Upload folder using huggingface_hub

openai_whisper-tiny/AudioEncoder.mlmodelc/analytics/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:37dbcb7943d1c5b92e0ae956102df6f69f343d1c6584f9f1511e2044fa8890f8
+size 243

openai_whisper-tiny/AudioEncoder.mlmodelc/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:abdbc66c120f462ae758502376c141c105e094780c537f997be62a8d577e775a
+size 347

openai_whisper-tiny/AudioEncoder.mlmodelc/metadata.json

@@ -0,0 +1,70 @@
+[
+  {
+    "metadataOutputVersion" : "3.0",
+    "storagePrecision" : "Float16",
+    "outputSchema" : [
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 384 × 1 × 1500)",
+        "shortDescription" : "",
+        "shape" : "[1, 384, 1, 1500]",
+        "name" : "encoder_output_embeds",
+        "type" : "MultiArray"
+      }
+    ],
+    "modelParameters" : [
+
+    ],
+    "specificationVersion" : 7,
+    "mlProgramOperationTypeHistogram" : {
+      "Concat" : 28,
+      "Ios16.add" : 9,
+      "Ios16.mul" : 144,
+      "SliceByIndex" : 216,
+      "Transpose" : 4,
+      "Ios16.batchNorm" : 9,
+      "Ios16.gelu" : 6,
+      "Ios16.einsum" : 288,
+      "Ios16.softmax" : 144,
+      "Ios16.layerNorm" : 9,
+      "Ios16.conv" : 26
+    },
+    "computePrecision" : "Mixed (Float16, Int32)",
+    "isUpdatable" : "0",
+    "stateSchema" : [
+
+    ],
+    "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==2.6.0",
+      "com.github.apple.coremltools.version" : "8.2"
+    },
+    "inputSchema" : [
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 80 × 1 × 3000)",
+        "shortDescription" : "",
+        "shape" : "[1, 80, 1, 3000]",
+        "name" : "melspectrogram_features",
+        "type" : "MultiArray"
+      }
+    ],
+    "generatedClassName" : "AudioEncoder",
+    "method" : "predict"
+  }
+]

openai_whisper-tiny/AudioEncoder.mlmodelc/weights/weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:52267fceba5c3f1bacb93027ec8e6edf72f096030d6e2ade687a616d67aeceea
+size 16422784

openai_whisper-tiny/MelSpectrogram.mlcomputeplan.json

@@ -0,0 +1,195 @@
+{
+    "1_ios16.reshape_shape_x": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU"
+        ],
+        "cost": 2.2206
+    },
+    "5_pad_pad_x_constant_val_mode": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU"
+        ],
+        "cost": 2.2216
+    },
+    "7_ios16.reshape_shape_x": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU"
+        ],
+        "cost": 2.2225
+    },
+    "9_expand_dims_x_axes": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU"
+        ],
+        "cost": 2.2225
+    },
+    "12_expand_dims_x_axes": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU"
+        ],
+        "cost": 2.2225
+    },
+    "18_ios16.conv_groups_weight_x_dilations_strides_pad_pad_type": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU"
+        ],
+        "cost": 21.5236
+    },
+    "24_ios16.conv_groups_weight_x_dilations_strides_pad_pad_type": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU"
+        ],
+        "cost": 21.5236
+    },
+    "26_squeeze_x_axes": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 2.7906
+    },
+    "28_squeeze_x_axes": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 2.7906
+    },
+    "29_ios16.square_x": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 2.7906
+    },
+    "30_ios16.square_x": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 2.7906
+    },
+    "31_ios16.add_x_y": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 10.1532
+    },
+    "32_identity_x": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 2.7906
+    },
+    "36_slice_by_index_begin_end_end_mask_x": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 2.7901
+    },
+    "40_ios16.matmul_y_transpose_y_transpose_x_x": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 4.0374
+    },
+    "42_ios16.add_x_y": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 2.0199
+    },
+    "44_ios16.log_epsilon_x": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 1.1103
+    },
+    "46_ios16.mul_x_y": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 2.9367
+    },
+    "48_ios16.reduce_max_x_keep_dims": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.5552
+    },
+    "50_ios16.sub_x_y": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0
+    },
+    "51_ios16.maximum_x_y": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 1.1103
+    },
+    "53_ios16.add_x_y": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 2.0199
+    },
+    "55_ios16.mul_x_y": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 2.9367
+    },
+    "57_expand_dims_x_axes": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 1.1103
+    },
+    "59_expand_dims_x_axes": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 1.1103
+    }
+}

openai_whisper-tiny/MelSpectrogram.mlmodelc/analytics/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c4f367993f0198e9858a4d89fb054318982c91a9bb5946e29231421c2f1100b9
+size 243

openai_whisper-tiny/MelSpectrogram.mlmodelc/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:806321f1034184a10b04dc50816219dec8ae9789698712050c81edecb9bb5aa7
+size 328

openai_whisper-tiny/MelSpectrogram.mlmodelc/metadata.json

@@ -0,0 +1,74 @@
+[
+  {
+    "metadataOutputVersion" : "3.0",
+    "storagePrecision" : "Float16",
+    "outputSchema" : [
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 80 × 1 × 3000)",
+        "shortDescription" : "",
+        "shape" : "[1, 80, 1, 3000]",
+        "name" : "melspectrogram_features",
+        "type" : "MultiArray"
+      }
+    ],
+    "modelParameters" : [
+
+    ],
+    "specificationVersion" : 7,
+    "mlProgramOperationTypeHistogram" : {
+      "Ios16.reshape" : 2,
+      "Ios16.mul" : 2,
+      "SliceByIndex" : 1,
+      "Ios16.sub" : 1,
+      "Ios16.log" : 1,
+      "Ios16.square" : 2,
+      "Ios16.add" : 3,
+      "Squeeze" : 2,
+      "Ios16.matmul" : 1,
+      "Ios16.conv" : 2,
+      "Ios16.maximum" : 1,
+      "ExpandDims" : 4,
+      "Ios16.reduceMax" : 1,
+      "Identity" : 1,
+      "Pad" : 1
+    },
+    "computePrecision" : "Mixed (Float16, Int32)",
+    "isUpdatable" : "0",
+    "stateSchema" : [
+
+    ],
+    "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" : "8.2",
+      "com.github.apple.coremltools.source" : "torch==2.6.0"
+    },
+    "inputSchema" : [
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 480000)",
+        "shortDescription" : "",
+        "shape" : "[480000]",
+        "name" : "audio",
+        "type" : "MultiArray"
+      }
+    ],
+    "generatedClassName" : "MelSpectrogram",
+    "method" : "predict"
+  }
+]

