Add models

Llama-3.2-1B-Instruct_chunk1.mlmodelc/analytics/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c1fdd81f7aa992caa1f946bcb46d36fce0df7cd32f7ad9ca046e84e0bcfa9fa7
+size 243

Llama-3.2-1B-Instruct_chunk1.mlmodelc/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:029f313416e6f6e57a6e277e61397fb98e2c1b3c88aaf17c11507cb80f63b720
+size 407

Llama-3.2-1B-Instruct_chunk1.mlmodelc/metadata.json

@@ -0,0 +1,105 @@
+[
+  {
+    "metadataOutputVersion" : "3.0",
+    "storagePrecision" : "Float16",
+    "outputSchema" : [
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 2048 × 8 × 8)",
+        "shortDescription" : "",
+        "shape" : "[1, 2048, 8, 8]",
+        "name" : "x",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 64 × 64)",
+        "shortDescription" : "",
+        "shape" : "[64, 64]",
+        "name" : "cos",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 64 × 64)",
+        "shortDescription" : "",
+        "shape" : "[64, 64]",
+        "name" : "sin",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 64)",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 64]",
+        "name" : "mask",
+        "type" : "MultiArray"
+      }
+    ],
+    "modelParameters" : [
+
+    ],
+    "specificationVersion" : 7,
+    "mlProgramOperationTypeHistogram" : {
+      "Select" : 2,
+      "Tile" : 2,
+      "Ios16.sub" : 3,
+      "Transpose" : 2,
+      "Ios16.gather" : 3,
+      "ExpandDims" : 3,
+      "Ios16.reshape" : 1,
+      "Ios16.maximum" : 1,
+      "Ios16.less" : 2
+    },
+    "computePrecision" : "Mixed (Float16, Int32)",
+    "isUpdatable" : "0",
+    "availability" : {
+      "macOS" : "13.0",
+      "tvOS" : "16.0",
+      "visionOS" : "1.0",
+      "watchOS" : "9.0",
+      "iOS" : "16.0",
+      "macCatalyst" : "16.0"
+    },
+    "modelType" : {
+      "name" : "MLModelType_mlProgram"
+    },
+    "userDefinedMetadata" : {
+      "com.github.apple.coremltools.source_dialect" : "TorchScript",
+      "com.github.apple.coremltools.source" : "torch==2.1.0",
+      "com.github.apple.coremltools.version" : "8.0b1"
+    },
+    "inputSchema" : [
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Int32",
+        "formattedType" : "MultiArray (Int32 1 × 64)",
+        "shortDescription" : "",
+        "shape" : "[1, 64]",
+        "name" : "input_ids",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Int32",
+        "formattedType" : "MultiArray (Int32 1)",
+        "shortDescription" : "",
+        "shape" : "[1]",
+        "name" : "full_sequence_length",
+        "type" : "MultiArray"
+      }
+    ],
+    "generatedClassName" : "Llama_3_2_1B_Instruct_2024_10_10_23_56_41_chunk1",
+    "method" : "predict"
+  }
+]

