Delete folder FsmnVad.mlmodelc with huggingface_hub

FsmnVad.mlmodelc/analytics/coremldata.bin

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:43a93af27ff0d3de4350b046af669f1d19991f6ee00ea75dff62e518e5e176a6
-size 243

FsmnVad.mlmodelc/coremldata.bin

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:075fba34ad2315b9d7d07b99c0dfca42c14fa604e98f7363fff749fab0f1eba1
-size 315

FsmnVad.mlmodelc/model.mil

@@ -1,122 +0,0 @@
-program(1.0)
-[buildInfo = dict<tensor<string, []>, tensor<string, []>>({{"coremlc-component-MIL", "3520.4.1"}, {"coremlc-version", "3520.5.1"}, {"coremltools-component-torch", "2.5.1"}, {"coremltools-source-dialect", "TorchScript"}, {"coremltools-version", "8.3.0"}})]
-{
-    func main<ios17>(tensor<fp32, [1, 3000, 400]> feats) {
-            tensor<int32, []> var_3 = const()[name = tensor<string, []>("op_3"), val = tensor<int32, []>(-1)];
-            tensor<int32, []> var_12 = const()[name = tensor<string, []>("op_12"), val = tensor<int32, []>(2)];
-            tensor<string, []> feats_to_fp16_dtype_0 = const()[name = tensor<string, []>("feats_to_fp16_dtype_0"), val = tensor<string, []>("fp16")];
-            tensor<fp16, [140, 400]> net_in_linear1_linear_weight_to_fp16 = const()[name = tensor<string, []>("net_in_linear1_linear_weight_to_fp16"), val = tensor<fp16, [140, 400]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(64)))];
-            tensor<fp16, [140]> net_in_linear1_linear_bias_to_fp16 = const()[name = tensor<string, []>("net_in_linear1_linear_bias_to_fp16"), val = tensor<fp16, [140]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(112128)))];
-            tensor<fp16, [1, 3000, 400]> feats_to_fp16 = cast(dtype = feats_to_fp16_dtype_0, x = feats)[name = tensor<string, []>("cast_1")];
-            tensor<fp16, [1, 3000, 140]> linear_0_cast_fp16 = linear(bias = net_in_linear1_linear_bias_to_fp16, weight = net_in_linear1_linear_weight_to_fp16, x = feats_to_fp16)[name = tensor<string, []>("linear_0_cast_fp16")];
-            tensor<fp16, [250, 140]> net_in_linear2_linear_weight_to_fp16 = const()[name = tensor<string, []>("net_in_linear2_linear_weight_to_fp16"), val = tensor<fp16, [250, 140]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(112512)))];
-            tensor<fp16, [250]> net_in_linear2_linear_bias_to_fp16 = const()[name = tensor<string, []>("net_in_linear2_linear_bias_to_fp16"), val = tensor<fp16, [250]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(182592)))];
-            tensor<fp16, [1, 3000, 250]> linear_1_cast_fp16 = linear(bias = net_in_linear2_linear_bias_to_fp16, weight = net_in_linear2_linear_weight_to_fp16, x = linear_0_cast_fp16)[name = tensor<string, []>("linear_1_cast_fp16")];
-            tensor<fp16, [1, 3000, 250]> input_5_cast_fp16 = relu(x = linear_1_cast_fp16)[name = tensor<string, []>("input_5_cast_fp16")];
-            tensor<fp16, [128, 250]> net_fsmn_0_linear_linear_weight_to_fp16 = const()[name = tensor<string, []>("net_fsmn_0_linear_linear_weight_to_fp16"), val = tensor<fp16, [128, 250]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(183168)))];
-            tensor<fp16, [128]> linear_2_bias_0_to_fp16 = const()[name = tensor<string, []>("linear_2_bias_0_to_fp16"), val = tensor<fp16, [128]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(247232)))];
-            tensor<fp16, [1, 3000, 128]> linear_2_cast_fp16 = linear(bias = linear_2_bias_0_to_fp16, weight = net_fsmn_0_linear_linear_weight_to_fp16, x = input_5_cast_fp16)[name = tensor<string, []>("linear_2_cast_fp16")];
