Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

config.json

@@ -0,0 +1,4 @@
+{
+  "tokenizer_class": "AutoTokenizer",
+  "model_type": "phi3"
+}

meta.yaml

@@ -0,0 +1,24 @@
+model_info:
+  name: anemll-5a149550068a1eb93398160d8953f5f56c3603e9-ctx2048
+  version: 0.3.0
+  description: |
+    Demonstarates running 5a149550068a1eb93398160d8953f5f56c3603e9 on Apple Neural Engine
+    Context length: 2048
+    Batch size: 64
+    Chunks: 2
+  license: MIT
+  author: Anemll
+  framework: Core ML
+  language: Python
+  parameters:
+    context_length: 2048
+    batch_size: 64
+    lut_embeddings: 8
+    lut_ffn: 4
+    lut_lmhead: 6
+    num_chunks: 2
+    model_prefix: phi4
+    embeddings: phi4_embeddings_lut8.mlmodelc
+    lm_head: phi4_lm_head_lut6.mlmodelc
+    ffn: phi4_FFN_PF_lut4.mlmodelc
+    split_lm_head: 1

phi4_FFN_PF_lut4_chunk_01of02.mlmodelc/analytics/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9100ee1f47c46d82899674cb8dd1267b7b3be5159f56687398744760d68e748a
+size 243

phi4_FFN_PF_lut4_chunk_01of02.mlmodelc/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1dea759822bbf8031a58a2da1da1d9cca24edc09a4ba2c8a1d49917d5cc02716
+size 743