Llama-3.2-1B-Instruct_chunk1.mlmodelc/model.mil

@@ -0,0 +1,50 @@
+program(1.0)
+[buildInfo = dict<tensor<string, []>, tensor<string, []>>({{"coremlc-component-MIL", "3304.5.2"}, {"coremlc-version", "3304.6.2"}, {"coremltools-component-torch", "2.1.0"}, {"coremltools-source-dialect", "TorchScript"}, {"coremltools-version", "8.0b1"}})]
+{
+    func main<ios16>(tensor<int32, [1]> full_sequence_length, tensor<int32, [1, 64]> input_ids) {
+            tensor<int32, [1]> T = const()[name = tensor<string, []>("T"), val = tensor<int32, [1]>([64])];
+            tensor<int32, []> x_1_axis_0 = const()[name = tensor<string, []>("x_1_axis_0"), val = tensor<int32, []>(0)];
+            tensor<int32, []> x_1_batch_dims_0 = const()[name = tensor<string, []>("x_1_batch_dims_0"), val = tensor<int32, []>(0)];
+            tensor<fp16, [128256, 2048]> wte_weight_to_fp16 = const()[name = tensor<string, []>("wte_weight_to_fp16"), val = tensor<fp16, [128256, 2048]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(64)))];
+            tensor<fp16, [1, 64, 2048]> x_1_cast_fp16 = gather(axis = x_1_axis_0, batch_dims = x_1_batch_dims_0, indices = input_ids, x = wte_weight_to_fp16)[name = tensor<string, []>("x_1_cast_fp16")];
+            tensor<int32, [3]> x_perm_0 = const()[name = tensor<string, []>("x_perm_0"), val = tensor<int32, [3]>([0, 2, 1])];
+            tensor<int32, [4]> var_27 = const()[name = tensor<string, []>("op_27"), val = tensor<int32, [4]>([1, 2048, -1, 8])];
+            tensor<fp16, [1, 2048, 64]> x_cast_fp16 = transpose(perm = x_perm_0, x = x_1_cast_fp16)[name = tensor<string, []>("transpose_1")];
+            tensor<fp16, [1, 2048, 8, 8]> x = reshape(shape = var_27, x = x_cast_fp16)[name = tensor<string, []>("op_28_cast_fp16")];
+            tensor<int32, [1]> pos_offset = sub(x = T, y = full_sequence_length)[name = tensor<string, []>("pos_offset")];
+            tensor<int32, [64]> var_36 = const()[name = tensor<string, []>("op_36"), val = tensor<int32, [64]>([0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63])];
+            tensor<int32, [64]> input_pos_1 = sub(x = var_36, y = pos_offset)[name = tensor<string, []>("input_pos_1")];
+            tensor<int32, [64]> var_44 = const()[name = tensor<string, []>("op_44"), val = tensor<int32, [64]>([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0])];
+            tensor<int32, [64]> input_pos = maximum(x = input_pos_1, y = var_44)[name = tensor<string, []>("input_pos")];
+            tensor<int32, []> var_55 = const()[name = tensor<string, []>("op_55"), val = tensor<int32, []>(1)];
+            tensor<int32, []> cos_batch_dims_0 = const()[name = tensor<string, []>("cos_batch_dims_0"), val = tensor<int32, []>(0)];
+            tensor<fp16, [64, 512]> var_54_to_fp16 = const()[name = tensor<string, []>("op_54_to_fp16"), val = tensor<fp16, [64, 512]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(525336704)))];
+            tensor<fp16, [64, 64]> cos = gather(axis = var_55, batch_dims = cos_batch_dims_0, indices = input_pos, x = var_54_to_fp16)[name = tensor<string, []>("cos_cast_fp16")];
+            tensor<int32, []> var_66 = const()[name = tensor<string, []>("op_66"), val = tensor<int32, []>(1)];
+            tensor<int32, []> sin_batch_dims_0 = const()[name = tensor<string, []>("sin_batch_dims_0"), val = tensor<int32, []>(0)];
+            tensor<fp16, [64, 512]> var_65_to_fp16 = const()[name = tensor<string, []>("op_65_to_fp16"), val = tensor<fp16, [64, 512]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(525402304)))];
+            tensor<fp16, [64, 64]> sin = gather(axis = var_66, batch_dims = sin_batch_dims_0, indices = input_pos, x = var_65_to_fp16)[name = tensor<string, []>("sin_cast_fp16")];
+            tensor<int32, [64, 1]> var_102 = const()[name = tensor<string, []>("op_102"), val = tensor<int32, [64, 1]>([[0], [1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31], [32], [33], [34], [35], [36], [37], [38], [39], [40], [41], [42], [43], [44], [45], [46], [47], [48], [49], [50], [51], [52], [53], [54], [55], [56], [57], [58], [59], [60], [61], [62], [63]])];
+            tensor<bool, [64, 1]> var_105 = less(x = var_102, y = pos_offset)[name = tensor<string, []>("op_105")];
+            tensor<int32, [2]> var_105_after_broadcast_reps_0 = const()[name = tensor<string, []>("op_105_after_broadcast_reps_0"), val = tensor<int32, [2]>([1, 512])];
+            tensor<bool, [64, 512]> var_105_after_broadcast = tile(reps = var_105_after_broadcast_reps_0, x = var_105)[name = tensor<string, []>("op_105_after_broadcast")];
+            tensor<fp16, [64, 512]> all_mask_to_fp16 = const()[name = tensor<string, []>("all_mask_to_fp16"), val = tensor<fp16, [64, 512]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(525467904)))];
+            tensor<fp16, [64, 512]> m_1_to_fp16 = const()[name = tensor<string, []>("m_1_to_fp16"), val = tensor<fp16, [64, 512]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(525533504)))];
+            tensor<fp16, [64, 512]> m_3_cast_fp16 = select(a = all_mask_to_fp16, b = m_1_to_fp16, cond = var_105_after_broadcast)[name = tensor<string, []>("m_3_cast_fp16")];
+            tensor<int32, [512]> var_115 = const()[name = tensor<string, []>("op_115"), val = tensor<int32, [512]>([0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 158, 159, 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 188, 189, 190, 191, 192, 193, 194, 195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212, 213, 214, 215, 216, 217, 218, 219, 220, 221, 222, 223, 224, 225, 226, 227, 228, 229, 230, 231, 232, 233, 234, 235, 236, 237, 238, 239, 240, 241, 242, 243, 244, 245, 246, 247, 248, 249, 250, 251, 252, 253, 254, 255, 256, 257, 258, 259, 260, 261, 262, 263, 264, 265, 266, 267, 268, 269, 270, 271, 272, 273, 274, 275, 276, 277, 278, 279, 280, 281, 282, 283, 284, 285, 286, 287, 288, 289, 290, 291, 292, 293, 294, 295, 296, 297, 298, 299, 300, 301, 302, 303, 304, 305, 306, 307, 308, 309, 310, 311, 312, 313, 314, 315, 316, 317, 318, 319, 320, 321, 322, 323, 324, 325, 326, 327, 328, 329, 330, 331, 332, 333, 334, 335, 336, 337, 338, 339, 340, 341, 342, 343, 344, 345, 346, 347, 348, 349, 350, 351, 352, 353, 354, 355, 356, 357, 358, 359, 360, 361, 362, 363, 364, 365, 366, 367, 368, 369, 370, 371, 372, 373, 374, 375, 376, 377, 378, 379, 380, 381, 382, 383, 384, 385, 386, 387, 388, 389, 390, 391, 392, 393, 394, 395, 396, 397, 398, 399, 400, 401, 402, 403, 404, 405, 406, 407, 408, 409, 410, 411, 412, 413, 414, 415, 416, 417, 418, 419, 420, 421, 422, 423, 424, 425, 426, 427, 428, 429, 430, 431, 432, 433, 434, 435, 436, 437, 438, 439, 440, 441, 442, 443, 444, 445, 446, 447, 448, 449, 450, 451, 452, 453, 454, 455, 456, 457, 458, 459, 460, 461, 462, 463, 464, 465, 466, 467, 468, 469, 470, 471, 472, 473, 474, 475, 476, 477, 478, 479, 480, 481, 482, 483, 484, 485, 486, 487, 488, 489, 490, 491, 492, 493, 494, 495, 496, 497, 498, 499, 500, 501, 502, 503, 504, 505, 506, 507, 508, 509, 510, 511])];
+            tensor<int32, []> var_116 = const()[name = tensor<string, []>("op_116"), val = tensor<int32, []>(512)];
+            tensor<int32, [1]> var_118 = sub(x = var_116, y = full_sequence_length)[name = tensor<string, []>("op_118")];
+            tensor<bool, [512]> var_119 = less(x = var_115, y = var_118)[name = tensor<string, []>("op_119")];
+            tensor<int32, [1]> expand_dims_0_axes_0 = const()[name = tensor<string, []>("expand_dims_0_axes_0"), val = tensor<int32, [1]>([0])];
+            tensor<bool, [1, 512]> expand_dims_0 = expand_dims(axes = expand_dims_0_axes_0, x = var_119)[name = tensor<string, []>("expand_dims_0")];
+            tensor<int32, [2]> var_119_after_broadcast_reps_0 = const()[name = tensor<string, []>("op_119_after_broadcast_reps_0"), val = tensor<int32, [2]>([64, 1])];
+            tensor<bool, [64, 512]> var_119_after_broadcast = tile(reps = var_119_after_broadcast_reps_0, x = expand_dims_0)[name = tensor<string, []>("op_119_after_broadcast")];
+            tensor<fp16, [64, 512]> m_cast_fp16 = select(a = all_mask_to_fp16, b = m_3_cast_fp16, cond = var_119_after_broadcast)[name = tensor<string, []>("m_cast_fp16")];
+            tensor<int32, [1]> var_122_axes_0 = const()[name = tensor<string, []>("op_122_axes_0"), val = tensor<int32, [1]>([0])];
+            tensor<fp16, [1, 64, 512]> var_122_cast_fp16 = expand_dims(axes = var_122_axes_0, x = m_cast_fp16)[name = tensor<string, []>("op_122_cast_fp16")];
+            tensor<int32, [1]> mask_axes_0 = const()[name = tensor<string, []>("mask_axes_0"), val = tensor<int32, [1]>([0])];
+            tensor<fp16, [1, 1, 64, 512]> mask_cast_fp16 = expand_dims(axes = mask_axes_0, x = var_122_cast_fp16)[name = tensor<string, []>("mask_cast_fp16")];
+            tensor<int32, [4]> var_129 = const()[name = tensor<string, []>("op_129"), val = tensor<int32, [4]>([0, 3, 1, 2])];
+            tensor<fp16, [1, 512, 1, 64]> mask = transpose(perm = var_129, x = mask_cast_fp16)[name = tensor<string, []>("transpose_0")];
+        } -> (x, cos, sin, mask);
+}

Llama-3.2-1B-Instruct_chunk1.mlmodelc/weights/weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5080c5c470cc23733c394cdef3dc1da1f940a45cce75d6bf9587d70f7933ae8e
+size 525599104

Llama-3.2-1B-Instruct_chunk2.mlmodelc/analytics/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aa232688f37d1b639e14da601159147837ea3a6d7861e440adc8913fb9b3246f
+size 243

Llama-3.2-1B-Instruct_chunk2.mlmodelc/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bbed3683f2e4450c8c82941040938b8db75f66e3348e5f757e67745654a5b4e4
+size 931