-            tensor<int32, [1]> x_1_axes_0 = const()[name = tensor<string, []>("x_1_axes_0"), val = tensor<int32, [1]>([1])];
-            tensor<fp16, [1, 1, 3000, 128]> x_1_cast_fp16 = expand_dims(axes = x_1_axes_0, x = linear_2_cast_fp16)[name = tensor<string, []>("x_1_cast_fp16")];
-            tensor<int32, [4]> var_45 = const()[name = tensor<string, []>("op_45"), val = tensor<int32, [4]>([0, 3, 2, 1])];
-            tensor<bool, []> y_left_1_interleave_0 = const()[name = tensor<string, []>("y_left_1_interleave_0"), val = tensor<bool, []>(false)];
-            tensor<fp16, [1, 128, 19, 1]> const_2_to_fp16 = const()[name = tensor<string, []>("const_2_to_fp16"), val = tensor<fp16, [1, 128, 19, 1]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(247552)))];
-            tensor<fp16, [1, 128, 3000, 1]> x_per_1_cast_fp16 = transpose(perm = var_45, x = x_1_cast_fp16)[name = tensor<string, []>("transpose_7")];
-            tensor<fp16, [1, 128, 3019, 1]> y_left_1_cast_fp16 = concat(axis = var_12, interleave = y_left_1_interleave_0, values = (const_2_to_fp16, x_per_1_cast_fp16))[name = tensor<string, []>("y_left_1_cast_fp16")];
-            tensor<string, []> y_left_3_pad_type_0 = const()[name = tensor<string, []>("y_left_3_pad_type_0"), val = tensor<string, []>("valid")];
-            tensor<int32, []> y_left_3_groups_0 = const()[name = tensor<string, []>("y_left_3_groups_0"), val = tensor<int32, []>(128)];
-            tensor<int32, [2]> y_left_3_strides_0 = const()[name = tensor<string, []>("y_left_3_strides_0"), val = tensor<int32, [2]>([1, 1])];
-            tensor<int32, [4]> y_left_3_pad_0 = const()[name = tensor<string, []>("y_left_3_pad_0"), val = tensor<int32, [4]>([0, 0, 0, 0])];
-            tensor<int32, [2]> y_left_3_dilations_0 = const()[name = tensor<string, []>("y_left_3_dilations_0"), val = tensor<int32, [2]>([1, 1])];
-            tensor<fp16, [128, 1, 20, 1]> net_fsmn_0_fsmn_block_conv_left_weight_to_fp16 = const()[name = tensor<string, []>("net_fsmn_0_fsmn_block_conv_left_weight_to_fp16"), val = tensor<fp16, [128, 1, 20, 1]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(252480)))];
-            tensor<fp16, [1, 128, 3000, 1]> y_left_3_cast_fp16 = conv(dilations = y_left_3_dilations_0, groups = y_left_3_groups_0, pad = y_left_3_pad_0, pad_type = y_left_3_pad_type_0, strides = y_left_3_strides_0, weight = net_fsmn_0_fsmn_block_conv_left_weight_to_fp16, x = y_left_1_cast_fp16)[name = tensor<string, []>("y_left_3_cast_fp16")];
-            tensor<fp16, [1, 128, 3000, 1]> out_1_cast_fp16 = add(x = x_per_1_cast_fp16, y = y_left_3_cast_fp16)[name = tensor<string, []>("out_1_cast_fp16")];
-            tensor<int32, [4]> var_57 = const()[name = tensor<string, []>("op_57"), val = tensor<int32, [4]>([0, 3, 2, 1])];
-            tensor<int32, [1]> input_7_axes_0 = const()[name = tensor<string, []>("input_7_axes_0"), val = tensor<int32, [1]>([1])];
-            tensor<fp16, [1, 1, 3000, 128]> out_per_1_cast_fp16 = transpose(perm = var_57, x = out_1_cast_fp16)[name = tensor<string, []>("transpose_6")];
-            tensor<fp16, [1, 3000, 128]> input_7_cast_fp16 = squeeze(axes = input_7_axes_0, x = out_per_1_cast_fp16)[name = tensor<string, []>("input_7_cast_fp16")];