phi4_FFN_PF_lut4_chunk_01of02.mlmodelc/metadata.json

@@ -0,0 +1,299 @@
+[
+  {
+    "metadataOutputVersion" : "3.0",
+    "userDefinedMetadata" : {
+      "com.github.apple.coremltools.source_dialect" : "TorchScript",
+      "com.github.apple.coremltools.source" : "torch==2.8.0",
+      "com.github.apple.coremltools.version" : "9.0b1",
+      "com.anemll.info" : "Converted with Anemll v0.3.4"
+    },
+    "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 × 3072)",
+        "shortDescription" : "",
+        "shape" : "[1, 1, 3072]",
+        "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 × 2048)",
+        "shortDescription" : "",
+        "shape" : "[1, 1, 1, 2048]",
+        "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 × 3072)",
+        "shortDescription" : "",
+        "shape" : "[1, 1, 3072]",
+        "name" : "output_hidden_states",
+        "type" : "MultiArray"
+      }
+    ],
+    "modelParameters" : [
+
+    ],
+    "storagePrecision" : "Mixed (Float16, Int32, Palettized (11 bits), Palettized (13 bits), Palettized (15 bits), Palettized (7 bits), UInt4)",
+    "method" : "predict",
+    "functions" : [
+      {
+        "inputSchema" : [
+          {
+            "hasShapeFlexibility" : "0",
+            "isOptional" : "0",
+            "dataType" : "Float16",
+            "formattedType" : "MultiArray (Float16 1 × 1 × 3072)",
+            "shortDescription" : "",
+            "shape" : "[1, 1, 3072]",
+            "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 × 2048)",
+            "shortDescription" : "",
+            "shape" : "[1, 1, 1, 2048]",
+            "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, Float32, Int32)",
+        "storagePrecision" : "Mixed (Float16, Int32, Palettized (11 bits), Palettized (13 bits), Palettized (15 bits), Palettized (7 bits), UInt4)",
+        "stateSchema" : [
+
+        ],
+        "outputSchema" : [
+          {
+            "hasShapeFlexibility" : "0",
+            "isOptional" : "0",
+            "dataType" : "Float16",
+            "formattedType" : "MultiArray (Float16 1 × 1 × 3072)",
+            "shortDescription" : "",
+            "shape" : "[1, 1, 3072]",
+            "name" : "output_hidden_states",
+            "type" : "MultiArray"
+          }
+        ],
+        "name" : "infer",
+        "mlProgramOperationTypeHistogram" : {
+          "Ios18.expandDims" : 69,
+          "Ios18.mul" : 194,
+          "Ios18.cos" : 1,
+          "Ios18.softmax" : 16,
+          "Ios18.matmul" : 33,
+          "Ios18.rsqrt" : 33,
+          "Identity" : 2,
+          "Ios16.reduceMean" : 33,
+          "Ios18.sin" : 1,
+          "Split" : 16,
+          "Select" : 1,
+          "Tile" : 32,
+          "Ios18.add" : 113,
+          "Ios16.fillLike" : 1,
+          "Ios18.sliceUpdate" : 32,
+          "Ios18.reshape" : 99,
+          "Ios18.constexprLutToDense" : 97,
+          "Ios18.conv" : 96,
+          "Ios18.concat" : 65,
+          "Ios18.cast" : 1,
+          "Ios18.transpose" : 18,
+          "Ios18.pow" : 33,
+          "Ios18.silu" : 16,
+          "Ios18.lessEqual" : 1,
+          "Ios18.sliceByIndex" : 160,
+          "Ios18.squeeze" : 16
+        }
+      },
+      {
+        "inputSchema" : [
+          {
+            "hasShapeFlexibility" : "0",
+            "isOptional" : "0",
+            "dataType" : "Float16",
+            "formattedType" : "MultiArray (Float16 1 × 64 × 3072)",
+            "shortDescription" : "",
+            "shape" : "[1, 64, 3072]",
+            "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 × 2048)",
+            "shortDescription" : "",
+            "shape" : "[1, 1, 64, 2048]",
+            "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, Float32, Int32)",
+        "storagePrecision" : "Mixed (Float16, Palettized (11 bits), Palettized (13 bits), Palettized (15 bits), Palettized (7 bits), UInt4)",
+        "stateSchema" : [
+
+        ],
+        "outputSchema" : [
+          {
+            "hasShapeFlexibility" : "0",
+            "isOptional" : "0",
+            "dataType" : "Float16",
+            "formattedType" : "MultiArray (Float16 1 × 64 × 3072)",
+            "shortDescription" : "",
+            "shape" : "[1, 64, 3072]",
+            "name" : "output_hidden_states",
+            "type" : "MultiArray"
+          }
+        ],
+        "name" : "prefill",
+        "mlProgramOperationTypeHistogram" : {
+          "Ios18.expandDims" : 67,
+          "Ios18.softmax" : 16,
+          "Ios18.cos" : 1,
+          "Ios18.mul" : 192,
+          "Ios18.matmul" : 33,
+          "Ios18.rsqrt" : 32,
+          "Identity" : 2,
+          "Ios16.reduceMean" : 32,
+          "Ios18.sin" : 1,
+          "Split" : 16,
+          "Tile" : 32,
+          "Ios18.add" : 112,
+          "Ios18.sliceUpdate" : 32,
+          "Ios18.reshape" : 98,
+          "Ios18.constexprLutToDense" : 97,
+          "Ios18.conv" : 96,
+          "Ios18.concat" : 65,
+          "Ios18.cast" : 1,
+          "Ios18.transpose" : 80,
+          "Ios18.pow" : 32,
+          "Ios18.silu" : 16,
+          "Ios18.sliceByIndex" : 160,
+          "Ios18.squeeze" : 16
+        }
+      }
+    ],
+    "version" : "0.3.4",
+    "isUpdatable" : "0",
+    "defaultFunctionName" : "infer",
+    "specificationVersion" : 9,
+    "stateSchema" : [
+
+    ],
+    "computePrecision" : "Mixed (Float16, Float32, Int32)",
+    "mlProgramOperationTypeHistogram" : {
+      "Ios18.expandDims" : 69,
+      "Ios18.mul" : 194,
+      "Ios18.cos" : 1,
+      "Ios18.softmax" : 16,
+      "Ios18.matmul" : 33,
+      "Ios18.rsqrt" : 33,
+      "Identity" : 2,
+      "Ios16.reduceMean" : 33,
+      "Ios18.sin" : 1,
+      "Split" : 16,
+      "Select" : 1,
+      "Tile" : 32,
+      "Ios18.add" : 113,
+      "Ios16.fillLike" : 1,
+      "Ios18.sliceUpdate" : 32,
+      "Ios18.reshape" : 99,
+      "Ios18.constexprLutToDense" : 97,
+      "Ios18.conv" : 96,
+      "Ios18.concat" : 65,
+      "Ios18.cast" : 1,
+      "Ios18.transpose" : 18,
+      "Ios18.pow" : 33,
+      "Ios18.silu" : 16,
+      "Ios18.lessEqual" : 1,
+      "Ios18.sliceByIndex" : 160,
+      "Ios18.squeeze" : 16
+    },
+    "shortDescription" : "Anemll Model: Multifunction Combined",
+    "generatedClassName" : "phi4_FFN_PF_lut4_chunk_01of02",
+    "author" : "Converted with Anemll v0.3.4",
+    "modelType" : {
+      "name" : "MLModelType_mlProgram"
+    }
+  }
+]

phi4_FFN_PF_lut4_chunk_01of02.mlmodelc/weights/weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c9b1a80fd917817620853cba6084cc50181eb31bebd4b51c03242c3a60714cfa
+size 942252864

phi4_FFN_PF_lut4_chunk_01of02.mlpackage/Data/com.apple.CoreML/model.mlmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9ab2caaac8f312247b29b5373e405721bb43fc2afd9981a0d0969c8aeb367071
+size 946777