openai_whisper-tiny/MelSpectrogram.mlmodelc/model.mil

@@ -0,0 +1,66 @@
+program(1.0)
+[buildInfo = dict<tensor<string, []>, tensor<string, []>>({{"coremlc-component-MIL", "3402.3.2"}, {"coremlc-version", "3402.4.1"}, {"coremltools-component-torch", "2.6.0"}, {"coremltools-source-dialect", "TorchScript"}, {"coremltools-version", "8.2"}})]
+{
+    func main<ios16>(tensor<fp16, [480000]> audio) {
+            tensor<int32, [3]> var_10 = const()[name = tensor<string, []>("op_10"), val = tensor<int32, [3]>([1, 1, 480000])];
+            tensor<fp16, [1, 1, 480000]> input_1_cast_fp16 = reshape(shape = var_10, x = audio)[name = tensor<string, []>("input_1_cast_fp16")];
+            tensor<int32, [6]> input_3_pad_0 = const()[name = tensor<string, []>("input_3_pad_0"), val = tensor<int32, [6]>([0, 0, 0, 0, 200, 200])];
+            tensor<string, []> input_3_mode_0 = const()[name = tensor<string, []>("input_3_mode_0"), val = tensor<string, []>("reflect")];
+            tensor<fp16, []> const_1_to_fp16 = const()[name = tensor<string, []>("const_1_to_fp16"), val = tensor<fp16, []>(0x0p+0)];
+            tensor<fp16, [1, 1, 480400]> input_3_cast_fp16 = pad(constant_val = const_1_to_fp16, mode = input_3_mode_0, pad = input_3_pad_0, x = input_1_cast_fp16)[name = tensor<string, []>("input_3_cast_fp16")];
+            tensor<int32, [1]> var_22 = const()[name = tensor<string, []>("op_22"), val = tensor<int32, [1]>([480400])];
+            tensor<fp16, [480400]> input_cast_fp16 = reshape(shape = var_22, x = input_3_cast_fp16)[name = tensor<string, []>("input_cast_fp16")];
+            tensor<int32, [1]> expand_dims_0_axes_0 = const()[name = tensor<string, []>("expand_dims_0_axes_0"), val = tensor<int32, [1]>([0])];
+            tensor<fp16, [1, 480400]> expand_dims_0_cast_fp16 = expand_dims(axes = expand_dims_0_axes_0, x = input_cast_fp16)[name = tensor<string, []>("expand_dims_0_cast_fp16")];
+            tensor<int32, [1]> expand_dims_3 = const()[name = tensor<string, []>("expand_dims_3"), val = tensor<int32, [1]>([160])];
+            tensor<int32, [1]> expand_dims_4_axes_0 = const()[name = tensor<string, []>("expand_dims_4_axes_0"), val = tensor<int32, [1]>([1])];
+            tensor<fp16, [1, 1, 480400]> expand_dims_4_cast_fp16 = expand_dims(axes = expand_dims_4_axes_0, x = expand_dims_0_cast_fp16)[name = tensor<string, []>("expand_dims_4_cast_fp16")];
+            tensor<string, []> conv_0_pad_type_0 = const()[name = tensor<string, []>("conv_0_pad_type_0"), val = tensor<string, []>("valid")];
+            tensor<int32, [2]> conv_0_pad_0 = const()[name = tensor<string, []>("conv_0_pad_0"), val = tensor<int32, [2]>([0, 0])];
+            tensor<int32, [1]> conv_0_dilations_0 = const()[name = tensor<string, []>("conv_0_dilations_0"), val = tensor<int32, [1]>([1])];
+            tensor<int32, []> conv_0_groups_0 = const()[name = tensor<string, []>("conv_0_groups_0"), val = tensor<int32, []>(1)];
+            tensor<fp16, [201, 1, 400]> expand_dims_1_to_fp16 = const()[name = tensor<string, []>("expand_dims_1_to_fp16"), val = tensor<fp16, [201, 1, 400]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(64)))];
+            tensor<fp16, [1, 201, 3001]> conv_0_cast_fp16 = conv(dilations = conv_0_dilations_0, groups = conv_0_groups_0, pad = conv_0_pad_0, pad_type = conv_0_pad_type_0, strides = expand_dims_3, weight = expand_dims_1_to_fp16, x = expand_dims_4_cast_fp16)[name = tensor<string, []>("conv_0_cast_fp16")];
+            tensor<string, []> conv_1_pad_type_0 = const()[name = tensor<string, []>("conv_1_pad_type_0"), val = tensor<string, []>("valid")];
+            tensor<int32, [2]> conv_1_pad_0 = const()[name = tensor<string, []>("conv_1_pad_0"), val = tensor<int32, [2]>([0, 0])];
+            tensor<int32, [1]> conv_1_dilations_0 = const()[name = tensor<string, []>("conv_1_dilations_0"), val = tensor<int32, [1]>([1])];
+            tensor<int32, []> conv_1_groups_0 = const()[name = tensor<string, []>("conv_1_groups_0"), val = tensor<int32, []>(1)];
+            tensor<fp16, [201, 1, 400]> expand_dims_2_to_fp16 = const()[name = tensor<string, []>("expand_dims_2_to_fp16"), val = tensor<fp16, [201, 1, 400]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(160960)))];
+            tensor<fp16, [1, 201, 3001]> conv_1_cast_fp16 = conv(dilations = conv_1_dilations_0, groups = conv_1_groups_0, pad = conv_1_pad_0, pad_type = conv_1_pad_type_0, strides = expand_dims_3, weight = expand_dims_2_to_fp16, x = expand_dims_4_cast_fp16)[name = tensor<string, []>("conv_1_cast_fp16")];
+            tensor<int32, [1]> squeeze_0_axes_0 = const()[name = tensor<string, []>("squeeze_0_axes_0"), val = tensor<int32, [1]>([0])];
+            tensor<fp16, [201, 3001]> squeeze_0_cast_fp16 = squeeze(axes = squeeze_0_axes_0, x = conv_0_cast_fp16)[name = tensor<string, []>("squeeze_0_cast_fp16")];
+            tensor<int32, [1]> squeeze_1_axes_0 = const()[name = tensor<string, []>("squeeze_1_axes_0"), val = tensor<int32, [1]>([0])];
+            tensor<fp16, [201, 3001]> squeeze_1_cast_fp16 = squeeze(axes = squeeze_1_axes_0, x = conv_1_cast_fp16)[name = tensor<string, []>("squeeze_1_cast_fp16")];
+            tensor<fp16, [201, 3001]> square_0_cast_fp16 = square(x = squeeze_0_cast_fp16)[name = tensor<string, []>("square_0_cast_fp16")];
+            tensor<fp16, [201, 3001]> square_1_cast_fp16 = square(x = squeeze_1_cast_fp16)[name = tensor<string, []>("square_1_cast_fp16")];
+            tensor<fp16, [201, 3001]> add_1_cast_fp16 = add(x = square_0_cast_fp16, y = square_1_cast_fp16)[name = tensor<string, []>("add_1_cast_fp16")];
+            tensor<fp16, [201, 3001]> magnitudes_1_cast_fp16 = identity(x = add_1_cast_fp16)[name = tensor<string, []>("magnitudes_1_cast_fp16")];
+            tensor<int32, [2]> magnitudes_begin_0 = const()[name = tensor<string, []>("magnitudes_begin_0"), val = tensor<int32, [2]>([0, 0])];
+            tensor<int32, [2]> magnitudes_end_0 = const()[name = tensor<string, []>("magnitudes_end_0"), val = tensor<int32, [2]>([201, 3000])];
+            tensor<bool, [2]> magnitudes_end_mask_0 = const()[name = tensor<string, []>("magnitudes_end_mask_0"), val = tensor<bool, [2]>([true, false])];
+            tensor<fp16, [201, 3000]> magnitudes_cast_fp16 = slice_by_index(begin = magnitudes_begin_0, end = magnitudes_end_0, end_mask = magnitudes_end_mask_0, x = magnitudes_1_cast_fp16)[name = tensor<string, []>("magnitudes_cast_fp16")];
+            tensor<bool, []> mel_spec_1_transpose_x_0 = const()[name = tensor<string, []>("mel_spec_1_transpose_x_0"), val = tensor<bool, []>(false)];
+            tensor<bool, []> mel_spec_1_transpose_y_0 = const()[name = tensor<string, []>("mel_spec_1_transpose_y_0"), val = tensor<bool, []>(false)];
+            tensor<fp16, [80, 201]> mel_filters_to_fp16 = const()[name = tensor<string, []>("mel_filters_to_fp16"), val = tensor<fp16, [80, 201]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(321856)))];
+            tensor<fp16, [80, 3000]> mel_spec_1_cast_fp16 = matmul(transpose_x = mel_spec_1_transpose_x_0, transpose_y = mel_spec_1_transpose_y_0, x = mel_filters_to_fp16, y = magnitudes_cast_fp16)[name = tensor<string, []>("mel_spec_1_cast_fp16")];
+            tensor<fp16, []> var_41_to_fp16 = const()[name = tensor<string, []>("op_41_to_fp16"), val = tensor<fp16, []>(0x1p-24)];
+            tensor<fp16, [80, 3000]> mel_spec_cast_fp16 = add(x = mel_spec_1_cast_fp16, y = var_41_to_fp16)[name = tensor<string, []>("mel_spec_cast_fp16")];
+            tensor<fp16, []> log_0_epsilon_0_to_fp16 = const()[name = tensor<string, []>("log_0_epsilon_0_to_fp16"), val = tensor<fp16, []>(0x0p+0)];
+            tensor<fp16, [80, 3000]> log_0_cast_fp16 = log(epsilon = log_0_epsilon_0_to_fp16, x = mel_spec_cast_fp16)[name = tensor<string, []>("log_0_cast_fp16")];
+            tensor<fp16, []> mul_0_y_0_to_fp16 = const()[name = tensor<string, []>("mul_0_y_0_to_fp16"), val = tensor<fp16, []>(0x1.bccp-2)];
+            tensor<fp16, [80, 3000]> mul_0_cast_fp16 = mul(x = log_0_cast_fp16, y = mul_0_y_0_to_fp16)[name = tensor<string, []>("mul_0_cast_fp16")];
+            tensor<bool, []> var_44_keep_dims_0 = const()[name = tensor<string, []>("op_44_keep_dims_0"), val = tensor<bool, []>(false)];
+            tensor<fp16, []> var_44_cast_fp16 = reduce_max(keep_dims = var_44_keep_dims_0, x = mul_0_cast_fp16)[name = tensor<string, []>("op_44_cast_fp16")];
+            tensor<fp16, []> var_46_to_fp16 = const()[name = tensor<string, []>("op_46_to_fp16"), val = tensor<fp16, []>(0x1p+3)];
+            tensor<fp16, []> var_47_cast_fp16 = sub(x = var_44_cast_fp16, y = var_46_to_fp16)[name = tensor<string, []>("op_47_cast_fp16")];
+            tensor<fp16, [80, 3000]> log_spec_3_cast_fp16 = maximum(x = mul_0_cast_fp16, y = var_47_cast_fp16)[name = tensor<string, []>("log_spec_3_cast_fp16")];
+            tensor<fp16, []> var_50_to_fp16 = const()[name = tensor<string, []>("op_50_to_fp16"), val = tensor<fp16, []>(0x1p+2)];
+            tensor<fp16, [80, 3000]> var_51_cast_fp16 = add(x = log_spec_3_cast_fp16, y = var_50_to_fp16)[name = tensor<string, []>("op_51_cast_fp16")];
+            tensor<fp16, []> _inversed_log_spec_y_0_to_fp16 = const()[name = tensor<string, []>("_inversed_log_spec_y_0_to_fp16"), val = tensor<fp16, []>(0x1p-2)];
+            tensor<fp16, [80, 3000]> _inversed_log_spec_cast_fp16 = mul(x = var_51_cast_fp16, y = _inversed_log_spec_y_0_to_fp16)[name = tensor<string, []>("_inversed_log_spec_cast_fp16")];
+            tensor<int32, [1]> var_55_axes_0 = const()[name = tensor<string, []>("op_55_axes_0"), val = tensor<int32, [1]>([0])];
+            tensor<fp16, [1, 80, 3000]> var_55_cast_fp16 = expand_dims(axes = var_55_axes_0, x = _inversed_log_spec_cast_fp16)[name = tensor<string, []>("op_55_cast_fp16")];
+            tensor<int32, [1]> var_62_axes_0 = const()[name = tensor<string, []>("op_62_axes_0"), val = tensor<int32, [1]>([2])];
+            tensor<fp16, [1, 80, 1, 3000]> melspectrogram_features = expand_dims(axes = var_62_axes_0, x = var_55_cast_fp16)[name = tensor<string, []>("op_62_cast_fp16")];
+        } -> (melspectrogram_features);
+}