Llama-3.2-1B-Instruct_chunk2.mlmodelc/metadata.json

@@ -0,0 +1,258 @@
+[
+  {
+    "metadataOutputVersion" : "3.0",
+    "storagePrecision" : "Float16",
+    "outputSchema" : [
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 2048 × 8 × 8)",
+        "shortDescription" : "",
+        "shape" : "[1, 2048, 8, 8]",
+        "name" : "new_x",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 64 × 1 × 512)",
+        "shortDescription" : "",
+        "shape" : "[1, 64, 1, 512]",
+        "name" : "new_k_cache_0",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 64)",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 64]",
+        "name" : "new_v_cache_0",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 64 × 1 × 512)",
+        "shortDescription" : "",
+        "shape" : "[1, 64, 1, 512]",
+        "name" : "new_k_cache_1",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 64)",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 64]",
+        "name" : "new_v_cache_1",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 64 × 1 × 512)",
+        "shortDescription" : "",
+        "shape" : "[1, 64, 1, 512]",
+        "name" : "new_k_cache_2",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 64)",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 64]",
+        "name" : "new_v_cache_2",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 64 × 1 × 512)",
+        "shortDescription" : "",
+        "shape" : "[1, 64, 1, 512]",
+        "name" : "new_k_cache_3",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 64)",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 64]",
+        "name" : "new_v_cache_3",
+        "type" : "MultiArray"
+      }
+    ],
+    "modelParameters" : [
+
+    ],
+    "specificationVersion" : 7,
+    "mlProgramOperationTypeHistogram" : {
+      "Concat" : 28,
+      "Ios16.mul" : 172,
+      "SliceByIndex" : 208,
+      "Transpose" : 4,
+      "Ios16.einsum" : 256,
+      "Ios16.conv" : 28,
+      "Ios16.add" : 144,
+      "Ios16.realDiv" : 8,
+      "Ios16.softmax" : 128,
+      "Ios16.reduceL2Norm" : 8,
+      "Ios16.reshape" : 28,
+      "Ios16.silu" : 4
+    },
+    "computePrecision" : "Mixed (Float16, Int32)",
+    "isUpdatable" : "0",
+    "availability" : {
+      "macOS" : "13.0",
+      "tvOS" : "16.0",
+      "visionOS" : "1.0",
+      "watchOS" : "9.0",
+      "iOS" : "16.0",
+      "macCatalyst" : "16.0"
+    },
+    "modelType" : {
+      "name" : "MLModelType_mlProgram"
+    },
+    "userDefinedMetadata" : {
+      "com.github.apple.coremltools.source_dialect" : "TorchScript",
+      "com.github.apple.coremltools.source" : "torch==2.1.0",
+      "com.github.apple.coremltools.version" : "8.0b1"
+    },
+    "inputSchema" : [
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 2048 × 8 × 8)",
+        "shortDescription" : "",
+        "shape" : "[1, 2048, 8, 8]",
+        "name" : "x",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 64 × 64)",
+        "shortDescription" : "",
+        "shape" : "[64, 64]",
+        "name" : "cos",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 64 × 64)",
+        "shortDescription" : "",
+        "shape" : "[64, 64]",
+        "name" : "sin",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 64)",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 64]",
+        "name" : "mask",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 448 × 1 × 512)?",
+        "shortDescription" : "",
+        "shape" : "[1, 448, 1, 512]",
+        "name" : "k_cache_0",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 448)?",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 448]",
+        "name" : "v_cache_0",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 448 × 1 × 512)?",
+        "shortDescription" : "",
+        "shape" : "[1, 448, 1, 512]",
+        "name" : "k_cache_1",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 448)?",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 448]",
+        "name" : "v_cache_1",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 448 × 1 × 512)?",
+        "shortDescription" : "",
+        "shape" : "[1, 448, 1, 512]",
+        "name" : "k_cache_2",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 448)?",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 448]",
+        "name" : "v_cache_2",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 448 × 1 × 512)?",
+        "shortDescription" : "",
+        "shape" : "[1, 448, 1, 512]",
+        "name" : "k_cache_3",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 448)?",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 448]",
+        "name" : "v_cache_3",
+        "type" : "MultiArray"
+      }
+    ],
+    "generatedClassName" : "Llama_3_2_1B_Instruct_2024_10_10_23_56_41_chunk2",
+    "method" : "predict"
+  }
+]

Llama-3.2-1B-Instruct_chunk2.mlmodelc/weights/weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:23c3f4ed69a73c80f959d0e977ff7677c380a3e8f06a8af84f2003f5cdd81c5b
+size 486575936

Llama-3.2-1B-Instruct_chunk3.mlmodelc/analytics/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aa232688f37d1b639e14da601159147837ea3a6d7861e440adc8913fb9b3246f
+size 243

Llama-3.2-1B-Instruct_chunk3.mlmodelc/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:507886ccf77d43d177e993f226218bda5b20448ffeb13ac2e9f6a768ffb75b70
+size 931

Llama-3.2-1B-Instruct_chunk3.mlmodelc/metadata.json

@@ -0,0 +1,258 @@
+[
+  {
+    "metadataOutputVersion" : "3.0",
+    "storagePrecision" : "Float16",
+    "outputSchema" : [
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 2048 × 8 × 8)",
+        "shortDescription" : "",
+        "shape" : "[1, 2048, 8, 8]",
+        "name" : "new_x",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 64 × 1 × 512)",
+        "shortDescription" : "",
+        "shape" : "[1, 64, 1, 512]",
+        "name" : "new_k_cache_0",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 64)",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 64]",
+        "name" : "new_v_cache_0",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 64 × 1 × 512)",
+        "shortDescription" : "",
+        "shape" : "[1, 64, 1, 512]",
+        "name" : "new_k_cache_1",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 64)",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 64]",
+        "name" : "new_v_cache_1",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 64 × 1 × 512)",
+        "shortDescription" : "",
+        "shape" : "[1, 64, 1, 512]",
+        "name" : "new_k_cache_2",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 64)",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 64]",
+        "name" : "new_v_cache_2",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 64 × 1 × 512)",
+        "shortDescription" : "",
+        "shape" : "[1, 64, 1, 512]",
+        "name" : "new_k_cache_3",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 64)",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 64]",
+        "name" : "new_v_cache_3",
+        "type" : "MultiArray"
+      }
+    ],
+    "modelParameters" : [
+
+    ],
+    "specificationVersion" : 7,
+    "mlProgramOperationTypeHistogram" : {
+      "Concat" : 28,
+      "Ios16.mul" : 172,
+      "SliceByIndex" : 208,
+      "Transpose" : 4,
+      "Ios16.einsum" : 256,
+      "Ios16.conv" : 28,
+      "Ios16.add" : 144,
+      "Ios16.realDiv" : 8,
+      "Ios16.softmax" : 128,
+      "Ios16.reduceL2Norm" : 8,
+      "Ios16.reshape" : 28,
+      "Ios16.silu" : 4
+    },
+    "computePrecision" : "Mixed (Float16, Int32)",
+    "isUpdatable" : "0",
+    "availability" : {
+      "macOS" : "13.0",
+      "tvOS" : "16.0",
+      "visionOS" : "1.0",
+      "watchOS" : "9.0",
+      "iOS" : "16.0",
+      "macCatalyst" : "16.0"
+    },
+    "modelType" : {
+      "name" : "MLModelType_mlProgram"
+    },
+    "userDefinedMetadata" : {
+      "com.github.apple.coremltools.source_dialect" : "TorchScript",
+      "com.github.apple.coremltools.source" : "torch==2.1.0",
+      "com.github.apple.coremltools.version" : "8.0b1"
+    },
+    "inputSchema" : [
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 2048 × 8 × 8)",
+        "shortDescription" : "",
+        "shape" : "[1, 2048, 8, 8]",
+        "name" : "x",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 64 × 64)",
+        "shortDescription" : "",
+        "shape" : "[64, 64]",
+        "name" : "cos",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 64 × 64)",
+        "shortDescription" : "",
+        "shape" : "[64, 64]",
+        "name" : "sin",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 64)",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 64]",
+        "name" : "mask",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 448 × 1 × 512)?",
+        "shortDescription" : "",
+        "shape" : "[1, 448, 1, 512]",
+        "name" : "k_cache_0",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 448)?",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 448]",
+        "name" : "v_cache_0",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 448 × 1 × 512)?",
+        "shortDescription" : "",
+        "shape" : "[1, 448, 1, 512]",
+        "name" : "k_cache_1",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 448)?",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 448]",
+        "name" : "v_cache_1",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 448 × 1 × 512)?",
+        "shortDescription" : "",
+        "shape" : "[1, 448, 1, 512]",
+        "name" : "k_cache_2",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 448)?",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 448]",
+        "name" : "v_cache_2",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 448 × 1 × 512)?",
+        "shortDescription" : "",
+        "shape" : "[1, 448, 1, 512]",
+        "name" : "k_cache_3",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 448)?",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 448]",
+        "name" : "v_cache_3",
+        "type" : "MultiArray"
+      }
+    ],
+    "generatedClassName" : "Llama_3_2_1B_Instruct_2024_10_10_23_56_41_chunk3",
+    "method" : "predict"
+  }
+]

Llama-3.2-1B-Instruct_chunk3.mlmodelc/weights/weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eddf0e85ab72700254696c13b5cb50e1fd525167a03d6d99aed4ec0c11867a4e
+size 486575936