phi4_FFN_PF_lut4_chunk_01of02.mlpackage/Data/com.apple.CoreML/weights/weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c9b1a80fd917817620853cba6084cc50181eb31bebd4b51c03242c3a60714cfa
+size 942252864

phi4_FFN_PF_lut4_chunk_01of02.mlpackage/Manifest.json

@@ -0,0 +1,18 @@
+{
+    "fileFormatVersion": "1.0.0",
+    "itemInfoEntries": {
+        "3ED36C90-09E7-454E-993A-613AD9D0DBA7": {
+            "author": "com.apple.CoreML",
+            "description": "CoreML Model Specification",
+            "name": "model.mlmodel",
+            "path": "com.apple.CoreML/model.mlmodel"
+        },
+        "56CB52E7-E373-4A88-A618-FABF77025DEB": {
+            "author": "com.apple.CoreML",
+            "description": "CoreML Model Weights",
+            "name": "weights",
+            "path": "com.apple.CoreML/weights"
+        }
+    },
+    "rootModelIdentifier": "3ED36C90-09E7-454E-993A-613AD9D0DBA7"
+}

phi4_FFN_PF_lut4_chunk_02of02.mlmodelc/analytics/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d459fd207688db500187a3718f7e766ec41fcd451e5e5c621a05d084d401385a
+size 243

phi4_FFN_PF_lut4_chunk_02of02.mlmodelc/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:572b886153dff5be6f240438c1f30e1c5effc7454ae5f9915baa30719bbe7fe7
+size 743