openai_whisper-tiny/MelSpectrogram.mlmodelc/weights/weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:801024dbc7a89c677be1f8b285de3409e35f7d1786c9c8d9d0d6842ac57a1c83
+size 354080

openai_whisper-tiny/TextDecoder.mlcomputeplan.json

@@ -0,0 +1,1567 @@
+{
+    "3_ios16.gather_axis_batch_dims_indices_x": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU"
+        ],
+        "cost": 57.7926
+    },
+    "7_ios16.gather_axis_batch_dims_indices_x": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU"
+        ],
+        "cost": 0.4992
+    },
+    "8_ios16.add_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0016
+    },
+    "10_expand_dims_x_axes": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "12_expand_dims_x_axes": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "15_split_axis_x_split_sizes": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.7864
+    },
+    "18_split_axis_x_split_sizes": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.7864
+    },
+    "22_ios16.layer_norm_epsilon_x_axes": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "28_ios16.batch_norm_epsilon_gamma_mean_variance_x_beta": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "36_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0329
+    },
+    "43_ios16.conv_groups_weight_x_dilations_strides_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0328
+    },
+    "51_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0329
+    },
+    "53_expand_dims_x_axes": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0008
+    },
+    "55_expand_dims_x_axes": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0008
+    },
+    "56_ios16.mul_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0014
+    },
+    "57_ios16.add_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.6997
+    },
+    "58_ios16.mul_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0014
+    },
+    "59_ios16.add_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.6997
+    },
+    "61_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "63_ios16.mul_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0006
+    },
+    "65_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.3146
+    },
+    "68_ios16.matmul_y_transpose_y_transpose_x_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.1601
+    },
+    "70_expand_dims_x_axes": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0008
+    },
+    "72_expand_dims_x_axes": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0008
+    },
+    "73_ios16.add_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0064
+    },
+    "74_ios16.softmax_axis_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0049
+    },
+    "76_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.3146
+    },
+    "79_ios16.matmul_y_transpose_y_transpose_x_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.1601
+    },
+    "81_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "89_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0329
+    },
+    "90_ios16.add_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0016
+    },
+    "93_ios16.layer_norm_epsilon_x_axes": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "97_ios16.batch_norm_epsilon_gamma_mean_variance_x_beta": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "105_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0329
+    },
+    "112_ios16.conv_groups_weight_x_dilations_strides_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 1.9687
+    },
+    "120_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 1.9712
+    },
+    "122_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "124_ios16.mul_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0006
+    },
+    "126_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 1.0532
+    },
+    "129_ios16.matmul_y_transpose_y_transpose_x_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.5352
+    },
+    "130_ios16.softmax_axis_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0165
+    },
+    "132_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 1.0532
+    },
+    "135_ios16.matmul_y_transpose_y_transpose_x_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.5352
+    },
+    "137_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "145_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0329
+    },
+    "146_ios16.add_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0016
+    },
+    "149_ios16.layer_norm_epsilon_x_axes": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "153_ios16.batch_norm_epsilon_gamma_mean_variance_x_beta": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "161_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.1314
+    },
+    "163_ios16.gelu_mode_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0028
+    },
+    "171_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.1313
+    },
+    "172_ios16.add_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0016
+    },
+    "176_ios16.layer_norm_epsilon_x_axes": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "180_ios16.batch_norm_epsilon_gamma_mean_variance_x_beta": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "188_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0329
+    },
+    "195_ios16.conv_groups_weight_x_dilations_strides_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0328
+    },
+    "203_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0329
+    },
+    "204_ios16.mul_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0014
+    },
+    "205_ios16.add_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.6997
+    },
+    "206_ios16.mul_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0014
+    },
+    "207_ios16.add_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.6997
+    },
+    "209_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "211_ios16.mul_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0006
+    },
+    "213_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.3146
+    },
+    "216_ios16.matmul_y_transpose_y_transpose_x_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.1601
+    },
+    "217_ios16.add_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0064
+    },
+    "218_ios16.softmax_axis_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0049
+    },
+    "220_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.3146
+    },
+    "223_ios16.matmul_y_transpose_y_transpose_x_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.1601
+    },
+    "225_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "233_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0329
+    },
+    "234_ios16.add_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0016
+    },
+    "237_ios16.layer_norm_epsilon_x_axes": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "241_ios16.batch_norm_epsilon_gamma_mean_variance_x_beta": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "249_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0329
+    },
+    "256_ios16.conv_groups_weight_x_dilations_strides_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 1.9687
+    },
+    "264_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 1.9712
+    },
+    "266_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "268_ios16.mul_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0006
+    },
+    "270_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 1.0532
+    },
+    "273_ios16.matmul_y_transpose_y_transpose_x_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.5352
+    },
+    "274_ios16.softmax_axis_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0165
+    },
+    "276_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 1.0532
+    },
+    "279_ios16.matmul_y_transpose_y_transpose_x_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.5352
+    },
+    "281_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "289_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0329
+    },
+    "290_ios16.add_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0016
+    },
+    "293_ios16.layer_norm_epsilon_x_axes": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "297_ios16.batch_norm_epsilon_gamma_mean_variance_x_beta": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "305_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.1314
+    },
+    "307_ios16.gelu_mode_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0028
+    },
+    "315_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.1313
+    },
+    "316_ios16.add_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0016
+    },
+    "320_ios16.layer_norm_epsilon_x_axes": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "324_ios16.batch_norm_epsilon_gamma_mean_variance_x_beta": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "332_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0329
+    },
+    "339_ios16.conv_groups_weight_x_dilations_strides_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0328
+    },
+    "347_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0329
+    },
+    "348_ios16.mul_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0014
+    },
+    "349_ios16.add_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.6997
+    },
+    "350_ios16.mul_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0014
+    },
+    "351_ios16.add_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.6997
+    },
+    "353_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "355_ios16.mul_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0006
+    },
+    "357_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.3146
+    },
+    "360_ios16.matmul_y_transpose_y_transpose_x_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.1601
+    },
+    "361_ios16.add_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0064
+    },
+    "362_ios16.softmax_axis_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0049
+    },
+    "364_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.3146
+    },
+    "367_ios16.matmul_y_transpose_y_transpose_x_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.1601
+    },
+    "369_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "377_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0329
+    },
+    "378_ios16.add_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0016
+    },
+    "381_ios16.layer_norm_epsilon_x_axes": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "385_ios16.batch_norm_epsilon_gamma_mean_variance_x_beta": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "393_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0329
+    },
+    "400_ios16.conv_groups_weight_x_dilations_strides_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 1.9687
+    },
+    "408_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 1.9712
+    },
+    "410_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "412_ios16.mul_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0006
+    },
+    "414_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 1.0532
+    },
+    "417_ios16.matmul_y_transpose_y_transpose_x_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.5352
+    },
+    "418_ios16.softmax_axis_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0165
+    },
+    "420_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 1.0532
+    },
+    "423_ios16.matmul_y_transpose_y_transpose_x_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.5352
+    },
+    "425_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "433_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0329
+    },
+    "434_ios16.add_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0016
+    },
+    "437_ios16.layer_norm_epsilon_x_axes": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "441_ios16.batch_norm_epsilon_gamma_mean_variance_x_beta": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "449_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.1314
+    },
+    "451_ios16.gelu_mode_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0028
+    },
+    "459_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.1313
+    },
+    "460_ios16.add_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0016
+    },
+    "464_ios16.layer_norm_epsilon_x_axes": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "468_ios16.batch_norm_epsilon_gamma_mean_variance_x_beta": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "476_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0329
+    },
+    "483_ios16.conv_groups_weight_x_dilations_strides_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0328
+    },
+    "491_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0329
+    },
+    "492_ios16.mul_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0014
+    },
+    "493_ios16.add_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.6997
+    },
+    "494_ios16.mul_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0014
+    },
+    "495_ios16.add_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.6997
+    },
+    "497_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "499_ios16.mul_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0006
+    },
+    "501_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.3146
+    },
+    "504_ios16.matmul_y_transpose_y_transpose_x_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.1601
+    },
+    "505_ios16.add_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0064
+    },
+    "506_ios16.softmax_axis_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0049
+    },
+    "508_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.3146
+    },
+    "511_ios16.matmul_y_transpose_y_transpose_x_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.1601
+    },
+    "513_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "521_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0329
+    },
+    "522_ios16.add_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0016
+    },
+    "525_ios16.layer_norm_epsilon_x_axes": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "529_ios16.batch_norm_epsilon_gamma_mean_variance_x_beta": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "537_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0329
+    },
+    "544_ios16.conv_groups_weight_x_dilations_strides_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 1.9687
+    },
+    "552_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 1.9712
+    },
+    "554_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "556_ios16.mul_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0006
+    },
+    "558_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 1.0532
+    },
+    "561_ios16.matmul_y_transpose_y_transpose_x_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.5352
+    },
+    "562_ios16.softmax_axis_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0165
+    },
+    "564_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 1.0532
+    },
+    "567_ios16.matmul_y_transpose_y_transpose_x_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.5352
+    },
+    "569_ios16.reshape_shape_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "577_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0329
+    },
+    "578_ios16.add_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0016
+    },
+    "581_ios16.layer_norm_epsilon_x_axes": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "585_ios16.batch_norm_epsilon_gamma_mean_variance_x_beta": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "593_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.1314
+    },
+    "595_ios16.gelu_mode_x": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0028
+    },
+    "603_ios16.conv_groups_bias_weight_dilations_strides_x_pad_pad_type": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.1313
+    },
+    "604_ios16.add_x_y": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0016
+    },
+    "607_ios16.layer_norm_epsilon_x_axes": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "611_ios16.batch_norm_epsilon_gamma_mean_variance_x_beta": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "613_squeeze_x_axes": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0007
+    },
+    "616_transpose_x_perm": {
+        "dispatch": "ANE",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0013
+    },
+    "617_ios16.linear_weight_x_bias": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU"
+        ],
+        "cost": 0.069
+    },
+    "620_concat_axis_values_interleave": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0089
+    },
+    "623_concat_axis_values_interleave": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0089
+    },
+    "627_slice_by_index_begin_end_end_mask_x": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0115
+    },
+    "632_slice_by_index_begin_end_end_mask_squeeze_mask_x": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0033
+    },
+    "636_slice_by_index_begin_end_end_mask_x": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0115
+    },
+    "641_slice_by_index_begin_end_end_mask_squeeze_mask_x": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0033
+    },
+    "645_slice_by_index_begin_end_end_mask_x": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0115
+    },
+    "650_slice_by_index_begin_end_end_mask_squeeze_mask_x": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0033
+    },
+    "654_slice_by_index_begin_end_end_mask_x": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0115
+    },
+    "659_slice_by_index_begin_end_end_mask_squeeze_mask_x": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0033
+    },
+    "663_slice_by_index_begin_end_end_mask_x": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0115
+    },
+    "668_slice_by_index_begin_end_end_mask_squeeze_mask_x": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0033
+    },
+    "672_slice_by_index_begin_end_end_mask_x": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0115
+    },
+    "677_slice_by_index_begin_end_end_mask_squeeze_mask_x": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0033
+    },
+    "680_concat_axis_values_interleave": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0522
+    },
+    "683_ios16.reduce_mean_x_keep_dims_axes": {
+        "dispatch": "CPU",
+        "supported": [
+            "CPU",
+            "ANE"
+        ],
+        "cost": 0.0115
+    }
+}