Llama-3.2-1B-Instruct_chunk4.mlmodelc/analytics/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aa232688f37d1b639e14da601159147837ea3a6d7861e440adc8913fb9b3246f
+size 243

Llama-3.2-1B-Instruct_chunk4.mlmodelc/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:507886ccf77d43d177e993f226218bda5b20448ffeb13ac2e9f6a768ffb75b70
+size 931

Llama-3.2-1B-Instruct_chunk4.mlmodelc/metadata.json

@@ -0,0 +1,258 @@
+[
+  {
+    "metadataOutputVersion" : "3.0",
+    "storagePrecision" : "Float16",
+    "outputSchema" : [
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 2048 × 8 × 8)",
+        "shortDescription" : "",
+        "shape" : "[1, 2048, 8, 8]",
+        "name" : "new_x",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 64 × 1 × 512)",
+        "shortDescription" : "",
+        "shape" : "[1, 64, 1, 512]",
+        "name" : "new_k_cache_0",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 64)",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 64]",
+        "name" : "new_v_cache_0",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 64 × 1 × 512)",
+        "shortDescription" : "",
+        "shape" : "[1, 64, 1, 512]",
+        "name" : "new_k_cache_1",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 64)",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 64]",
+        "name" : "new_v_cache_1",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 64 × 1 × 512)",
+        "shortDescription" : "",
+        "shape" : "[1, 64, 1, 512]",
+        "name" : "new_k_cache_2",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 64)",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 64]",
+        "name" : "new_v_cache_2",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 64 × 1 × 512)",
+        "shortDescription" : "",
+        "shape" : "[1, 64, 1, 512]",
+        "name" : "new_k_cache_3",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 64)",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 64]",
+        "name" : "new_v_cache_3",
+        "type" : "MultiArray"
+      }
+    ],
+    "modelParameters" : [
+
+    ],
+    "specificationVersion" : 7,
+    "mlProgramOperationTypeHistogram" : {
+      "Concat" : 28,
+      "Ios16.mul" : 172,
+      "SliceByIndex" : 208,
+      "Transpose" : 4,
+      "Ios16.einsum" : 256,
+      "Ios16.conv" : 28,
+      "Ios16.add" : 144,
+      "Ios16.realDiv" : 8,
+      "Ios16.softmax" : 128,
+      "Ios16.reduceL2Norm" : 8,
+      "Ios16.reshape" : 28,
+      "Ios16.silu" : 4
+    },
+    "computePrecision" : "Mixed (Float16, Int32)",
+    "isUpdatable" : "0",
+    "availability" : {
+      "macOS" : "13.0",
+      "tvOS" : "16.0",
+      "visionOS" : "1.0",
+      "watchOS" : "9.0",
+      "iOS" : "16.0",
+      "macCatalyst" : "16.0"
+    },
+    "modelType" : {
+      "name" : "MLModelType_mlProgram"
+    },
+    "userDefinedMetadata" : {
+      "com.github.apple.coremltools.source_dialect" : "TorchScript",
+      "com.github.apple.coremltools.version" : "8.0b1",
+      "com.github.apple.coremltools.source" : "torch==2.1.0"
+    },
+    "inputSchema" : [
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 2048 × 8 × 8)",
+        "shortDescription" : "",
+        "shape" : "[1, 2048, 8, 8]",
+        "name" : "x",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 64 × 64)",
+        "shortDescription" : "",
+        "shape" : "[64, 64]",
+        "name" : "cos",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 64 × 64)",
+        "shortDescription" : "",
+        "shape" : "[64, 64]",
+        "name" : "sin",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 64)",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 64]",
+        "name" : "mask",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 448 × 1 × 512)?",
+        "shortDescription" : "",
+        "shape" : "[1, 448, 1, 512]",
+        "name" : "k_cache_0",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 448)?",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 448]",
+        "name" : "v_cache_0",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 448 × 1 × 512)?",
+        "shortDescription" : "",
+        "shape" : "[1, 448, 1, 512]",
+        "name" : "k_cache_1",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 448)?",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 448]",
+        "name" : "v_cache_1",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 448 × 1 × 512)?",
+        "shortDescription" : "",
+        "shape" : "[1, 448, 1, 512]",
+        "name" : "k_cache_2",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 448)?",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 448]",
+        "name" : "v_cache_2",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 448 × 1 × 512)?",
+        "shortDescription" : "",
+        "shape" : "[1, 448, 1, 512]",
+        "name" : "k_cache_3",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 448)?",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 448]",
+        "name" : "v_cache_3",
+        "type" : "MultiArray"
+      }
+    ],
+    "generatedClassName" : "Llama_3_2_1B_Instruct_2024_10_10_23_56_41_chunk4",
+    "method" : "predict"
+  }
+]

Llama-3.2-1B-Instruct_chunk4.mlmodelc/weights/weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1d882d8368974eb82dcef65c0fa5664b29da2ec73bdaf1d2798f5f923761197b
+size 486575936

Llama-3.2-1B-Instruct_chunk5.mlmodelc/analytics/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aa232688f37d1b639e14da601159147837ea3a6d7861e440adc8913fb9b3246f
+size 243

Llama-3.2-1B-Instruct_chunk5.mlmodelc/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bbed3683f2e4450c8c82941040938b8db75f66e3348e5f757e67745654a5b4e4
+size 931

Llama-3.2-1B-Instruct_chunk5.mlmodelc/metadata.json

@@ -0,0 +1,258 @@
+[
+  {
+    "metadataOutputVersion" : "3.0",
+    "storagePrecision" : "Float16",
+    "outputSchema" : [
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 2048 × 8 × 8)",
+        "shortDescription" : "",
+        "shape" : "[1, 2048, 8, 8]",
+        "name" : "new_x",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 64 × 1 × 512)",
+        "shortDescription" : "",
+        "shape" : "[1, 64, 1, 512]",
+        "name" : "new_k_cache_0",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 64)",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 64]",
+        "name" : "new_v_cache_0",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 64 × 1 × 512)",
+        "shortDescription" : "",
+        "shape" : "[1, 64, 1, 512]",
+        "name" : "new_k_cache_1",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 64)",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 64]",
+        "name" : "new_v_cache_1",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 64 × 1 × 512)",
+        "shortDescription" : "",
+        "shape" : "[1, 64, 1, 512]",
+        "name" : "new_k_cache_2",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 64)",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 64]",
+        "name" : "new_v_cache_2",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 64 × 1 × 512)",
+        "shortDescription" : "",
+        "shape" : "[1, 64, 1, 512]",
+        "name" : "new_k_cache_3",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 64)",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 64]",
+        "name" : "new_v_cache_3",
+        "type" : "MultiArray"
+      }
+    ],
+    "modelParameters" : [
+
+    ],
+    "specificationVersion" : 7,
+    "mlProgramOperationTypeHistogram" : {
+      "Concat" : 28,
+      "Ios16.mul" : 172,
+      "SliceByIndex" : 208,
+      "Transpose" : 4,
+      "Ios16.einsum" : 256,
+      "Ios16.conv" : 28,
+      "Ios16.add" : 144,
+      "Ios16.realDiv" : 8,
+      "Ios16.softmax" : 128,
+      "Ios16.reduceL2Norm" : 8,
+      "Ios16.reshape" : 28,
+      "Ios16.silu" : 4
+    },
+    "computePrecision" : "Mixed (Float16, Int32)",
+    "isUpdatable" : "0",
+    "availability" : {
+      "macOS" : "13.0",
+      "tvOS" : "16.0",
+      "visionOS" : "1.0",
+      "watchOS" : "9.0",
+      "iOS" : "16.0",
+      "macCatalyst" : "16.0"
+    },
+    "modelType" : {
+      "name" : "MLModelType_mlProgram"
+    },
+    "userDefinedMetadata" : {
+      "com.github.apple.coremltools.source_dialect" : "TorchScript",
+      "com.github.apple.coremltools.source" : "torch==2.1.0",
+      "com.github.apple.coremltools.version" : "8.0b1"
+    },
+    "inputSchema" : [
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 2048 × 8 × 8)",
+        "shortDescription" : "",
+        "shape" : "[1, 2048, 8, 8]",
+        "name" : "x",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 64 × 64)",
+        "shortDescription" : "",
+        "shape" : "[64, 64]",
+        "name" : "cos",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 64 × 64)",
+        "shortDescription" : "",
+        "shape" : "[64, 64]",
+        "name" : "sin",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 64)",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 64]",
+        "name" : "mask",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 448 × 1 × 512)?",
+        "shortDescription" : "",
+        "shape" : "[1, 448, 1, 512]",
+        "name" : "k_cache_0",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 448)?",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 448]",
+        "name" : "v_cache_0",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 448 × 1 × 512)?",
+        "shortDescription" : "",
+        "shape" : "[1, 448, 1, 512]",
+        "name" : "k_cache_1",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 448)?",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 448]",
+        "name" : "v_cache_1",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 448 × 1 × 512)?",
+        "shortDescription" : "",
+        "shape" : "[1, 448, 1, 512]",
+        "name" : "k_cache_2",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 448)?",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 448]",
+        "name" : "v_cache_2",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 448 × 1 × 512)?",
+        "shortDescription" : "",
+        "shape" : "[1, 448, 1, 512]",
+        "name" : "k_cache_3",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "1",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 448)?",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 448]",
+        "name" : "v_cache_3",
+        "type" : "MultiArray"
+      }
+    ],
+    "generatedClassName" : "Llama_3_2_1B_Instruct_2024_10_10_23_56_41_chunk5",
+    "method" : "predict"
+  }
+]