phi4_FFN_PF_lut4_chunk_02of02.mlmodelc/metadata.json

@@ -0,0 +1,299 @@
+[
+  {
+    "metadataOutputVersion" : "3.0",
+    "userDefinedMetadata" : {
+      "com.github.apple.coremltools.version" : "9.0b1",
+      "com.github.apple.coremltools.source" : "torch==2.8.0",
+      "com.anemll.info" : "Converted with Anemll v0.3.4",
+      "com.github.apple.coremltools.source_dialect" : "TorchScript"
+    },
+    "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 × 3072)",
+        "shortDescription" : "",
+        "shape" : "[1, 1, 3072]",
+        "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 × 2048)",
+        "shortDescription" : "",
+        "shape" : "[1, 1, 1, 2048]",
+        "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 × 3072)",
+        "shortDescription" : "",
+        "shape" : "[1, 1, 3072]",
+        "name" : "output_hidden_states",
+        "type" : "MultiArray"
+      }
+    ],
+    "modelParameters" : [
+
+    ],
+    "storagePrecision" : "Mixed (Float16, Int32, Palettized (11 bits), Palettized (13 bits), Palettized (15 bits), Palettized (7 bits), UInt4)",
+    "method" : "predict",
+    "functions" : [
+      {
+        "inputSchema" : [
+          {
+            "hasShapeFlexibility" : "0",
+            "isOptional" : "0",
+            "dataType" : "Float16",
+            "formattedType" : "MultiArray (Float16 1 × 1 × 3072)",
+            "shortDescription" : "",
+            "shape" : "[1, 1, 3072]",
+            "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 × 2048)",
+            "shortDescription" : "",
+            "shape" : "[1, 1, 1, 2048]",
+            "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, Float32, Int32)",
+        "storagePrecision" : "Mixed (Float16, Int32, Palettized (11 bits), Palettized (13 bits), Palettized (15 bits), Palettized (7 bits), UInt4)",
+        "stateSchema" : [
+
+        ],
+        "outputSchema" : [
+          {
+            "hasShapeFlexibility" : "0",
+            "isOptional" : "0",
+            "dataType" : "Float16",
+            "formattedType" : "MultiArray (Float16 1 × 1 × 3072)",
+            "shortDescription" : "",
+            "shape" : "[1, 1, 3072]",
+            "name" : "output_hidden_states",
+            "type" : "MultiArray"
+          }
+        ],
+        "name" : "infer",
+        "mlProgramOperationTypeHistogram" : {
+          "Ios18.expandDims" : 69,
+          "Ios18.mul" : 194,
+          "Ios18.cos" : 1,
+          "Ios18.softmax" : 16,
+          "Ios18.matmul" : 33,
+          "Ios18.rsqrt" : 33,
+          "Identity" : 2,
+          "Ios16.reduceMean" : 33,
+          "Ios18.sin" : 1,
+          "Split" : 16,
+          "Select" : 1,
+          "Tile" : 32,
+          "Ios18.add" : 113,
+          "Ios16.fillLike" : 1,
+          "Ios18.sliceUpdate" : 32,
+          "Ios18.reshape" : 99,
+          "Ios18.constexprLutToDense" : 97,
+          "Ios18.conv" : 96,
+          "Ios18.concat" : 65,
+          "Ios18.cast" : 1,
+          "Ios18.transpose" : 18,
+          "Ios18.pow" : 33,
+          "Ios18.silu" : 16,
+          "Ios18.lessEqual" : 1,
+          "Ios18.sliceByIndex" : 160,
+          "Ios18.squeeze" : 16
+        }
+      },
+      {
+        "inputSchema" : [
+          {
+            "hasShapeFlexibility" : "0",
+            "isOptional" : "0",
+            "dataType" : "Float16",
+            "formattedType" : "MultiArray (Float16 1 × 64 × 3072)",
+            "shortDescription" : "",
+            "shape" : "[1, 64, 3072]",
+            "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 × 2048)",
+            "shortDescription" : "",
+            "shape" : "[1, 1, 64, 2048]",
+            "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, Float32, Int32)",
+        "storagePrecision" : "Mixed (Float16, Palettized (11 bits), Palettized (13 bits), Palettized (15 bits), Palettized (7 bits), UInt4)",
+        "stateSchema" : [
+
+        ],
+        "outputSchema" : [
+          {
+            "hasShapeFlexibility" : "0",
+            "isOptional" : "0",
+            "dataType" : "Float16",
+            "formattedType" : "MultiArray (Float16 1 × 64 × 3072)",
+            "shortDescription" : "",
+            "shape" : "[1, 64, 3072]",
+            "name" : "output_hidden_states",
+            "type" : "MultiArray"
+          }
+        ],
+        "name" : "prefill",
+        "mlProgramOperationTypeHistogram" : {
+          "Ios18.expandDims" : 67,
+          "Ios18.softmax" : 16,
+          "Ios18.cos" : 1,
+          "Ios18.mul" : 192,
+          "Ios18.matmul" : 33,
+          "Ios18.rsqrt" : 32,
+          "Identity" : 2,
+          "Ios16.reduceMean" : 32,
+          "Ios18.sin" : 1,
+          "Split" : 16,
+          "Tile" : 32,
+          "Ios18.add" : 112,
+          "Ios18.sliceUpdate" : 32,
+          "Ios18.reshape" : 98,
+          "Ios18.constexprLutToDense" : 97,
+          "Ios18.conv" : 96,
+          "Ios18.concat" : 65,
+          "Ios18.cast" : 1,
+          "Ios18.transpose" : 80,
+          "Ios18.pow" : 32,
+          "Ios18.silu" : 16,
+          "Ios18.sliceByIndex" : 160,
+          "Ios18.squeeze" : 16
+        }
+      }
+    ],
+    "version" : "0.3.4",
+    "isUpdatable" : "0",
+    "defaultFunctionName" : "infer",
+    "specificationVersion" : 9,
+    "stateSchema" : [
+
+    ],
+    "computePrecision" : "Mixed (Float16, Float32, Int32)",
+    "mlProgramOperationTypeHistogram" : {
+      "Ios18.expandDims" : 69,
+      "Ios18.mul" : 194,
+      "Ios18.cos" : 1,
+      "Ios18.softmax" : 16,
+      "Ios18.matmul" : 33,
+      "Ios18.rsqrt" : 33,
+      "Identity" : 2,
+      "Ios16.reduceMean" : 33,
+      "Ios18.sin" : 1,
+      "Split" : 16,
+      "Select" : 1,
+      "Tile" : 32,
+      "Ios18.add" : 113,
+      "Ios16.fillLike" : 1,
+      "Ios18.sliceUpdate" : 32,
+      "Ios18.reshape" : 99,
+      "Ios18.constexprLutToDense" : 97,
+      "Ios18.conv" : 96,
+      "Ios18.concat" : 65,
+      "Ios18.cast" : 1,
+      "Ios18.transpose" : 18,
+      "Ios18.pow" : 33,
+      "Ios18.silu" : 16,
+      "Ios18.lessEqual" : 1,
+      "Ios18.sliceByIndex" : 160,
+      "Ios18.squeeze" : 16
+    },
+    "shortDescription" : "Anemll Model: Multifunction Combined",
+    "generatedClassName" : "phi4_FFN_PF_lut4_chunk_02of02",
+    "author" : "Converted with Anemll v0.3.4",
+    "modelType" : {
+      "name" : "MLModelType_mlProgram"
+    }
+  }
+]

phi4_FFN_PF_lut4_chunk_02of02.mlmodelc/weights/weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c7cddf51a23eccf8537aff0f49164d219a60b12dcfabab1a51b297eed9104d0d
+size 942252864

phi4_FFN_PF_lut4_chunk_02of02.mlpackage/Data/com.apple.CoreML/model.mlmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0deac5f27f9d44bdea9ad7c01884078b050749a8a146535c70762679ac9f04de
+size 947140