openai_whisper-tiny/TextDecoder.mlmodelc/analytics/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:70c25a4908f1b0082f2bd62989f2120eb08addebca634439c9207b6ee776154a
+size 243

openai_whisper-tiny/TextDecoder.mlmodelc/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:940113f3791f56fae5e918f1fce8102f6f903cbd7b8cb5a9e49b2d011d4239b5
+size 633

openai_whisper-tiny/TextDecoder.mlmodelc/metadata.json

@@ -0,0 +1,167 @@
+[
+  {
+    "metadataOutputVersion" : "3.0",
+    "storagePrecision" : "Float16",
+    "outputSchema" : [
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 1 × 51865)",
+        "shortDescription" : "",
+        "shape" : "[1, 1, 51865]",
+        "name" : "logits",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 1536 × 1 × 1)",
+        "shortDescription" : "",
+        "shape" : "[1, 1536, 1, 1]",
+        "name" : "key_cache_updates",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 1536 × 1 × 1)",
+        "shortDescription" : "",
+        "shape" : "[1, 1536, 1, 1]",
+        "name" : "value_cache_updates",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 1500)",
+        "shortDescription" : "",
+        "shape" : "[1, 1500]",
+        "name" : "alignment_heads_weights",
+        "type" : "MultiArray"
+      }
+    ],
+    "modelParameters" : [
+
+    ],
+    "specificationVersion" : 7,
+    "mlProgramOperationTypeHistogram" : {
+      "Ios16.linear" : 1,
+      "Concat" : 3,
+      "Ios16.reduceMean" : 1,
+      "Ios16.mul" : 16,
+      "Ios16.layerNorm" : 13,
+      "SliceByIndex" : 12,
+      "Transpose" : 1,
+      "Ios16.conv" : 40,
+      "Ios16.add" : 25,
+      "Squeeze" : 1,
+      "Ios16.matmul" : 16,
+      "Ios16.softmax" : 8,
+      "Ios16.gelu" : 4,
+      "ExpandDims" : 6,
+      "Ios16.batchNorm" : 13,
+      "Split" : 2,
+      "Ios16.gather" : 2,
+      "Ios16.reshape" : 32
+    },
+    "computePrecision" : "Mixed (Float16, Int32)",
+    "isUpdatable" : "0",
+    "stateSchema" : [
+
+    ],
+    "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==2.6.0",
+      "com.github.apple.coremltools.version" : "8.2"
+    },
+    "inputSchema" : [
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Int32",
+        "formattedType" : "MultiArray (Int32 1)",
+        "shortDescription" : "",
+        "shape" : "[1]",
+        "name" : "input_ids",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Int32",
+        "formattedType" : "MultiArray (Int32 1)",
+        "shortDescription" : "",
+        "shape" : "[1]",
+        "name" : "cache_length",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 1536 × 1 × 448)",
+        "shortDescription" : "",
+        "shape" : "[1, 1536, 1, 448]",
+        "name" : "key_cache",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 1536 × 1 × 448)",
+        "shortDescription" : "",
+        "shape" : "[1, 1536, 1, 448]",
+        "name" : "value_cache",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 448)",
+        "shortDescription" : "",
+        "shape" : "[1, 448]",
+        "name" : "kv_cache_update_mask",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 384 × 1 × 1500)",
+        "shortDescription" : "",
+        "shape" : "[1, 384, 1, 1500]",
+        "name" : "encoder_output_embeds",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 448)",
+        "shortDescription" : "",
+        "shape" : "[1, 448]",
+        "name" : "decoder_key_padding_mask",
+        "type" : "MultiArray"
+      }
+    ],
+    "generatedClassName" : "TextDecoder",
+    "method" : "predict"
+  }
+]

openai_whisper-tiny/TextDecoder.mlmodelc/weights/weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:778d39fa5b45d8a628ba68b2037c7be19f7d22d9116bf8cfc06013272c7f6def
+size 59216434