-            tensor<fp16, [250, 128]> net_fsmn_0_affine_linear_weight_to_fp16 = const()[name = tensor<string, []>("net_fsmn_0_affine_linear_weight_to_fp16"), val = tensor<fp16, [250, 128]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(257664)))];
-            tensor<fp16, [250]> net_fsmn_0_affine_linear_bias_to_fp16 = const()[name = tensor<string, []>("net_fsmn_0_affine_linear_bias_to_fp16"), val = tensor<fp16, [250]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(321728)))];
-            tensor<fp16, [1, 3000, 250]> linear_3_cast_fp16 = linear(bias = net_fsmn_0_affine_linear_bias_to_fp16, weight = net_fsmn_0_affine_linear_weight_to_fp16, x = input_7_cast_fp16)[name = tensor<string, []>("linear_3_cast_fp16")];
-            tensor<fp16, [1, 3000, 250]> input_11_cast_fp16 = relu(x = linear_3_cast_fp16)[name = tensor<string, []>("input_11_cast_fp16")];
-            tensor<fp16, [128, 250]> net_fsmn_1_linear_linear_weight_to_fp16 = const()[name = tensor<string, []>("net_fsmn_1_linear_linear_weight_to_fp16"), val = tensor<fp16, [128, 250]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(322304)))];
-            tensor<fp16, [1, 3000, 128]> linear_4_cast_fp16 = linear(bias = linear_2_bias_0_to_fp16, weight = net_fsmn_1_linear_linear_weight_to_fp16, x = input_11_cast_fp16)[name = tensor<string, []>("linear_4_cast_fp16")];
-            tensor<int32, [1]> x_3_axes_0 = const()[name = tensor<string, []>("x_3_axes_0"), val = tensor<int32, [1]>([1])];
-            tensor<fp16, [1, 1, 3000, 128]> x_3_cast_fp16 = expand_dims(axes = x_3_axes_0, x = linear_4_cast_fp16)[name = tensor<string, []>("x_3_cast_fp16")];
-            tensor<int32, [4]> var_77 = const()[name = tensor<string, []>("op_77"), val = tensor<int32, [4]>([0, 3, 2, 1])];
-            tensor<bool, []> y_left_5_interleave_0 = const()[name = tensor<string, []>("y_left_5_interleave_0"), val = tensor<bool, []>(false)];
-            tensor<fp16, [1, 128, 3000, 1]> x_per_3_cast_fp16 = transpose(perm = var_77, x = x_3_cast_fp16)[name = tensor<string, []>("transpose_5")];
-            tensor<fp16, [1, 128, 3019, 1]> y_left_5_cast_fp16 = concat(axis = var_12, interleave = y_left_5_interleave_0, values = (const_2_to_fp16, x_per_3_cast_fp16))[name = tensor<string, []>("y_left_5_cast_fp16")];
-            tensor<string, []> y_left_7_pad_type_0 = const()[name = tensor<string, []>("y_left_7_pad_type_0"), val = tensor<string, []>("valid")];
-            tensor<int32, []> y_left_7_groups_0 = const()[name = tensor<string, []>("y_left_7_groups_0"), val = tensor<int32, []>(128)];
-            tensor<int32, [2]> y_left_7_strides_0 = const()[name = tensor<string, []>("y_left_7_strides_0"), val = tensor<int32, [2]>([1, 1])];
-            tensor<int32, [4]> y_left_7_pad_0 = const()[name = tensor<string, []>("y_left_7_pad_0"), val = tensor<int32, [4]>([0, 0, 0, 0])];
-            tensor<int32, [2]> y_left_7_dilations_0 = const()[name = tensor<string, []>("y_left_7_dilations_0"), val = tensor<int32, [2]>([1, 1])];
-            tensor<fp16, [128, 1, 20, 1]> net_fsmn_1_fsmn_block_conv_left_weight_to_fp16 = const()[name = tensor<string, []>("net_fsmn_1_fsmn_block_conv_left_weight_to_fp16"), val = tensor<fp16, [128, 1, 20, 1]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(386368)))];
-            tensor<fp16, [1, 128, 3000, 1]> y_left_7_cast_fp16 = conv(dilations = y_left_7_dilations_0, groups = y_left_7_groups_0, pad = y_left_7_pad_0, pad_type = y_left_7_pad_type_0, strides = y_left_7_strides_0, weight = net_fsmn_1_fsmn_block_conv_left_weight_to_fp16, x = y_left_5_cast_fp16)[name = tensor<string, []>("y_left_7_cast_fp16")];