phi4_FFN_PF_lut4_chunk_02of02.mlpackage/Data/com.apple.CoreML/weights/weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c7cddf51a23eccf8537aff0f49164d219a60b12dcfabab1a51b297eed9104d0d
+size 942252864

phi4_FFN_PF_lut4_chunk_02of02.mlpackage/Manifest.json

@@ -0,0 +1,18 @@
+{
+    "fileFormatVersion": "1.0.0",
+    "itemInfoEntries": {
+        "30CD2E36-07C5-46C1-A067-C7AF3DF75B14": {
+            "author": "com.apple.CoreML",
+            "description": "CoreML Model Specification",
+            "name": "model.mlmodel",
+            "path": "com.apple.CoreML/model.mlmodel"
+        },
+        "4B6D69A1-DA28-4E09-B239-28E2E3CAD832": {
+            "author": "com.apple.CoreML",
+            "description": "CoreML Model Weights",
+            "name": "weights",
+            "path": "com.apple.CoreML/weights"
+        }
+    },
+    "rootModelIdentifier": "30CD2E36-07C5-46C1-A067-C7AF3DF75B14"
+}

phi4_FFN_lut4_chunk_01of02.mlpackage/Data/com.apple.CoreML/model.mlmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:586eb631c94781bc7fb29b0dba9f94aed96c38cc9886abc87e283e62fb597837
+size 465885

phi4_FFN_lut4_chunk_01of02.mlpackage/Data/com.apple.CoreML/weights/weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7ce2dab2f6ca3968850368574439596cbc90af797aa8b887ad7bf556c3975c27
+size 941536064

phi4_FFN_lut4_chunk_01of02.mlpackage/Manifest.json

@@ -0,0 +1,18 @@
+{
+    "fileFormatVersion": "1.0.0",
+    "itemInfoEntries": {
+        "10C45FAA-C192-4B04-A068-94CC88D24529": {
+            "author": "com.apple.CoreML",
+            "description": "CoreML Model Specification",
+            "name": "model.mlmodel",
+            "path": "com.apple.CoreML/model.mlmodel"
+        },
+        "CD944379-EDFB-49EF-BF36-AD1F74C8E637": {
+            "author": "com.apple.CoreML",
+            "description": "CoreML Model Weights",
+            "name": "weights",
+            "path": "com.apple.CoreML/weights"
+        }
+    },
+    "rootModelIdentifier": "10C45FAA-C192-4B04-A068-94CC88D24529"
+}

phi4_FFN_lut4_chunk_02of02.mlpackage/Data/com.apple.CoreML/model.mlmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:01da65b3bc03491a84d0f48d5de5ad0355492b8d176244ce4fcf65f5b128ebe4
+size 466038

phi4_FFN_lut4_chunk_02of02.mlpackage/Data/com.apple.CoreML/weights/weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f4b93c86e74522162838912ec96960a71fb56953b9a9cc173fe9d2f02b93938e
+size 941536064

phi4_FFN_lut4_chunk_02of02.mlpackage/Manifest.json

@@ -0,0 +1,18 @@
+{
+    "fileFormatVersion": "1.0.0",
+    "itemInfoEntries": {
+        "07A9AE1F-BD5F-43F9-9786-BF4AD9AA7067": {
+            "author": "com.apple.CoreML",
+            "description": "CoreML Model Specification",
+            "name": "model.mlmodel",
+            "path": "com.apple.CoreML/model.mlmodel"
+        },
+        "9335DBFF-822F-4875-917B-186684CE3051": {
+            "author": "com.apple.CoreML",
+            "description": "CoreML Model Weights",
+            "name": "weights",
+            "path": "com.apple.CoreML/weights"
+        }
+    },
+    "rootModelIdentifier": "07A9AE1F-BD5F-43F9-9786-BF4AD9AA7067"
+}

phi4_embeddings_lut8.mlmodelc/analytics/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d3c9e2c0ee2c9a3a599d109ae298f4474e53fba48d154a57b4a84513f70c6f0a
+size 243

phi4_embeddings_lut8.mlmodelc/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cc923709421bf358759efcb64df7f19bf8e30bcfbf8aa0dc878c7fa229cf4b6c
+size 383