openai_whisper-tiny/config.json

@@ -0,0 +1 @@
+{"_name_or_path": "openai/whisper-tiny", "activation_dropout": 0.0, "activation_function": "gelu", "architectures": ["WhisperForConditionalGeneration"], "attention_dropout": 0.0, "begin_suppress_tokens": [220, 50257], "bos_token_id": 50257, "d_model": 384, "decoder_attention_heads": 6, "decoder_ffn_dim": 1536, "decoder_layerdrop": 0.0, "decoder_layers": 4, "decoder_start_token_id": 50258, "dropout": 0.0, "encoder_attention_heads": 6, "encoder_ffn_dim": 1536, "encoder_layerdrop": 0.0, "encoder_layers": 4, "eos_token_id": 50257, "forced_decoder_ids": [[1, 50259], [2, 50359], [3, 50363]], "init_std": 0.02, "is_encoder_decoder": true, "max_length": 448, "max_source_positions": 1500, "max_target_positions": 448, "model_type": "whisper", "num_hidden_layers": 4, "num_mel_bins": 80, "pad_token_id": 50257, "scale_embedding": false, "suppress_tokens": [1, 2, 7, 8, 9, 10, 14, 25, 26, 27, 28, 29, 31, 58, 59, 60, 61, 62, 63, 90, 91, 92, 93, 359, 503, 522, 542, 873, 893, 902, 918, 922, 931, 1350, 1853, 1982, 2460, 2627, 3246, 3253, 3268, 3536, 3846, 3961, 4183, 4667, 6585, 6647, 7273, 9061, 9383, 10428, 10929, 11938, 12033, 12331, 12562, 13793, 14157, 14635, 15265, 15618, 16553, 16604, 18362, 18956, 20075, 21675, 22520, 26130, 26161, 26435, 28279, 29464, 31650, 32302, 32470, 36865, 42863, 47425, 49870, 50254, 50258, 50358, 50359, 50360, 50361, 50362], "torch_dtype": "float32", "transformers_version": "4.27.0.dev0", "use_cache": true, "vocab_size": 51865}

openai_whisper-tiny/generation_config.json

@@ -0,0 +1 @@
+{"alignment_heads": [[2, 2], [3, 0], [3, 2], [3, 3], [3, 4], [3, 5]], "begin_suppress_tokens": [220, 50257], "bos_token_id": 50257, "decoder_start_token_id": 50258, "eos_token_id": 50257, "forced_decoder_ids": [[1, null], [2, 50359]], "is_multilingual": true, "lang_to_id": {"<|af|>": 50327, "<|am|>": 50334, "<|ar|>": 50272, "<|as|>": 50350, "<|az|>": 50304, "<|ba|>": 50355, "<|be|>": 50330, "<|bg|>": 50292, "<|bn|>": 50302, "<|bo|>": 50347, "<|br|>": 50309, "<|bs|>": 50315, "<|ca|>": 50270, "<|cs|>": 50283, "<|cy|>": 50297, "<|da|>": 50285, "<|de|>": 50261, "<|el|>": 50281, "<|en|>": 50259, "<|es|>": 50262, "<|et|>": 50307, "<|eu|>": 50310, "<|fa|>": 50300, "<|fi|>": 50277, "<|fo|>": 50338, "<|fr|>": 50265, "<|gl|>": 50319, "<|gu|>": 50333, "<|haw|>": 50352, "<|ha|>": 50354, "<|he|>": 50279, "<|hi|>": 50276, "<|hr|>": 50291, "<|ht|>": 50339, "<|hu|>": 50286, "<|hy|>": 50312, "<|id|>": 50275, "<|is|>": 50311, "<|it|>": 50274, "<|ja|>": 50266, "<|jw|>": 50356, "<|ka|>": 50329, "<|kk|>": 50316, "<|km|>": 50323, "<|kn|>": 50306, "<|ko|>": 50264, "<|la|>": 50294, "<|lb|>": 50345, "<|ln|>": 50353, "<|lo|>": 50336, "<|lt|>": 50293, "<|lv|>": 50301, "<|mg|>": 50349, "<|mi|>": 50295, "<|mk|>": 50308, "<|ml|>": 50296, "<|mn|>": 50314, "<|mr|>": 50320, "<|ms|>": 50282, "<|mt|>": 50343, "<|my|>": 50346, "<|ne|>": 50313, "<|nl|>": 50271, "<|nn|>": 50342, "<|no|>": 50288, "<|oc|>": 50328, "<|pa|>": 50321, "<|pl|>": 50269, "<|ps|>": 50340, "<|pt|>": 50267, "<|ro|>": 50284, "<|ru|>": 50263, "<|sa|>": 50344, "<|sd|>": 50332, "<|si|>": 50322, "<|sk|>": 50298, "<|sl|>": 50305, "<|sn|>": 50324, "<|so|>": 50326, "<|sq|>": 50317, "<|sr|>": 50303, "<|su|>": 50357, "<|sv|>": 50273, "<|sw|>": 50318, "<|ta|>": 50287, "<|te|>": 50299, "<|tg|>": 50331, "<|th|>": 50289, "<|tk|>": 50341, "<|tl|>": 50348, "<|tr|>": 50268, "<|tt|>": 50351, "<|uk|>": 50280, "<|ur|>": 50290, "<|uz|>": 50337, "<|vi|>": 50278, "<|yi|>": 50335, "<|yo|>": 50325, "<|zh|>": 50260}, "max_initial_timestamp_index": 50, "max_length": 448, "no_timestamps_token_id": 50363, "pad_token_id": 50257, "prev_sot_token_id": 50361, "return_timestamps": false, "suppress_tokens": [1, 2, 7, 8, 9, 10, 14, 25, 26, 27, 28, 29, 31, 58, 59, 60, 61, 62, 63, 90, 91, 92, 93, 359, 503, 522, 542, 873, 893, 902, 918, 922, 931, 1350, 1853, 1982, 2460, 2627, 3246, 3253, 3268, 3536, 3846, 3961, 4183, 4667, 6585, 6647, 7273, 9061, 9383, 10428, 10929, 11938, 12033, 12331, 12562, 13793, 14157, 14635, 15265, 15618, 16553, 16604, 18362, 18956, 20075, 21675, 22520, 26130, 26161, 26435, 28279, 29464, 31650, 32302, 32470, 36865, 42863, 47425, 49870, 50254, 50258, 50358, 50359, 50360, 50361, 50362], "task_to_id": {"transcribe": 50359, "translate": 50358}, "transformers_version": "4.31.0.dev0"}

openai_whisper___1/AudioEncoder.mlmodelc/analytics/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0b25820e5b2ab0b0686b4bea147fb217d1d1bface45170ff4ffde01fa6864ae2
+size 243

openai_whisper___1/AudioEncoder.mlmodelc/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:142c33ade402fe41952059f175eb855093dfe09b5d2b84624a31e3a9952ed47d
+size 347

openai_whisper___1/AudioEncoder.mlmodelc/metadata.json

@@ -0,0 +1,69 @@
+[
+  {
+    "metadataOutputVersion" : "3.0",
+    "storagePrecision" : "Float16",
+    "outputSchema" : [
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 384 × 1 × 1500)",
+        "shortDescription" : "",
+        "shape" : "[1, 384, 1, 1500]",
+        "name" : "encoder_output_embeds",
+        "type" : "MultiArray"
+      }
+    ],
+    "modelParameters" : [
+
+    ],
+    "specificationVersion" : 7,
+    "mlProgramOperationTypeHistogram" : {
+      "Concat" : 28,
+      "Ios16.rsqrt" : 9,
+      "Ios16.mul" : 114,
+      "SliceByIndex" : 168,
+      "Ios16.sub" : 9,
+      "Transpose" : 4,
+      "Ios16.einsum" : 192,
+      "Ios16.conv" : 26,
+      "Ios16.add" : 18,
+      "Ios16.reduceMean" : 18,
+      "Ios16.softmax" : 96,
+      "Ios16.gelu" : 6,
+      "Ios16.batchNorm" : 9
+    },
+    "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==2.2.1",
+      "com.github.apple.coremltools.version" : "7.1"
+    },
+    "inputSchema" : [
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 80 × 1 × 3000)",
+        "shortDescription" : "",
+        "shape" : "[1, 80, 1, 3000]",
+        "name" : "melspectrogram_features",
+        "type" : "MultiArray"
+      }
+    ],
+    "generatedClassName" : "AudioEncoder",
+    "method" : "predict"
+  }
+]

openai_whisper___1/AudioEncoder.mlmodelc/model.mlmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:030d64a3ddd296d6f709691a66a870aab7ee9f19e5fe07e8086245fb85302802
+size 54965