Llama-3.2-1B-Instruct_chunk5.mlmodelc/weights/weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cb54b3abf03d6e5002315bdfbf933e239c11369de782783287db9219e238cf8e
+size 486575936

Llama-3.2-1B-Instruct_chunk6.mlmodelc/analytics/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:23aa1a8a2b6bca88beeecf08d6281cdeb43aad33eb02cbe10ebff7eede7ed329
+size 243

Llama-3.2-1B-Instruct_chunk6.mlmodelc/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2740a88eec733758ae31768866e8550009ce36eb5177a09938d42cb18a095d05
+size 311

Llama-3.2-1B-Instruct_chunk6.mlmodelc/metadata.json

@@ -0,0 +1,64 @@
+[
+  {
+    "metadataOutputVersion" : "3.0",
+    "storagePrecision" : "Float16",
+    "outputSchema" : [
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 64 × 128256)",
+        "shortDescription" : "",
+        "shape" : "[1, 64, 128256]",
+        "name" : "logits",
+        "type" : "MultiArray"
+      }
+    ],
+    "modelParameters" : [
+
+    ],
+    "specificationVersion" : 7,
+    "mlProgramOperationTypeHistogram" : {
+      "Concat" : 2,
+      "Ios16.mul" : 2,
+      "Squeeze" : 1,
+      "Transpose" : 1,
+      "Ios16.reshape" : 10,
+      "Ios16.matmul" : 8,
+      "Ios16.realDiv" : 1,
+      "Ios16.reduceL2Norm" : 1
+    },
+    "computePrecision" : "Mixed (Float16, Int32)",
+    "isUpdatable" : "0",
+    "availability" : {
+      "macOS" : "13.0",
+      "tvOS" : "16.0",
+      "visionOS" : "1.0",
+      "watchOS" : "9.0",
+      "iOS" : "16.0",
+      "macCatalyst" : "16.0"
+    },
+    "modelType" : {
+      "name" : "MLModelType_mlProgram"
+    },
+    "userDefinedMetadata" : {
+      "com.github.apple.coremltools.source_dialect" : "TorchScript",
+      "com.github.apple.coremltools.source" : "torch==2.1.0",
+      "com.github.apple.coremltools.version" : "8.0b1"
+    },
+    "inputSchema" : [
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 2048 × 8 × 8)",
+        "shortDescription" : "",
+        "shape" : "[1, 2048, 8, 8]",
+        "name" : "x",
+        "type" : "MultiArray"
+      }
+    ],
+    "generatedClassName" : "Llama_3_2_1B_Instruct_2024_10_10_23_56_41_chunk6",
+    "method" : "predict"
+  }
+]