phi4_embeddings_lut8.mlmodelc/metadata.json

@@ -0,0 +1,65 @@
+[
+  {
+    "metadataOutputVersion" : "3.0",
+    "storagePrecision" : "Mixed (Float16, Palettized (23 bits), UInt8)",
+    "outputSchema" : [
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 1 × 3072)",
+        "shortDescription" : "",
+        "shape" : "[1, 1, 3072]",
+        "name" : "hidden_states",
+        "type" : "MultiArray"
+      }
+    ],
+    "modelParameters" : [
+
+    ],
+    "specificationVersion" : 9,
+    "mlProgramOperationTypeHistogram" : {
+      "Ios18.greaterEqual" : 2,
+      "Ios18.constexprLutToDense" : 1,
+      "Ios18.add" : 2,
+      "Select" : 2,
+      "Ios18.gather" : 1
+    },
+    "computePrecision" : "Mixed (Float16, Int32)",
+    "isUpdatable" : "0",
+    "stateSchema" : [
+
+    ],
+    "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" : "2025-09-28",
+      "com.github.apple.coremltools.source" : "torch==2.8.0",
+      "com.github.apple.coremltools.version" : "9.0b1",
+      "com.github.apple.coremltools.source_dialect" : "TorchScript"
+    },
+    "inputSchema" : [
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Int32",
+        "formattedType" : "MultiArray (Int32 1 × 1)",
+        "shortDescription" : "",
+        "shape" : "[1, 1]",
+        "name" : "input_ids",
+        "type" : "MultiArray"
+      }
+    ],
+    "generatedClassName" : "phi4_embeddings_lut8",
+    "method" : "predict"
+  }
+]

phi4_embeddings_lut8.mlmodelc/model.mil

@@ -0,0 +1,21 @@
+program(1.3)
+[buildInfo = dict<string, string>({{"coremlc-component-MIL", "3500.14.1"}, {"coremlc-version", "3500.32.1"}})]
+{
+    func main<ios18>(tensor<int32, [1, 1]> input_ids) {
+            int32 var_6_batch_dims_0 = const()[name = string("op_6_batch_dims_0"), val = int32(0)];
+            bool var_6_validate_indices_0 = const()[name = string("op_6_validate_indices_0"), val = bool(false)];
+            tensor<fp16, [200064, 3072]> embed_weight_to_fp16_palettized = constexpr_lut_to_dense(indices = tensor<uint8, [200064, 3072]>(BLOBFILE(path = string("@model_path/weights/weight.bin"), offset = uint64(64))), lut = tensor<fp16, [25008, 1, 256, 1]>(BLOBFILE(path = string("@model_path/weights/weight.bin"), offset = uint64(614596736))))[name = string("embed_weight_to_fp16_palettized")];
+            int32 greater_equal_0_y_0 = const()[name = string("greater_equal_0_y_0"), val = int32(0)];
+            tensor<bool, [1, 1]> greater_equal_0 = greater_equal(x = input_ids, y = greater_equal_0_y_0)[name = string("greater_equal_0")];
+            int32 slice_by_index_0 = const()[name = string("slice_by_index_0"), val = int32(200064)];
+            tensor<int32, [1, 1]> add_0 = add(x = input_ids, y = slice_by_index_0)[name = string("add_0")];
+            tensor<int32, [1, 1]> select_0 = select(a = input_ids, b = add_0, cond = greater_equal_0)[name = string("select_0")];
+            int32 greater_equal_0_y_0_1 = const()[name = string("greater_equal_0_y_0_1"), val = int32(0)];
+            tensor<bool, [1, 1]> greater_equal_0_1 = greater_equal(x = select_0, y = greater_equal_0_y_0_1)[name = string("greater_equal_0_1")];
+            int32 slice_by_index_0_1 = const()[name = string("slice_by_index_0_1"), val = int32(200064)];
+            tensor<int32, [1, 1]> add_0_1 = add(x = select_0, y = slice_by_index_0_1)[name = string("add_0_1")];
+            tensor<int32, [1, 1]> select_0_1 = select(a = select_0, b = add_0_1, cond = greater_equal_0_1)[name = string("select_0_1")];
+            int32 op_6_cast_fp16_axis_0 = const()[name = string("op_6_cast_fp16_axis_0"), val = int32(0)];
+            tensor<fp16, [1, 1, 3072]> hidden_states = gather(axis = op_6_cast_fp16_axis_0, batch_dims = var_6_batch_dims_0, indices = select_0_1, validate_indices = var_6_validate_indices_0, x = embed_weight_to_fp16_palettized)[name = string("op_6_cast_fp16")];
+        } -> (hidden_states);
+}

phi4_embeddings_lut8.mlmodelc/weights/weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:471ccac8a8d66e8f8a024ce8b98f037dd31fa3bac450fdb53f1c8e3308ee011a
+size 627400896

phi4_embeddings_lut8.mlpackage/Data/com.apple.CoreML/model.mlmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:92af5fae585670c8e37cafbf73971a02bd75b8c102ec6ab9af0aaa723aa672b0
+size 2526

phi4_embeddings_lut8.mlpackage/Data/com.apple.CoreML/weights/weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:471ccac8a8d66e8f8a024ce8b98f037dd31fa3bac450fdb53f1c8e3308ee011a
+size 627400896