openai_whisper___1/AudioEncoder.mlmodelc/weights/weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bcd0879f6d1c61832765c7ec05d883d0dcbf1504057b13095fd315484196fc5e
+size 16422784

openai_whisper___1/MelSpectrogram.mlmodelc/analytics/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7f77e6457285248f99cd7aa3fd4cc2efbb17733e63e7023ac53abe1f95785d07
+size 243

openai_whisper___1/MelSpectrogram.mlmodelc/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dabdc5aa69f6ef4d97dc9499f5c30514e00e96b53b750b33a5a6471363c71662
+size 328

openai_whisper___1/MelSpectrogram.mlmodelc/metadata.json

@@ -0,0 +1,71 @@
+[
+  {
+    "metadataOutputVersion" : "3.0",
+    "storagePrecision" : "Float16",
+    "outputSchema" : [
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 80 × 1 × 3000)",
+        "shortDescription" : "",
+        "shape" : "[1, 80, 1, 3000]",
+        "name" : "melspectrogram_features",
+        "type" : "MultiArray"
+      }
+    ],
+    "modelParameters" : [
+
+    ],
+    "specificationVersion" : 7,
+    "mlProgramOperationTypeHistogram" : {
+      "Pad" : 1,
+      "Ios16.mul" : 2,
+      "SliceByIndex" : 1,
+      "Ios16.sub" : 1,
+      "Ios16.log" : 1,
+      "Ios16.conv" : 2,
+      "Ios16.add" : 3,
+      "Ios16.square" : 2,
+      "Ios16.matmul" : 1,
+      "Squeeze" : 2,
+      "Ios16.maximum" : 1,
+      "ExpandDims" : 4,
+      "Ios16.reduceMax" : 1,
+      "Identity" : 1,
+      "Ios16.reshape" : 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.version" : "7.1",
+      "com.github.apple.coremltools.source" : "torch==2.2.1"
+    },
+    "inputSchema" : [
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 480000)",
+        "shortDescription" : "",
+        "shape" : "[480000]",
+        "name" : "audio",
+        "type" : "MultiArray"
+      }
+    ],
+    "generatedClassName" : "MelSpectrogram",
+    "method" : "predict"
+  }
+]