Llama-3.2-1B-Instruct_chunk6.mlmodelc/model.mil

@@ -0,0 +1,77 @@
+program(1.0)
+[buildInfo = dict<tensor<string, []>, tensor<string, []>>({{"coremlc-component-MIL", "3304.5.2"}, {"coremlc-version", "3304.6.2"}, {"coremltools-component-torch", "2.1.0"}, {"coremltools-source-dialect", "TorchScript"}, {"coremltools-version", "8.0b1"}})]
+{
+    func main<ios16>(tensor<fp16, [1, 2048, 8, 8]> x) {
+            tensor<bool, []> var_6 = const()[name = tensor<string, []>("op_6"), val = tensor<bool, []>(true)];
+            tensor<int32, []> var_9 = const()[name = tensor<string, []>("op_9"), val = tensor<int32, []>(1)];
+            tensor<bool, []> x_eps_interleave_0 = const()[name = tensor<string, []>("x_eps_interleave_0"), val = tensor<bool, []>(false)];
+            tensor<fp16, [1, 1, 8, 8]> eps_chan_to_fp16 = const()[name = tensor<string, []>("eps_chan_to_fp16"), val = tensor<fp16, [1, 1, 8, 8]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(64)))];
+            tensor<fp16, [1, 2049, 8, 8]> x_eps_cast_fp16 = concat(axis = var_9, interleave = x_eps_interleave_0, values = (x, eps_chan_to_fp16))[name = tensor<string, []>("x_eps_cast_fp16")];
+            tensor<int32, [1]> norm_x_axes_0 = const()[name = tensor<string, []>("norm_x_axes_0"), val = tensor<int32, [1]>([1])];
+            tensor<fp16, [1, 1, 8, 8]> norm_x_cast_fp16 = reduce_l2_norm(axes = norm_x_axes_0, keep_dims = var_6, x = x_eps_cast_fp16)[name = tensor<string, []>("norm_x_cast_fp16")];
+            tensor<fp16, [1, 2048, 8, 8]> x_normed_1_cast_fp16 = real_div(x = x, y = norm_x_cast_fp16)[name = tensor<string, []>("x_normed_1_cast_fp16")];
+            tensor<fp16, []> var_34_to_fp16 = const()[name = tensor<string, []>("op_34_to_fp16"), val = tensor<fp16, []>(0x1.6ap+5)];
+            tensor<fp16, [1, 2048, 8, 8]> x_normed_3_cast_fp16 = mul(x = x_normed_1_cast_fp16, y = var_34_to_fp16)[name = tensor<string, []>("x_normed_3_cast_fp16")];
+            tensor<fp16, [1, 2048, 1, 1]> ln_f_weight_to_fp16 = const()[name = tensor<string, []>("ln_f_weight_to_fp16"), val = tensor<fp16, [1, 2048, 1, 1]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(256)))];
+            tensor<fp16, [1, 2048, 8, 8]> x_5_cast_fp16 = mul(x = x_normed_3_cast_fp16, y = ln_f_weight_to_fp16)[name = tensor<string, []>("x_5_cast_fp16")];
+            tensor<int32, [4]> var_48 = const()[name = tensor<string, []>("op_48"), val = tensor<int32, [4]>([1, 2048, 1, -1])];
+            tensor<fp16, [1, 2048, 1, 64]> x_cast_fp16 = reshape(shape = var_48, x = x_5_cast_fp16)[name = tensor<string, []>("x_cast_fp16")];
+            tensor<int32, [1]> var_51_axes_0 = const()[name = tensor<string, []>("op_51_axes_0"), val = tensor<int32, [1]>([2])];
+            tensor<fp16, [1, 2048, 64]> var_51_cast_fp16 = squeeze(axes = var_51_axes_0, x = x_cast_fp16)[name = tensor<string, []>("op_51_cast_fp16")];
+            tensor<int32, [3]> var_54_perm_0 = const()[name = tensor<string, []>("op_54_perm_0"), val = tensor<int32, [3]>([0, 2, 1])];
+            tensor<int32, [2]> concat_4 = const()[name = tensor<string, []>("concat_4"), val = tensor<int32, [2]>([64, 2048])];
+            tensor<fp16, [1, 64, 2048]> var_54_cast_fp16 = transpose(perm = var_54_perm_0, x = var_51_cast_fp16)[name = tensor<string, []>("transpose_16")];
+            tensor<fp16, [64, 2048]> reshape_0_cast_fp16 = reshape(shape = concat_4, x = var_54_cast_fp16)[name = tensor<string, []>("reshape_0_cast_fp16")];
+            tensor<bool, []> matmul_0_transpose_x_0 = const()[name = tensor<string, []>("matmul_0_transpose_x_0"), val = tensor<bool, []>(false)];
+            tensor<bool, []> matmul_0_transpose_y_0 = const()[name = tensor<string, []>("matmul_0_transpose_y_0"), val = tensor<bool, []>(false)];
+            tensor<fp16, [2048, 16384]> transpose_1_to_fp16 = const()[name = tensor<string, []>("transpose_1_to_fp16"), val = tensor<fp16, [2048, 16384]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(4416)))];
+            tensor<fp16, [64, 16384]> matmul_0_cast_fp16 = matmul(transpose_x = matmul_0_transpose_x_0, transpose_y = matmul_0_transpose_y_0, x = reshape_0_cast_fp16, y = transpose_1_to_fp16)[name = tensor<string, []>("matmul_0_cast_fp16")];
+            tensor<int32, [3]> concat_8 = const()[name = tensor<string, []>("concat_8"), val = tensor<int32, [3]>([1, 64, 16384])];
+            tensor<fp16, [1, 64, 16384]> reshape_2_cast_fp16 = reshape(shape = concat_8, x = matmul_0_cast_fp16)[name = tensor<string, []>("reshape_2_cast_fp16")];
+            tensor<bool, []> matmul_1_transpose_x_0 = const()[name = tensor<string, []>("matmul_1_transpose_x_0"), val = tensor<bool, []>(false)];
+            tensor<bool, []> matmul_1_transpose_y_0 = const()[name = tensor<string, []>("matmul_1_transpose_y_0"), val = tensor<bool, []>(false)];
+            tensor<fp16, [2048, 16384]> transpose_3_to_fp16 = const()[name = tensor<string, []>("transpose_3_to_fp16"), val = tensor<fp16, [2048, 16384]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(67113344)))];
+            tensor<fp16, [64, 16384]> matmul_1_cast_fp16 = matmul(transpose_x = matmul_1_transpose_x_0, transpose_y = matmul_1_transpose_y_0, x = reshape_0_cast_fp16, y = transpose_3_to_fp16)[name = tensor<string, []>("matmul_1_cast_fp16")];
+            tensor<int32, [3]> concat_16 = const()[name = tensor<string, []>("concat_16"), val = tensor<int32, [3]>([1, 64, 16384])];
+            tensor<fp16, [1, 64, 16384]> reshape_5_cast_fp16 = reshape(shape = concat_16, x = matmul_1_cast_fp16)[name = tensor<string, []>("reshape_5_cast_fp16")];
+            tensor<bool, []> matmul_2_transpose_x_0 = const()[name = tensor<string, []>("matmul_2_transpose_x_0"), val = tensor<bool, []>(false)];
+            tensor<bool, []> matmul_2_transpose_y_0 = const()[name = tensor<string, []>("matmul_2_transpose_y_0"), val = tensor<bool, []>(false)];
+            tensor<fp16, [2048, 16384]> transpose_5_to_fp16 = const()[name = tensor<string, []>("transpose_5_to_fp16"), val = tensor<fp16, [2048, 16384]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(134222272)))];
+            tensor<fp16, [64, 16384]> matmul_2_cast_fp16 = matmul(transpose_x = matmul_2_transpose_x_0, transpose_y = matmul_2_transpose_y_0, x = reshape_0_cast_fp16, y = transpose_5_to_fp16)[name = tensor<string, []>("matmul_2_cast_fp16")];
+            tensor<int32, [3]> concat_24 = const()[name = tensor<string, []>("concat_24"), val = tensor<int32, [3]>([1, 64, 16384])];
+            tensor<fp16, [1, 64, 16384]> reshape_8_cast_fp16 = reshape(shape = concat_24, x = matmul_2_cast_fp16)[name = tensor<string, []>("reshape_8_cast_fp16")];
+            tensor<bool, []> matmul_3_transpose_x_0 = const()[name = tensor<string, []>("matmul_3_transpose_x_0"), val = tensor<bool, []>(false)];
+            tensor<bool, []> matmul_3_transpose_y_0 = const()[name = tensor<string, []>("matmul_3_transpose_y_0"), val = tensor<bool, []>(false)];
+            tensor<fp16, [2048, 16384]> transpose_7_to_fp16 = const()[name = tensor<string, []>("transpose_7_to_fp16"), val = tensor<fp16, [2048, 16384]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(201331200)))];
+            tensor<fp16, [64, 16384]> matmul_3_cast_fp16 = matmul(transpose_x = matmul_3_transpose_x_0, transpose_y = matmul_3_transpose_y_0, x = reshape_0_cast_fp16, y = transpose_7_to_fp16)[name = tensor<string, []>("matmul_3_cast_fp16")];
+            tensor<int32, [3]> concat_32 = const()[name = tensor<string, []>("concat_32"), val = tensor<int32, [3]>([1, 64, 16384])];
+            tensor<fp16, [1, 64, 16384]> reshape_11_cast_fp16 = reshape(shape = concat_32, x = matmul_3_cast_fp16)[name = tensor<string, []>("reshape_11_cast_fp16")];
+            tensor<bool, []> matmul_4_transpose_x_0 = const()[name = tensor<string, []>("matmul_4_transpose_x_0"), val = tensor<bool, []>(false)];
+            tensor<bool, []> matmul_4_transpose_y_0 = const()[name = tensor<string, []>("matmul_4_transpose_y_0"), val = tensor<bool, []>(false)];
+            tensor<fp16, [2048, 16384]> transpose_9_to_fp16 = const()[name = tensor<string, []>("transpose_9_to_fp16"), val = tensor<fp16, [2048, 16384]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(268440128)))];
+            tensor<fp16, [64, 16384]> matmul_4_cast_fp16 = matmul(transpose_x = matmul_4_transpose_x_0, transpose_y = matmul_4_transpose_y_0, x = reshape_0_cast_fp16, y = transpose_9_to_fp16)[name = tensor<string, []>("matmul_4_cast_fp16")];
+            tensor<int32, [3]> concat_40 = const()[name = tensor<string, []>("concat_40"), val = tensor<int32, [3]>([1, 64, 16384])];
+            tensor<fp16, [1, 64, 16384]> reshape_14_cast_fp16 = reshape(shape = concat_40, x = matmul_4_cast_fp16)[name = tensor<string, []>("reshape_14_cast_fp16")];
+            tensor<bool, []> matmul_5_transpose_x_0 = const()[name = tensor<string, []>("matmul_5_transpose_x_0"), val = tensor<bool, []>(false)];
+            tensor<bool, []> matmul_5_transpose_y_0 = const()[name = tensor<string, []>("matmul_5_transpose_y_0"), val = tensor<bool, []>(false)];
+            tensor<fp16, [2048, 16384]> transpose_11_to_fp16 = const()[name = tensor<string, []>("transpose_11_to_fp16"), val = tensor<fp16, [2048, 16384]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(335549056)))];
+            tensor<fp16, [64, 16384]> matmul_5_cast_fp16 = matmul(transpose_x = matmul_5_transpose_x_0, transpose_y = matmul_5_transpose_y_0, x = reshape_0_cast_fp16, y = transpose_11_to_fp16)[name = tensor<string, []>("matmul_5_cast_fp16")];
+            tensor<int32, [3]> concat_48 = const()[name = tensor<string, []>("concat_48"), val = tensor<int32, [3]>([1, 64, 16384])];
+            tensor<fp16, [1, 64, 16384]> reshape_17_cast_fp16 = reshape(shape = concat_48, x = matmul_5_cast_fp16)[name = tensor<string, []>("reshape_17_cast_fp16")];
+            tensor<bool, []> matmul_6_transpose_x_0 = const()[name = tensor<string, []>("matmul_6_transpose_x_0"), val = tensor<bool, []>(false)];
+            tensor<bool, []> matmul_6_transpose_y_0 = const()[name = tensor<string, []>("matmul_6_transpose_y_0"), val = tensor<bool, []>(false)];
+            tensor<fp16, [2048, 16384]> transpose_13_to_fp16 = const()[name = tensor<string, []>("transpose_13_to_fp16"), val = tensor<fp16, [2048, 16384]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(402657984)))];
+            tensor<fp16, [64, 16384]> matmul_6_cast_fp16 = matmul(transpose_x = matmul_6_transpose_x_0, transpose_y = matmul_6_transpose_y_0, x = reshape_0_cast_fp16, y = transpose_13_to_fp16)[name = tensor<string, []>("matmul_6_cast_fp16")];
+            tensor<int32, [3]> concat_56 = const()[name = tensor<string, []>("concat_56"), val = tensor<int32, [3]>([1, 64, 16384])];
+            tensor<fp16, [1, 64, 16384]> reshape_20_cast_fp16 = reshape(shape = concat_56, x = matmul_6_cast_fp16)[name = tensor<string, []>("reshape_20_cast_fp16")];
+            tensor<bool, []> matmul_7_transpose_x_0 = const()[name = tensor<string, []>("matmul_7_transpose_x_0"), val = tensor<bool, []>(false)];
+            tensor<bool, []> matmul_7_transpose_y_0 = const()[name = tensor<string, []>("matmul_7_transpose_y_0"), val = tensor<bool, []>(false)];
+            tensor<fp16, [2048, 13568]> transpose_15_to_fp16 = const()[name = tensor<string, []>("transpose_15_to_fp16"), val = tensor<fp16, [2048, 13568]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(469766912)))];
+            tensor<fp16, [64, 13568]> matmul_7_cast_fp16 = matmul(transpose_x = matmul_7_transpose_x_0, transpose_y = matmul_7_transpose_y_0, x = reshape_0_cast_fp16, y = transpose_15_to_fp16)[name = tensor<string, []>("matmul_7_cast_fp16")];
+            tensor<int32, [3]> concat_64 = const()[name = tensor<string, []>("concat_64"), val = tensor<int32, [3]>([1, 64, 13568])];
+            tensor<fp16, [1, 64, 13568]> reshape_23_cast_fp16 = reshape(shape = concat_64, x = matmul_7_cast_fp16)[name = tensor<string, []>("reshape_23_cast_fp16")];
+            tensor<int32, []> var_99 = const()[name = tensor<string, []>("op_99"), val = tensor<int32, []>(-1)];
+            tensor<bool, []> var_100_interleave_0 = const()[name = tensor<string, []>("op_100_interleave_0"), val = tensor<bool, []>(false)];
+            tensor<fp16, [1, 64, 128256]> logits = concat(axis = var_99, interleave = var_100_interleave_0, values = (reshape_2_cast_fp16, reshape_5_cast_fp16, reshape_8_cast_fp16, reshape_11_cast_fp16, reshape_14_cast_fp16, reshape_17_cast_fp16, reshape_20_cast_fp16, reshape_23_cast_fp16))[name = tensor<string, []>("op_100_cast_fp16")];
+        } -> (logits);
+}

