name" : "model_model_kv_cache_0", | |
| "type" : "State" | |
| } | |
| ], | |
| "outputSchema" : [ | |
| { | |
| "hasShapeFlexibility" : "0", | |
| "isOptional" : "0", | |
| "dataType" : "Float16", | |
| "formattedType" : "MultiArray (Float16 1 × 64 × 2048)", | |
| "shortDescription" : "", | |
| "shape" : "[1, 64, 2048]", | |
| "name" : "output_hidden_states", | |
| "type" : "MultiArray" | |
| } | |
| ], | |
| "name" : "prefill", | |
| "mlProgramOperationTypeHistogram" : { | |
| "Ios18.expandDims" : 32, | |
| "Ios18.mul" : 80, | |
| "Ios18.matmul" : 16, | |
| "Ios16.reduceMean" : 16, | |
| "Ios18.exp" : 8, | |
| "Ios18.realDiv" : 8, | |
| "Ios18.greaterEqual" : 1, | |
| "Select" : 1, | |
| "Ios18.readState" : 17, | |
| "Tile" : 16, | |
| "Ios18.gather" : 2, | |
| "Ios18.add" : 42, | |
| "Ios18.layerNorm" : 16, | |
| "Ios18.sliceUpdate" : 16, | |
| "Ios18.writeState" : 16, | |
| "Ios18.reshape" : 66, | |
| "Ios16.reduceMax" : 8, | |
| "Ios16.reduceSum" : 8, | |
| "Ios18.constexprLutToDense" : 56, | |
| "Ios18.conv" : 48, | |
| "Ios18.concat" : 48, | |
| "Ios18.transpose" : 58, | |
| "Ios18.sub" : 40, | |
| "Ios18.linear" : 8, | |
| "Ios18.silu" : 8, | |
| "Ios18.sliceByIndex" : 50, | |
| "Ios18.squeeze" : 24 | |
| } | |
| } | |
| ], | |
| "version" : "0.1.2", | |
| "isUpdatable" : "0", | |
| "defaultFunctionName" : "infer", | |
| "specificationVersion" : 9, | |
| "stateSchema" : [ | |
| { | |
| "dataType" : "Float16", | |
| "isOptional" : "0", | |
| "formattedType" : "State (Float16 32 × 8 × 1546 × 64)", | |
| "shortDescription" : "", | |
| "shape" : "[32, 8, 1546, 64]", | |
| "name" : "model_model_kv_cache_0", | |
| "type" : "State" | |
| } | |
| ], | |
| "computePrecision" : "Mixed (Float16, Int32)", | |
| "mlProgramOperationTypeHistogram" : { | |
| "Ios18.expandDims" : 32, | |
| "Ios18.mul" : 80, | |
| "Ios18.matmul" : 16, | |
| "Identity" : 1, | |
| "Ios16.reduceMean" : 16, | |
| "Ios18.exp" : 8, | |
| "Ios18.realDiv" : 8, | |
| "Ios18.greaterEqual" : 1, | |
| "Select" : 1, | |
| "Ios18.readState" : 17, | |
| "Tile" : 16, | |
| "Ios18.gather" : 2, | |
| "Ios18.add" : 42, | |
| "Ios18.layerNorm" : 16, | |
| "Ios18.sliceUpdate" : 16, | |
| "Ios18.writeState" : 16, | |
| "Ios18.reshape" : 50, | |
| "Ios16.reduceMax" : 8, | |
| "Ios16.reduceSum" : 8, | |
| "Ios18.constexprLutToDense" : 56, | |
| "Ios18.conv" : 48, | |
| "Ios18.concat" : 48, | |
| "Ios18.transpose" : 32, | |
| "Ios18.sub" : 40, | |
| "Ios18.linear" : 8, | |
| "Ios18.silu" : 8, | |
| "Ios18.sliceByIndex" : 50, | |
| "Ios18.squeeze" : 24 | |
| }, | |
| "shortDescription" : "Anemll Model: Multifunction FFN+Prefill", | |
| "generatedClassName" : "llama_FFN_PF_lut4_chunk_01of02", | |
| "author" : "Converted with Anemll v0.1.2", | |
| "modelType" : { | |
| "name" : "MLModelType_mlProgram" | |
| } | |
| } | |
| ] |