openai_whisper___1/MelSpectrogram.mlmodelc/model.mil

@@ -0,0 +1,66 @@
+program(1.0)
+[buildInfo = dict<tensor<string, []>, tensor<string, []>>({{"coremlc-component-MIL", "5.33.5"}, {"coremlc-version", "1877.40.3"}, {"coremltools-component-torch", "2.2.1"}, {"coremltools-source-dialect", "TorchScript"}, {"coremltools-version", "7.1"}})]
+{
+    func main<ios16>(tensor<fp16, [480000]> audio) {
+            tensor<int32, [3]> var_10 = const()[name = tensor<string, []>("op_10"), val = tensor<int32, [3]>([1, 1, 480000])];
+            tensor<fp16, [1, 1, 480000]> input_1_cast_fp16 = reshape(shape = var_10, x = audio)[name = tensor<string, []>("input_1_cast_fp16")];
+            tensor<int32, [6]> input_3_pad_0 = const()[name = tensor<string, []>("input_3_pad_0"), val = tensor<int32, [6]>([0, 0, 0, 0, 200, 200])];
+            tensor<string, []> input_3_mode_0 = const()[name = tensor<string, []>("input_3_mode_0"), val = tensor<string, []>("reflect")];
+            tensor<fp16, []> input_3_constant_val_0_to_fp16 = const()[name = tensor<string, []>("input_3_constant_val_0_to_fp16"), val = tensor<fp16, []>(0x0p+0)];
+            tensor<fp16, [1, 1, 480400]> input_3_cast_fp16 = pad(constant_val = input_3_constant_val_0_to_fp16, mode = input_3_mode_0, pad = input_3_pad_0, x = input_1_cast_fp16)[name = tensor<string, []>("input_3_cast_fp16")];
+            tensor<int32, [1]> var_22 = const()[name = tensor<string, []>("op_22"), val = tensor<int32, [1]>([480400])];
+            tensor<fp16, [480400]> input_cast_fp16 = reshape(shape = var_22, x = input_3_cast_fp16)[name = tensor<string, []>("input_cast_fp16")];
+            tensor<int32, [1]> expand_dims_0_axes_0 = const()[name = tensor<string, []>("expand_dims_0_axes_0"), val = tensor<int32, [1]>([0])];
+            tensor<fp16, [1, 480400]> expand_dims_0_cast_fp16 = expand_dims(axes = expand_dims_0_axes_0, x = input_cast_fp16)[name = tensor<string, []>("expand_dims_0_cast_fp16")];
+            tensor<int32, [1]> expand_dims_3 = const()[name = tensor<string, []>("expand_dims_3"), val = tensor<int32, [1]>([160])];
+            tensor<int32, [1]> expand_dims_4_axes_0 = const()[name = tensor<string, []>("expand_dims_4_axes_0"), val = tensor<int32, [1]>([1])];
+            tensor<fp16, [1, 1, 480400]> expand_dims_4_cast_fp16 = expand_dims(axes = expand_dims_4_axes_0, x = expand_dims_0_cast_fp16)[name = tensor<string, []>("expand_dims_4_cast_fp16")];
+            tensor<string, []> conv_0_pad_type_0 = const()[name = tensor<string, []>("conv_0_pad_type_0"), val = tensor<string, []>("valid")];
+            tensor<int32, [2]> conv_0_pad_0 = const()[name = tensor<string, []>("conv_0_pad_0"), val = tensor<int32, [2]>([0, 0])];
+            tensor<int32, [1]> conv_0_dilations_0 = const()[name = tensor<string, []>("conv_0_dilations_0"), val = tensor<int32, [1]>([1])];
+            tensor<int32, []> conv_0_groups_0 = const()[name = tensor<string, []>("conv_0_groups_0"), val = tensor<int32, []>(1)];
+            tensor<fp16, [201, 1, 400]> expand_dims_1_to_fp16 = const()[name = tensor<string, []>("expand_dims_1_to_fp16"), val = tensor<fp16, [201, 1, 400]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(64)))];
+            tensor<fp16, [1, 201, 3001]> conv_0_cast_fp16 = conv(dilations = conv_0_dilations_0, groups = conv_0_groups_0, pad = conv_0_pad_0, pad_type = conv_0_pad_type_0, strides = expand_dims_3, weight = expand_dims_1_to_fp16, x = expand_dims_4_cast_fp16)[name = tensor<string, []>("conv_0_cast_fp16")];
+            tensor<string, []> conv_1_pad_type_0 = const()[name = tensor<string, []>("conv_1_pad_type_0"), val = tensor<string, []>("valid")];
+            tensor<int32, [2]> conv_1_pad_0 = const()[name = tensor<string, []>("conv_1_pad_0"), val = tensor<int32, [2]>([0, 0])];
+            tensor<int32, [1]> conv_1_dilations_0 = const()[name = tensor<string, []>("conv_1_dilations_0"), val = tensor<int32, [1]>([1])];
+            tensor<int32, []> conv_1_groups_0 = const()[name = tensor<string, []>("conv_1_groups_0"), val = tensor<int32, []>(1)];
+            tensor<fp16, [201, 1, 400]> expand_dims_2_to_fp16 = const()[name = tensor<string, []>("expand_dims_2_to_fp16"), val = tensor<fp16, [201, 1, 400]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(160960)))];
+            tensor<fp16, [1, 201, 3001]> conv_1_cast_fp16 = conv(dilations = conv_1_dilations_0, groups = conv_1_groups_0, pad = conv_1_pad_0, pad_type = conv_1_pad_type_0, strides = expand_dims_3, weight = expand_dims_2_to_fp16, x = expand_dims_4_cast_fp16)[name = tensor<string, []>("conv_1_cast_fp16")];
+            tensor<int32, [1]> squeeze_0_axes_0 = const()[name = tensor<string, []>("squeeze_0_axes_0"), val = tensor<int32, [1]>([0])];
+            tensor<fp16, [201, 3001]> squeeze_0_cast_fp16 = squeeze(axes = squeeze_0_axes_0, x = conv_0_cast_fp16)[name = tensor<string, []>("squeeze_0_cast_fp16")];
+            tensor<int32, [1]> squeeze_1_axes_0 = const()[name = tensor<string, []>("squeeze_1_axes_0"), val = tensor<int32, [1]>([0])];
+            tensor<fp16, [201, 3001]> squeeze_1_cast_fp16 = squeeze(axes = squeeze_1_axes_0, x = conv_1_cast_fp16)[name = tensor<string, []>("squeeze_1_cast_fp16")];
+            tensor<fp16, [201, 3001]> square_0_cast_fp16 = square(x = squeeze_0_cast_fp16)[name = tensor<string, []>("square_0_cast_fp16")];
+            tensor<fp16, [201, 3001]> square_1_cast_fp16 = square(x = squeeze_1_cast_fp16)[name = tensor<string, []>("square_1_cast_fp16")];
+            tensor<fp16, [201, 3001]> add_1_cast_fp16 = add(x = square_0_cast_fp16, y = square_1_cast_fp16)[name = tensor<string, []>("add_1_cast_fp16")];
+            tensor<fp16, [201, 3001]> magnitudes_1_cast_fp16 = identity(x = add_1_cast_fp16)[name = tensor<string, []>("magnitudes_1_cast_fp16")];
+            tensor<int32, [2]> magnitudes_begin_0 = const()[name = tensor<string, []>("magnitudes_begin_0"), val = tensor<int32, [2]>([0, 0])];
+            tensor<int32, [2]> magnitudes_end_0 = const()[name = tensor<string, []>("magnitudes_end_0"), val = tensor<int32, [2]>([201, 3000])];
+            tensor<bool, [2]> magnitudes_end_mask_0 = const()[name = tensor<string, []>("magnitudes_end_mask_0"), val = tensor<bool, [2]>([true, false])];
+            tensor<fp16, [201, 3000]> magnitudes_cast_fp16 = slice_by_index(begin = magnitudes_begin_0, end = magnitudes_end_0, end_mask = magnitudes_end_mask_0, x = magnitudes_1_cast_fp16)[name = tensor<string, []>("magnitudes_cast_fp16")];
+            tensor<bool, []> mel_spec_1_transpose_x_0 = const()[name = tensor<string, []>("mel_spec_1_transpose_x_0"), val = tensor<bool, []>(false)];
+            tensor<bool, []> mel_spec_1_transpose_y_0 = const()[name = tensor<string, []>("mel_spec_1_transpose_y_0"), val = tensor<bool, []>(false)];
+            tensor<fp16, [80, 201]> mel_filters_to_fp16 = const()[name = tensor<string, []>("mel_filters_to_fp16"), val = tensor<fp16, [80, 201]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(321856)))];
+            tensor<fp16, [80, 3000]> mel_spec_1_cast_fp16 = matmul(transpose_x = mel_spec_1_transpose_x_0, transpose_y = mel_spec_1_transpose_y_0, x = mel_filters_to_fp16, y = magnitudes_cast_fp16)[name = tensor<string, []>("mel_spec_1_cast_fp16")];
+            tensor<fp16, []> var_41_to_fp16 = const()[name = tensor<string, []>("op_41_to_fp16"), val = tensor<fp16, []>(0x1p-24)];
+            tensor<fp16, [80, 3000]> mel_spec_cast_fp16 = add(x = mel_spec_1_cast_fp16, y = var_41_to_fp16)[name = tensor<string, []>("mel_spec_cast_fp16")];
+            tensor<fp16, []> log_0_epsilon_0_to_fp16 = const()[name = tensor<string, []>("log_0_epsilon_0_to_fp16"), val = tensor<fp16, []>(0x0p+0)];
+            tensor<fp16, [80, 3000]> log_0_cast_fp16 = log(epsilon = log_0_epsilon_0_to_fp16, x = mel_spec_cast_fp16)[name = tensor<string, []>("log_0_cast_fp16")];
+            tensor<fp16, []> mul_0_y_0_to_fp16 = const()[name = tensor<string, []>("mul_0_y_0_to_fp16"), val = tensor<fp16, []>(0x1.bccp-2)];
+            tensor<fp16, [80, 3000]> mul_0_cast_fp16 = mul(x = log_0_cast_fp16, y = mul_0_y_0_to_fp16)[name = tensor<string, []>("mul_0_cast_fp16")];
+            tensor<bool, []> var_44_keep_dims_0 = const()[name = tensor<string, []>("op_44_keep_dims_0"), val = tensor<bool, []>(false)];
+            tensor<fp16, []> var_44_cast_fp16 = reduce_max(keep_dims = var_44_keep_dims_0, x = mul_0_cast_fp16)[name = tensor<string, []>("op_44_cast_fp16")];
+            tensor<fp16, []> var_46_to_fp16 = const()[name = tensor<string, []>("op_46_to_fp16"), val = tensor<fp16, []>(0x1p+3)];
+            tensor<fp16, []> var_47_cast_fp16 = sub(x = var_44_cast_fp16, y = var_46_to_fp16)[name = tensor<string, []>("op_47_cast_fp16")];
+            tensor<fp16, [80, 3000]> log_spec_3_cast_fp16 = maximum(x = mul_0_cast_fp16, y = var_47_cast_fp16)[name = tensor<string, []>("log_spec_3_cast_fp16")];
+            tensor<fp16, []> var_50_to_fp16 = const()[name = tensor<string, []>("op_50_to_fp16"), val = tensor<fp16, []>(0x1p+2)];
+            tensor<fp16, [80, 3000]> var_51_cast_fp16 = add(x = log_spec_3_cast_fp16, y = var_50_to_fp16)[name = tensor<string, []>("op_51_cast_fp16")];
+            tensor<fp16, []> _inversed_log_spec_y_0_to_fp16 = const()[name = tensor<string, []>("_inversed_log_spec_y_0_to_fp16"), val = tensor<fp16, []>(0x1p-2)];
+            tensor<fp16, [80, 3000]> _inversed_log_spec_cast_fp16 = mul(x = var_51_cast_fp16, y = _inversed_log_spec_y_0_to_fp16)[name = tensor<string, []>("_inversed_log_spec_cast_fp16")];
+            tensor<int32, [1]> var_55_axes_0 = const()[name = tensor<string, []>("op_55_axes_0"), val = tensor<int32, [1]>([0])];
+            tensor<fp16, [1, 80, 3000]> var_55_cast_fp16 = expand_dims(axes = var_55_axes_0, x = _inversed_log_spec_cast_fp16)[name = tensor<string, []>("op_55_cast_fp16")];
+            tensor<int32, [1]> var_62_axes_0 = const()[name = tensor<string, []>("op_62_axes_0"), val = tensor<int32, [1]>([2])];
+            tensor<fp16, [1, 80, 1, 3000]> melspectrogram_features = expand_dims(axes = var_62_axes_0, x = var_55_cast_fp16)[name = tensor<string, []>("op_62_cast_fp16")];
+        } -> (melspectrogram_features);
+}

openai_whisper___1/MelSpectrogram.mlmodelc/weights/weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5b65b76f4e1dab57239e3946f6ab1314a7d1fdfa114485683dd04476ca62adb6
+size 354080

openai_whisper___1/TextDecoder.mlmodelc/analytics/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bfbe102ae5fb9368974a077f780441dd222fdfb0c7778c1df227ef6a73cbaada
+size 243

openai_whisper___1/TextDecoder.mlmodelc/coremldata.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:292f96416a33f9a80aaa62ead3dd5206aee6c5e6b3ac6cc02c059d38cbf04c6a
+size 633