Llama-3.2-1B-Instruct_chunk6.mlmodelc/weights/weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b4fb56007d0d4cb4a93aa67d2054ef3b3e2676a0358aff526a4ef5d66201b163
+size 525341504

cache-processor.mlmodelc/analytics/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d715da60630e06f07589e8fc3c2ed630f45943f1805cb6c078f284ee2655da88
+size 243

cache-processor.mlmodelc/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f0173108f39e072006d20029cbd37f4de85c21f1908ea1f7f433ffd68d8b42f3
+size 516

cache-processor.mlmodelc/metadata.json

@@ -0,0 +1,109 @@
+[
+  {
+    "metadataOutputVersion" : "3.0",
+    "outputSchema" : [
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 448 × 1 × 512)",
+        "shortDescription" : "",
+        "shape" : "[1, 448, 1, 512]",
+        "name" : "updated_k_cache",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 448)",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 448]",
+        "name" : "updated_v_cache",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16)",
+        "shortDescription" : "",
+        "shape" : "[]",
+        "name" : "ignore_me_im_only_here_so_this_runs_on_the_ane",
+        "type" : "MultiArray"
+      }
+    ],
+    "modelParameters" : [
+
+    ],
+    "specificationVersion" : 7,
+    "mlProgramOperationTypeHistogram" : {
+      "SliceByIndex" : 2,
+      "Ios16.mul" : 1,
+      "Concat" : 2,
+      "Ios16.reduceMin" : 1
+    },
+    "computePrecision" : "Mixed (Float16, Int32)",
+    "isUpdatable" : "0",
+    "availability" : {
+      "macOS" : "13.0",
+      "tvOS" : "16.0",
+      "visionOS" : "1.0",
+      "watchOS" : "9.0",
+      "iOS" : "16.0",
+      "macCatalyst" : "16.0"
+    },
+    "modelType" : {
+      "name" : "MLModelType_mlProgram"
+    },
+    "userDefinedMetadata" : {
+      "com.github.apple.coremltools.source_dialect" : "TorchScript",
+      "com.github.apple.coremltools.source" : "torch==2.1.0",
+      "com.github.apple.coremltools.version" : "8.0b1"
+    },
+    "inputSchema" : [
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 448 × 1 × 512)",
+        "shortDescription" : "",
+        "shape" : "[1, 448, 1, 512]",
+        "name" : "old_k_cache",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 64 × 1 × 512)",
+        "shortDescription" : "",
+        "shape" : "[1, 64, 1, 512]",
+        "name" : "new_k_cache",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 448)",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 448]",
+        "name" : "old_v_cache",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 512 × 1 × 64)",
+        "shortDescription" : "",
+        "shape" : "[1, 512, 1, 64]",
+        "name" : "new_v_cache",
+        "type" : "MultiArray"
+      }
+    ],
+    "generatedClassName" : "cache_processor",
+    "method" : "predict"
+  }
+]