-            tensor<fp16, [1, 128, 3000, 1]> out_3_cast_fp16 = add(x = x_per_3_cast_fp16, y = y_left_7_cast_fp16)[name = tensor<string, []>("out_3_cast_fp16")];
-            tensor<int32, [4]> var_89 = const()[name = tensor<string, []>("op_89"), val = tensor<int32, [4]>([0, 3, 2, 1])];
-            tensor<int32, [1]> input_13_axes_0 = const()[name = tensor<string, []>("input_13_axes_0"), val = tensor<int32, [1]>([1])];
-            tensor<fp16, [1, 1, 3000, 128]> out_per_3_cast_fp16 = transpose(perm = var_89, x = out_3_cast_fp16)[name = tensor<string, []>("transpose_4")];
-            tensor<fp16, [1, 3000, 128]> input_13_cast_fp16 = squeeze(axes = input_13_axes_0, x = out_per_3_cast_fp16)[name = tensor<string, []>("input_13_cast_fp16")];
-            tensor<fp16, [250, 128]> net_fsmn_1_affine_linear_weight_to_fp16 = const()[name = tensor<string, []>("net_fsmn_1_affine_linear_weight_to_fp16"), val = tensor<fp16, [250, 128]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(391552)))];
-            tensor<fp16, [250]> net_fsmn_1_affine_linear_bias_to_fp16 = const()[name = tensor<string, []>("net_fsmn_1_affine_linear_bias_to_fp16"), val = tensor<fp16, [250]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(455616)))];
-            tensor<fp16, [1, 3000, 250]> linear_5_cast_fp16 = linear(bias = net_fsmn_1_affine_linear_bias_to_fp16, weight = net_fsmn_1_affine_linear_weight_to_fp16, x = input_13_cast_fp16)[name = tensor<string, []>("linear_5_cast_fp16")];
-            tensor<fp16, [1, 3000, 250]> input_17_cast_fp16 = relu(x = linear_5_cast_fp16)[name = tensor<string, []>("input_17_cast_fp16")];
-            tensor<fp16, [128, 250]> net_fsmn_2_linear_linear_weight_to_fp16 = const()[name = tensor<string, []>("net_fsmn_2_linear_linear_weight_to_fp16"), val = tensor<fp16, [128, 250]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(456192)))];
-            tensor<fp16, [1, 3000, 128]> linear_6_cast_fp16 = linear(bias = linear_2_bias_0_to_fp16, weight = net_fsmn_2_linear_linear_weight_to_fp16, x = input_17_cast_fp16)[name = tensor<string, []>("linear_6_cast_fp16")];
-            tensor<int32, [1]> x_5_axes_0 = const()[name = tensor<string, []>("x_5_axes_0"), val = tensor<int32, [1]>([1])];
-            tensor<fp16, [1, 1, 3000, 128]> x_5_cast_fp16 = expand_dims(axes = x_5_axes_0, x = linear_6_cast_fp16)[name = tensor<string, []>("x_5_cast_fp16")];
-            tensor<int32, [4]> var_109 = const()[name = tensor<string, []>("op_109"), val = tensor<int32, [4]>([0, 3, 2, 1])];
-            tensor<bool, []> y_left_9_interleave_0 = const()[name = tensor<string, []>("y_left_9_interleave_0"), val = tensor<bool, []>(false)];
-            tensor<fp16, [1, 128, 3000, 1]> x_per_5_cast_fp16 = transpose(perm = var_109, x = x_5_cast_fp16)[name = tensor<string, []>("transpose_3")];
-            tensor<fp16, [1, 128, 3019, 1]> y_left_9_cast_fp16 = concat(axis = var_12, interleave = y_left_9_interleave_0, values = (const_2_to_fp16, x_per_5_cast_fp16))[name = tensor<string, []>("y_left_9_cast_fp16")];
-            tensor<string, []> y_left_11_pad_type_0 = const()[name = tensor<string, []>("y_left_11_pad_type_0"), val = tensor<string, []>("valid")];
-            tensor<int32, []> y_left_11_groups_0 = const()[name = tensor<string, []>("y_left_11_groups_0"), val = tensor<int32, []>(128)];