phi4_embeddings_lut8.mlpackage/Manifest.json

@@ -0,0 +1,18 @@
+{
+    "fileFormatVersion": "1.0.0",
+    "itemInfoEntries": {
+        "61B24E7A-D867-4391-A38F-50F28BF7D5FD": {
+            "author": "com.apple.CoreML",
+            "description": "CoreML Model Specification",
+            "name": "model.mlmodel",
+            "path": "com.apple.CoreML/model.mlmodel"
+        },
+        "7BCAECDE-96CA-47ED-ABB9-3F6559A9D4F4": {
+            "author": "com.apple.CoreML",
+            "description": "CoreML Model Weights",
+            "name": "weights",
+            "path": "com.apple.CoreML/weights"
+        }
+    },
+    "rootModelIdentifier": "61B24E7A-D867-4391-A38F-50F28BF7D5FD"
+}

phi4_lm_head_lut6.mlmodelc/analytics/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4d17ebd64d18ae52309635d5106ef6270ad518ededa11856b6c50bdc201c68e2
+size 243

phi4_lm_head_lut6.mlmodelc/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:22a2a16ffd4f1e687322b91095caa994c6a27c2b1616dc92ec3f3f5a429895c1
+size 383

phi4_lm_head_lut6.mlmodelc/metadata.json

@@ -0,0 +1,63 @@
+[
+  {
+    "metadataOutputVersion" : "3.0",
+    "storagePrecision" : "Mixed (Float16, Palettized (21 bits), UInt6)",
+    "outputSchema" : [
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 1 × 200064)",
+        "shortDescription" : "",
+        "shape" : "[1, 1, 200064]",
+        "name" : "logits",
+        "type" : "MultiArray"
+      }
+    ],
+    "modelParameters" : [
+
+    ],
+    "specificationVersion" : 9,
+    "mlProgramOperationTypeHistogram" : {
+      "Ios18.constexprLutToDense" : 1,
+      "Ios18.transpose" : 2,
+      "Ios18.conv" : 1
+    },
+    "computePrecision" : "Mixed (Float16, Int32)",
+    "isUpdatable" : "0",
+    "stateSchema" : [
+
+    ],
+    "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" : "2025-09-28",
+      "com.github.apple.coremltools.source" : "torch==2.8.0",
+      "com.github.apple.coremltools.version" : "9.0b1",
+      "com.github.apple.coremltools.source_dialect" : "TorchScript"
+    },
+    "inputSchema" : [
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 1 × 3072)",
+        "shortDescription" : "",
+        "shape" : "[1, 1, 3072]",
+        "name" : "hidden_states",
+        "type" : "MultiArray"
+      }
+    ],
+    "generatedClassName" : "phi4_lm_head_lut6",
+    "method" : "predict"
+  }
+]

phi4_lm_head_lut6.mlmodelc/model.mil

@@ -0,0 +1,17 @@
+program(1.3)
+[buildInfo = dict<string, string>({{"coremlc-component-MIL", "3500.14.1"}, {"coremlc-version", "3500.32.1"}})]
+{
+    func main<ios18>(tensor<fp16, [1, 1, 3072]> hidden_states) {
+            tensor<int32, [3]> var_6 = const()[name = string("op_6"), val = tensor<int32, [3]>([0, 2, 1])];
+            string var_22_pad_type_0 = const()[name = string("op_22_pad_type_0"), val = string("valid")];
+            int32 var_22_groups_0 = const()[name = string("op_22_groups_0"), val = int32(1)];
+            tensor<int32, [1]> var_22_strides_0 = const()[name = string("op_22_strides_0"), val = tensor<int32, [1]>([1])];
+            tensor<int32, [2]> var_22_pad_0 = const()[name = string("op_22_pad_0"), val = tensor<int32, [2]>([0, 0])];
+            tensor<int32, [1]> var_22_dilations_0 = const()[name = string("op_22_dilations_0"), val = tensor<int32, [1]>([1])];
+            tensor<fp16, [200064, 3072, 1]> squeeze_0_cast_fp16_to_fp32_to_fp16_palettized = constexpr_lut_to_dense(indices = tensor<uint6, [200064, 3072, 1]>(BLOBFILE(path = string("@model_path/weights/weight.bin"), offset = uint64(64))), lut = tensor<fp16, [25008, 1, 1, 64, 1]>(BLOBFILE(path = string("@model_path/weights/weight.bin"), offset = uint64(460947584))))[name = string("squeeze_0_cast_fp16_to_fp32_to_fp16_palettized")];
+            tensor<fp16, [1, 3072, 1]> var_7_cast_fp16 = transpose(perm = var_6, x = hidden_states)[name = string("transpose_1")];
+            tensor<fp16, [1, 200064, 1]> var_22_cast_fp16 = conv(dilations = var_22_dilations_0, groups = var_22_groups_0, pad = var_22_pad_0, pad_type = var_22_pad_type_0, strides = var_22_strides_0, weight = squeeze_0_cast_fp16_to_fp32_to_fp16_palettized, x = var_7_cast_fp16)[name = string("op_22_cast_fp16")];
+            tensor<int32, [3]> var_26 = const()[name = string("op_26"), val = tensor<int32, [3]>([0, 2, 1])];
+            tensor<fp16, [1, 1, 200064]> logits = transpose(perm = var_26, x = var_22_cast_fp16)[name = string("transpose_0")];
+        } -> (logits);
+}