cache-processor.mlmodelc/model.mil

@@ -0,0 +1,24 @@
+program(1.0)
+[buildInfo = dict<tensor<string, []>, tensor<string, []>>({{"coremlc-component-MIL", "3304.5.2"}, {"coremlc-version", "3304.6.2"}, {"coremltools-component-torch", "2.1.0"}, {"coremltools-source-dialect", "TorchScript"}, {"coremltools-version", "8.0b1"}})]
+{
+    func main<ios16>(tensor<fp16, [1, 64, 1, 512]> new_k_cache, tensor<fp16, [1, 512, 1, 64]> new_v_cache, tensor<fp16, [1, 448, 1, 512]> old_k_cache, tensor<fp16, [1, 512, 1, 448]> old_v_cache) {
+            tensor<int32, []> var_6 = const()[name = tensor<string, []>("op_6"), val = tensor<int32, []>(-3)];
+            tensor<bool, []> cat_k_1_interleave_0 = const()[name = tensor<string, []>("cat_k_1_interleave_0"), val = tensor<bool, []>(false)];
+            tensor<fp16, [1, 512, 1, 512]> cat_k_1_cast_fp16 = concat(axis = var_6, interleave = cat_k_1_interleave_0, values = (old_k_cache, new_k_cache))[name = tensor<string, []>("cat_k_1_cast_fp16")];
+            tensor<int32, []> var_9 = const()[name = tensor<string, []>("op_9"), val = tensor<int32, []>(-1)];
+            tensor<bool, []> cat_v_interleave_0 = const()[name = tensor<string, []>("cat_v_interleave_0"), val = tensor<bool, []>(false)];
+            tensor<fp16, [1, 512, 1, 512]> cat_v_cast_fp16 = concat(axis = var_9, interleave = cat_v_interleave_0, values = (old_v_cache, new_v_cache))[name = tensor<string, []>("cat_v_cast_fp16")];
+            tensor<int32, [4]> var_20_begin_0 = const()[name = tensor<string, []>("op_20_begin_0"), val = tensor<int32, [4]>([0, 64, 0, 0])];
+            tensor<int32, [4]> var_20_end_0 = const()[name = tensor<string, []>("op_20_end_0"), val = tensor<int32, [4]>([1, 512, 1, 512])];
+            tensor<bool, [4]> var_20_end_mask_0 = const()[name = tensor<string, []>("op_20_end_mask_0"), val = tensor<bool, [4]>([true, false, true, true])];
+            tensor<fp16, [1, 448, 1, 512]> updated_k_cache = slice_by_index(begin = var_20_begin_0, end = var_20_end_0, end_mask = var_20_end_mask_0, x = cat_k_1_cast_fp16)[name = tensor<string, []>("op_20_cast_fp16")];
+            tensor<int32, [4]> var_50_begin_0 = const()[name = tensor<string, []>("op_50_begin_0"), val = tensor<int32, [4]>([0, 0, 0, 64])];
+            tensor<int32, [4]> var_50_end_0 = const()[name = tensor<string, []>("op_50_end_0"), val = tensor<int32, [4]>([1, 512, 1, 512])];
+            tensor<bool, [4]> var_50_end_mask_0 = const()[name = tensor<string, []>("op_50_end_mask_0"), val = tensor<bool, [4]>([true, true, true, false])];
+            tensor<fp16, [1, 512, 1, 448]> updated_v_cache = slice_by_index(begin = var_50_begin_0, end = var_50_end_0, end_mask = var_50_end_mask_0, x = cat_v_cast_fp16)[name = tensor<string, []>("op_50_cast_fp16")];
+            tensor<fp16, []> var_51_promoted_to_fp16 = const()[name = tensor<string, []>("op_51_promoted_to_fp16"), val = tensor<fp16, []>(0x1p+1)];
+            tensor<fp16, [1, 448, 1, 512]> prod_cast_fp16 = mul(x = updated_k_cache, y = var_51_promoted_to_fp16)[name = tensor<string, []>("prod_cast_fp16")];
+            tensor<bool, []> var_53_keep_dims_0 = const()[name = tensor<string, []>("op_53_keep_dims_0"), val = tensor<bool, []>(false)];
+            tensor<fp16, []> ignore_me_im_only_here_so_this_runs_on_the_ane = reduce_min(keep_dims = var_53_keep_dims_0, x = prod_cast_fp16)[name = tensor<string, []>("op_53_cast_fp16")];
+        } -> (updated_k_cache, updated_v_cache, ignore_me_im_only_here_so_this_runs_on_the_ane);
+}

logit-processor.mlmodelc/analytics/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0ad03dc247f59282bf008d857db8620b0ad600eb939bfa2a4e8a78438e1c2573
+size 243

logit-processor.mlmodelc/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ccca55190c5da56bfc175471f3239eeeb7bffece8d38d565de9443edef9c9148
+size 378

logit-processor.mlmodelc/metadata.json

@@ -0,0 +1,58 @@
+[
+  {
+    "metadataOutputVersion" : "3.0",
+    "outputSchema" : [
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Int32",
+        "formattedType" : "MultiArray (Int32)",
+        "shortDescription" : "",
+        "shape" : "[]",
+        "name" : "argmax",
+        "type" : "MultiArray"
+      }
+    ],
+    "modelParameters" : [
+
+    ],
+    "specificationVersion" : 7,
+    "mlProgramOperationTypeHistogram" : {
+      "Ios16.reduceArgmax" : 1
+    },
+    "computePrecision" : "Mixed (Float16, Int32)",
+    "isUpdatable" : "0",
+    "availability" : {
+      "macOS" : "13.0",
+      "tvOS" : "16.0",
+      "visionOS" : "1.0",
+      "watchOS" : "9.0",
+      "iOS" : "16.0",
+      "macCatalyst" : "16.0"
+    },
+    "modelType" : {
+      "name" : "MLModelType_mlProgram"
+    },
+    "userDefinedMetadata" : {
+      "com.github.apple.coremltools.source_dialect" : "TorchScript",
+      "com.github.apple.coremltools.source" : "torch==2.1.0",
+      "com.github.apple.coremltools.version" : "8.0b1"
+    },
+    "inputSchema" : [
+      {
+        "shortDescription" : "",
+        "dataType" : "Float16",
+        "hasShapeFlexibility" : "1",
+        "isOptional" : "0",
+        "shapeFlexibility" : "1 × 511 × 32000 | 1 × 1 × 32000 | 1 × 2 × 32000 | 1 × 4 × 32000 | 1 × 64 × 32000 | 1 × 64 × 128256 | 1 × 512 × 32000",
+        "formattedType" : "MultiArray (Float16 1 × 511 × 32000)",
+        "type" : "MultiArray",
+        "shape" : "[1, 511, 32000]",
+        "name" : "logits",
+        "enumeratedShapes" : "[[1, 511, 32000], [1, 1, 32000], [1, 2, 32000], [1, 4, 32000], [1, 64, 32000], [1, 64, 128256], [1, 512, 32000]]"
+      }
+    ],
+    "generatedClassName" : "logit_processor",
+    "method" : "predict"
+  }
+]

logit-processor.mlmodelc/model.mil

@@ -0,0 +1,9 @@
+program(1.0)
+[buildInfo = dict<tensor<string, []>, tensor<string, []>>({{"coremlc-component-MIL", "3304.5.2"}, {"coremlc-version", "3304.6.2"}, {"coremltools-component-torch", "2.1.0"}, {"coremltools-source-dialect", "TorchScript"}, {"coremltools-version", "8.0b1"}})]
+{
+    func main<ios16>(tensor<fp16, [1, ?, ?]> logits) [FlexibleShapeInformation = tuple<tuple<tensor<string, []>, dict<tensor<string, []>, tensor<int32, [?]>>>, tuple<tensor<string, []>, dict<tensor<string, []>, dict<tensor<string, []>, tensor<int32, [?]>>>>>((("DefaultShapes", {{"logits", [1, 511, 32000]}}), ("EnumeratedShapes", {{"logits_1_1_1_1_32000_", {{"logits", [1, 1, 32000]}}}, {"logits_1_1_1_2_32000_", {{"logits", [1, 2, 32000]}}}, {"logits_1_1_1_4_32000_", {{"logits", [1, 4, 32000]}}}, {"logits_1_1_1_511_32000_", {{"logits", [1, 511, 32000]}}}, {"logits_1_1_1_512_32000_", {{"logits", [1, 512, 32000]}}}, {"logits_1_1_1_64_128256_", {{"logits", [1, 64, 128256]}}}, {"logits_1_1_1_64_32000_", {{"logits", [1, 64, 32000]}}}})))] {
+            tensor<int32, []> var_2 = const()[name = tensor<string, []>("op_2"), val = tensor<int32, []>(-1)];
+            tensor<bool, []> var_3 = const()[name = tensor<string, []>("op_3"), val = tensor<bool, []>(false)];
+            tensor<int32, [1, ?]> argmax = reduce_argmax(axis = var_2, keep_dims = var_3, x = logits)[name = tensor<string, []>("op_4_cast_fp16")];
+        } -> (argmax);
+}