-            tensor<int32, [2]> y_left_11_strides_0 = const()[name = tensor<string, []>("y_left_11_strides_0"), val = tensor<int32, [2]>([1, 1])];
-            tensor<int32, [4]> y_left_11_pad_0 = const()[name = tensor<string, []>("y_left_11_pad_0"), val = tensor<int32, [4]>([0, 0, 0, 0])];
-            tensor<int32, [2]> y_left_11_dilations_0 = const()[name = tensor<string, []>("y_left_11_dilations_0"), val = tensor<int32, [2]>([1, 1])];
-            tensor<fp16, [128, 1, 20, 1]> net_fsmn_2_fsmn_block_conv_left_weight_to_fp16 = const()[name = tensor<string, []>("net_fsmn_2_fsmn_block_conv_left_weight_to_fp16"), val = tensor<fp16, [128, 1, 20, 1]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(520256)))];
-            tensor<fp16, [1, 128, 3000, 1]> y_left_11_cast_fp16 = conv(dilations = y_left_11_dilations_0, groups = y_left_11_groups_0, pad = y_left_11_pad_0, pad_type = y_left_11_pad_type_0, strides = y_left_11_strides_0, weight = net_fsmn_2_fsmn_block_conv_left_weight_to_fp16, x = y_left_9_cast_fp16)[name = tensor<string, []>("y_left_11_cast_fp16")];
-            tensor<fp16, [1, 128, 3000, 1]> out_5_cast_fp16 = add(x = x_per_5_cast_fp16, y = y_left_11_cast_fp16)[name = tensor<string, []>("out_5_cast_fp16")];
-            tensor<int32, [4]> var_121 = const()[name = tensor<string, []>("op_121"), val = tensor<int32, [4]>([0, 3, 2, 1])];
-            tensor<int32, [1]> input_19_axes_0 = const()[name = tensor<string, []>("input_19_axes_0"), val = tensor<int32, [1]>([1])];
-            tensor<fp16, [1, 1, 3000, 128]> out_per_5_cast_fp16 = transpose(perm = var_121, x = out_5_cast_fp16)[name = tensor<string, []>("transpose_2")];
-            tensor<fp16, [1, 3000, 128]> input_19_cast_fp16 = squeeze(axes = input_19_axes_0, x = out_per_5_cast_fp16)[name = tensor<string, []>("input_19_cast_fp16")];
-            tensor<fp16, [250, 128]> net_fsmn_2_affine_linear_weight_to_fp16 = const()[name = tensor<string, []>("net_fsmn_2_affine_linear_weight_to_fp16"), val = tensor<fp16, [250, 128]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(525440)))];
-            tensor<fp16, [250]> net_fsmn_2_affine_linear_bias_to_fp16 = const()[name = tensor<string, []>("net_fsmn_2_affine_linear_bias_to_fp16"), val = tensor<fp16, [250]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(589504)))];
-            tensor<fp16, [1, 3000, 250]> linear_7_cast_fp16 = linear(bias = net_fsmn_2_affine_linear_bias_to_fp16, weight = net_fsmn_2_affine_linear_weight_to_fp16, x = input_19_cast_fp16)[name = tensor<string, []>("linear_7_cast_fp16")];
-            tensor<fp16, [1, 3000, 250]> input_23_cast_fp16 = relu(x = linear_7_cast_fp16)[name = tensor<string, []>("input_23_cast_fp16")];
-            tensor<fp16, [128, 250]> net_fsmn_3_linear_linear_weight_to_fp16 = const()[name = tensor<string, []>("net_fsmn_3_linear_linear_weight_to_fp16"), val = tensor<fp16, [128, 250]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(590080)))];
-            tensor<fp16, [1, 3000, 128]> linear_8_cast_fp16 = linear(bias = linear_2_bias_0_to_fp16, weight = net_fsmn_3_linear_linear_weight_to_fp16, x = input_23_cast_fp16)[name = tensor<string, []>("linear_8_cast_fp16")];
-            tensor<int32, [1]> x_axes_0 = const()[name = tensor<string, []>("x_axes_0"), val = tensor<int32, [1]>([1])];
-            tensor<fp16, [1, 1, 3000, 128]> x_cast_fp16 = expand_dims(axes = x_axes_0, x = linear_8_cast_fp16)[name = tensor<string, []>("x_cast_fp16")];