phi4_lm_head_lut6.mlmodelc/weights/weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d1620b615ef8cd96a1087ca6fba635eebcc22398c4174ad45ef18dde785d365a
+size 464148672

phi4_lm_head_lut6.mlpackage/Data/com.apple.CoreML/model.mlmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7e7803878a76f9ada016ee6d8016448971188293a7ae6694eedc1d4ef86fb583
+size 2080

phi4_lm_head_lut6.mlpackage/Data/com.apple.CoreML/weights/weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d1620b615ef8cd96a1087ca6fba635eebcc22398c4174ad45ef18dde785d365a
+size 464148672

phi4_lm_head_lut6.mlpackage/Manifest.json

@@ -0,0 +1,18 @@
+{
+    "fileFormatVersion": "1.0.0",
+    "itemInfoEntries": {
+        "3528B6F8-359A-47AB-9099-3E9C4121A65A": {
+            "author": "com.apple.CoreML",
+            "description": "CoreML Model Weights",
+            "name": "weights",
+            "path": "com.apple.CoreML/weights"
+        },
+        "5C3068C5-09B3-469D-9E30-D89DCD3242BA": {
+            "author": "com.apple.CoreML",
+            "description": "CoreML Model Specification",
+            "name": "model.mlmodel",
+            "path": "com.apple.CoreML/model.mlmodel"
+        }
+    },
+    "rootModelIdentifier": "5C3068C5-09B3-469D-9E30-D89DCD3242BA"
+}

phi4_prefill_lut4_chunk_01of02.mlpackage/Data/com.apple.CoreML/model.mlmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1c7dffe2693f54d774fd943549a8029ebf7997991bb633be036c3761b5e97f3f
+size 481032

phi4_prefill_lut4_chunk_01of02.mlpackage/Data/com.apple.CoreML/weights/weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ea0e5d1ec01f7d0a8c55c98384c211456ab42d35799ed274e3787bb86cf1d1b8
+size 942029504

phi4_prefill_lut4_chunk_01of02.mlpackage/Manifest.json

@@ -0,0 +1,18 @@
+{
+    "fileFormatVersion": "1.0.0",
+    "itemInfoEntries": {
+        "CF8A09C1-D506-45BD-80A3-4E36C2529334": {
+            "author": "com.apple.CoreML",
+            "description": "CoreML Model Weights",
+            "name": "weights",
+            "path": "com.apple.CoreML/weights"
+        },
+        "DA5F8CFD-5833-43DD-A9BD-0981D832D125": {
+            "author": "com.apple.CoreML",
+            "description": "CoreML Model Specification",
+            "name": "model.mlmodel",
+            "path": "com.apple.CoreML/model.mlmodel"
+        }
+    },
+    "rootModelIdentifier": "DA5F8CFD-5833-43DD-A9BD-0981D832D125"
+}

phi4_prefill_lut4_chunk_02of02.mlpackage/Data/com.apple.CoreML/model.mlmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1e0b8e9e132180bed768bc88e6aaa4a2056a0a6506815caf8d1090b1e6becd1f
+size 481242

phi4_prefill_lut4_chunk_02of02.mlpackage/Data/com.apple.CoreML/weights/weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7344a9ff4bcc5c1e124c7e110973d957cd6ddeafc603e9701b684076678969d4
+size 942029504

phi4_prefill_lut4_chunk_02of02.mlpackage/Manifest.json

@@ -0,0 +1,18 @@
+{
+    "fileFormatVersion": "1.0.0",
+    "itemInfoEntries": {
+        "05D6FC91-3A2E-4B17-975D-AF311B32C225": {
+            "author": "com.apple.CoreML",
+            "description": "CoreML Model Specification",
+            "name": "model.mlmodel",
+            "path": "com.apple.CoreML/model.mlmodel"
+        },
+        "A203C972-22D7-4829-B187-257027B2E637": {
+            "author": "com.apple.CoreML",
+            "description": "CoreML Model Weights",
+            "name": "weights",
+            "path": "com.apple.CoreML/weights"
+        }
+    },
+    "rootModelIdentifier": "05D6FC91-3A2E-4B17-975D-AF311B32C225"
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:382cc235b56c725945e149cc25f191da667c836655efd0857b004320e90e91ea
+size 15524095