openai_whisper___1/TextDecoder.mlmodelc/metadata.json

@@ -0,0 +1,165 @@
+[
+  {
+    "metadataOutputVersion" : "3.0",
+    "storagePrecision" : "Float16",
+    "outputSchema" : [
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 1 × 51865)",
+        "shortDescription" : "",
+        "shape" : "[1, 1, 51865]",
+        "name" : "logits",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 1536 × 1 × 1)",
+        "shortDescription" : "",
+        "shape" : "[1, 1536, 1, 1]",
+        "name" : "key_cache_updates",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 1536 × 1 × 1)",
+        "shortDescription" : "",
+        "shape" : "[1, 1536, 1, 1]",
+        "name" : "value_cache_updates",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 1500)",
+        "shortDescription" : "",
+        "shape" : "[1, 1500]",
+        "name" : "alignment_heads_weights",
+        "type" : "MultiArray"
+      }
+    ],
+    "modelParameters" : [
+
+    ],
+    "specificationVersion" : 7,
+    "mlProgramOperationTypeHistogram" : {
+      "Split" : 2,
+      "Concat" : 3,
+      "Ios16.rsqrt" : 13,
+      "Ios16.mul" : 50,
+      "Squeeze" : 1,
+      "SliceByIndex" : 12,
+      "Ios16.sub" : 14,
+      "Transpose" : 1,
+      "Ios16.conv" : 40,
+      "Ios16.add" : 38,
+      "Ios16.linear" : 1,
+      "Ios16.matmul" : 16,
+      "Ios16.gelu" : 4,
+      "Ios16.reduceMean" : 27,
+      "ExpandDims" : 6,
+      "Ios16.batchNorm" : 13,
+      "Ios16.gather" : 2,
+      "Ios16.reshape" : 32,
+      "Ios16.softmax" : 8
+    },
+    "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==2.2.1",
+      "com.github.apple.coremltools.version" : "7.1"
+    },
+    "inputSchema" : [
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Int32",
+        "formattedType" : "MultiArray (Int32 1)",
+        "shortDescription" : "",
+        "shape" : "[1]",
+        "name" : "input_ids",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Int32",
+        "formattedType" : "MultiArray (Int32 1)",
+        "shortDescription" : "",
+        "shape" : "[1]",
+        "name" : "cache_length",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 1536 × 1 × 224)",
+        "shortDescription" : "",
+        "shape" : "[1, 1536, 1, 224]",
+        "name" : "key_cache",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 1536 × 1 × 224)",
+        "shortDescription" : "",
+        "shape" : "[1, 1536, 1, 224]",
+        "name" : "value_cache",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 224)",
+        "shortDescription" : "",
+        "shape" : "[1, 224]",
+        "name" : "kv_cache_update_mask",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 384 × 1 × 1500)",
+        "shortDescription" : "",
+        "shape" : "[1, 384, 1, 1500]",
+        "name" : "encoder_output_embeds",
+        "type" : "MultiArray"
+      },
+      {
+        "hasShapeFlexibility" : "0",
+        "isOptional" : "0",
+        "dataType" : "Float16",
+        "formattedType" : "MultiArray (Float16 1 × 224)",
+        "shortDescription" : "",
+        "shape" : "[1, 224]",
+        "name" : "decoder_key_padding_mask",
+        "type" : "MultiArray"
+      }
+    ],
+    "generatedClassName" : "TextDecoder",
+    "method" : "predict"
+  }
+]

openai_whisper___1/TextDecoder.mlmodelc/model.mlmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1afdfc3a8f3e8d6afc46e1ecc5fb216eadccbf82d9c568e7dbd3955143a1cd0e
+size 113134

openai_whisper___1/TextDecoder.mlmodelc/weights/weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d0313e1a4ffa88538c141cc3c73e6eb0e3dc54db9d574b21c7c034de688e4951
+size 59216434

openai_whisper___1/config.json

@@ -0,0 +1 @@
+{"_name_or_path": "openai/whisper-tiny", "activation_dropout": 0.0, "activation_function": "gelu", "architectures": ["WhisperForConditionalGeneration"], "attention_dropout": 0.0, "begin_suppress_tokens": [220, 50257], "bos_token_id": 50257, "d_model": 384, "decoder_attention_heads": 6, "decoder_ffn_dim": 1536, "decoder_layerdrop": 0.0, "decoder_layers": 4, "decoder_start_token_id": 50258, "dropout": 0.0, "encoder_attention_heads": 6, "encoder_ffn_dim": 1536, "encoder_layerdrop": 0.0, "encoder_layers": 4, "eos_token_id": 50257, "forced_decoder_ids": [[1, 50259], [2, 50359], [3, 50363]], "init_std": 0.02, "is_encoder_decoder": true, "max_length": 448, "max_source_positions": 1500, "max_target_positions": 448, "model_type": "whisper", "num_hidden_layers": 4, "num_mel_bins": 80, "pad_token_id": 50257, "scale_embedding": false, "suppress_tokens": [1, 2, 7, 8, 9, 10, 14, 25, 26, 27, 28, 29, 31, 58, 59, 60, 61, 62, 63, 90, 91, 92, 93, 359, 503, 522, 542, 873, 893, 902, 918, 922, 931, 1350, 1853, 1982, 2460, 2627, 3246, 3253, 3268, 3536, 3846, 3961, 4183, 4667, 6585, 6647, 7273, 9061, 9383, 10428, 10929, 11938, 12033, 12331, 12562, 13793, 14157, 14635, 15265, 15618, 16553, 16604, 18362, 18956, 20075, 21675, 22520, 26130, 26161, 26435, 28279, 29464, 31650, 32302, 32470, 36865, 42863, 47425, 49870, 50254, 50258, 50358, 50359, 50360, 50361, 50362], "torch_dtype": "float32", "transformers_version": "4.27.0.dev0", "use_cache": true, "vocab_size": 51865}

openai_whisper___1/generation_config.json

@@ -0,0 +1 @@
+{"alignment_heads": [[2, 2], [3, 0], [3, 2], [3, 3], [3, 4], [3, 5]], "begin_suppress_tokens": [220, 50257], "bos_token_id": 50257, "decoder_start_token_id": 50258, "eos_token_id": 50257, "forced_decoder_ids": [[1, null], [2, 50359]], "is_multilingual": true, "lang_to_id": {"<|af|>": 50327, "<|am|>": 50334, "<|ar|>": 50272, "<|as|>": 50350, "<|az|>": 50304, "<|ba|>": 50355, "<|be|>": 50330, "<|bg|>": 50292, "<|bn|>": 50302, "<|bo|>": 50347, "<|br|>": 50309, "<|bs|>": 50315, "<|ca|>": 50270, "<|cs|>": 50283, "<|cy|>": 50297, "<|da|>": 50285, "<|de|>": 50261, "<|el|>": 50281, "<|en|>": 50259, "<|es|>": 50262, "<|et|>": 50307, "<|eu|>": 50310, "<|fa|>": 50300, "<|fi|>": 50277, "<|fo|>": 50338, "<|fr|>": 50265, "<|gl|>": 50319, "<|gu|>": 50333, "<|haw|>": 50352, "<|ha|>": 50354, "<|he|>": 50279, "<|hi|>": 50276, "<|hr|>": 50291, "<|ht|>": 50339, "<|hu|>": 50286, "<|hy|>": 50312, "<|id|>": 50275, "<|is|>": 50311, "<|it|>": 50274, "<|ja|>": 50266, "<|jw|>": 50356, "<|ka|>": 50329, "<|kk|>": 50316, "<|km|>": 50323, "<|kn|>": 50306, "<|ko|>": 50264, "<|la|>": 50294, "<|lb|>": 50345, "<|ln|>": 50353, "<|lo|>": 50336, "<|lt|>": 50293, "<|lv|>": 50301, "<|mg|>": 50349, "<|mi|>": 50295, "<|mk|>": 50308, "<|ml|>": 50296, "<|mn|>": 50314, "<|mr|>": 50320, "<|ms|>": 50282, "<|mt|>": 50343, "<|my|>": 50346, "<|ne|>": 50313, "<|nl|>": 50271, "<|nn|>": 50342, "<|no|>": 50288, "<|oc|>": 50328, "<|pa|>": 50321, "<|pl|>": 50269, "<|ps|>": 50340, "<|pt|>": 50267, "<|ro|>": 50284, "<|ru|>": 50263, "<|sa|>": 50344, "<|sd|>": 50332, "<|si|>": 50322, "<|sk|>": 50298, "<|sl|>": 50305, "<|sn|>": 50324, "<|so|>": 50326, "<|sq|>": 50317, "<|sr|>": 50303, "<|su|>": 50357, "<|sv|>": 50273, "<|sw|>": 50318, "<|ta|>": 50287, "<|te|>": 50299, "<|tg|>": 50331, "<|th|>": 50289, "<|tk|>": 50341, "<|tl|>": 50348, "<|tr|>": 50268, "<|tt|>": 50351, "<|uk|>": 50280, "<|ur|>": 50290, "<|uz|>": 50337, "<|vi|>": 50278, "<|yi|>": 50335, "<|yo|>": 50325, "<|zh|>": 50260}, "max_initial_timestamp_index": 50, "max_length": 448, "no_timestamps_token_id": 50363, "pad_token_id": 50257, "prev_sot_token_id": 50361, "return_timestamps": false, "suppress_tokens": [1, 2, 7, 8, 9, 10, 14, 25, 26, 27, 28, 29, 31, 58, 59, 60, 61, 62, 63, 90, 91, 92, 93, 359, 503, 522, 542, 873, 893, 902, 918, 922, 931, 1350, 1853, 1982, 2460, 2627, 3246, 3253, 3268, 3536, 3846, 3961, 4183, 4667, 6585, 6647, 7273, 9061, 9383, 10428, 10929, 11938, 12033, 12331, 12562, 13793, 14157, 14635, 15265, 15618, 16553, 16604, 18362, 18956, 20075, 21675, 22520, 26130, 26161, 26435, 28279, 29464, 31650, 32302, 32470, 36865, 42863, 47425, 49870, 50254, 50258, 50358, 50359, 50360, 50361, 50362], "task_to_id": {"transcribe": 50359, "translate": 50358}, "transformers_version": "4.31.0.dev0"}