-            tensor<int32, [4]> var_141 = const()[name = tensor<string, []>("op_141"), val = tensor<int32, [4]>([0, 3, 2, 1])];
-            tensor<bool, []> y_left_13_interleave_0 = const()[name = tensor<string, []>("y_left_13_interleave_0"), val = tensor<bool, []>(false)];
-            tensor<fp16, [1, 128, 3000, 1]> x_per_cast_fp16 = transpose(perm = var_141, x = x_cast_fp16)[name = tensor<string, []>("transpose_1")];
-            tensor<fp16, [1, 128, 3019, 1]> y_left_13_cast_fp16 = concat(axis = var_12, interleave = y_left_13_interleave_0, values = (const_2_to_fp16, x_per_cast_fp16))[name = tensor<string, []>("y_left_13_cast_fp16")];
-            tensor<string, []> y_left_pad_type_0 = const()[name = tensor<string, []>("y_left_pad_type_0"), val = tensor<string, []>("valid")];
-            tensor<int32, []> y_left_groups_0 = const()[name = tensor<string, []>("y_left_groups_0"), val = tensor<int32, []>(128)];
-            tensor<int32, [2]> y_left_strides_0 = const()[name = tensor<string, []>("y_left_strides_0"), val = tensor<int32, [2]>([1, 1])];
-            tensor<int32, [4]> y_left_pad_0 = const()[name = tensor<string, []>("y_left_pad_0"), val = tensor<int32, [4]>([0, 0, 0, 0])];
-            tensor<int32, [2]> y_left_dilations_0 = const()[name = tensor<string, []>("y_left_dilations_0"), val = tensor<int32, [2]>([1, 1])];
-            tensor<fp16, [128, 1, 20, 1]> net_fsmn_3_fsmn_block_conv_left_weight_to_fp16 = const()[name = tensor<string, []>("net_fsmn_3_fsmn_block_conv_left_weight_to_fp16"), val = tensor<fp16, [128, 1, 20, 1]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(654144)))];
-            tensor<fp16, [1, 128, 3000, 1]> y_left_cast_fp16 = conv(dilations = y_left_dilations_0, groups = y_left_groups_0, pad = y_left_pad_0, pad_type = y_left_pad_type_0, strides = y_left_strides_0, weight = net_fsmn_3_fsmn_block_conv_left_weight_to_fp16, x = y_left_13_cast_fp16)[name = tensor<string, []>("y_left_cast_fp16")];
-            tensor<fp16, [1, 128, 3000, 1]> out_cast_fp16 = add(x = x_per_cast_fp16, y = y_left_cast_fp16)[name = tensor<string, []>("out_cast_fp16")];
-            tensor<int32, [4]> var_153 = const()[name = tensor<string, []>("op_153"), val = tensor<int32, [4]>([0, 3, 2, 1])];
-            tensor<int32, [1]> input_25_axes_0 = const()[name = tensor<string, []>("input_25_axes_0"), val = tensor<int32, [1]>([1])];
-            tensor<fp16, [1, 1, 3000, 128]> out_per_cast_fp16 = transpose(perm = var_153, x = out_cast_fp16)[name = tensor<string, []>("transpose_0")];
-            tensor<fp16, [1, 3000, 128]> input_25_cast_fp16 = squeeze(axes = input_25_axes_0, x = out_per_cast_fp16)[name = tensor<string, []>("input_25_cast_fp16")];
-            tensor<fp16, [250, 128]> net_fsmn_3_affine_linear_weight_to_fp16 = const()[name = tensor<string, []>("net_fsmn_3_affine_linear_weight_to_fp16"), val = tensor<fp16, [250, 128]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(659328)))];
-            tensor<fp16, [250]> net_fsmn_3_affine_linear_bias_to_fp16 = const()[name = tensor<string, []>("net_fsmn_3_affine_linear_bias_to_fp16"), val = tensor<fp16, [250]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(723392)))];
-            tensor<fp16, [1, 3000, 250]> linear_9_cast_fp16 = linear(bias = net_fsmn_3_affine_linear_bias_to_fp16, weight = net_fsmn_3_affine_linear_weight_to_fp16, x = input_25_cast_fp16)[name = tensor<string, []>("linear_9_cast_fp16")];
-            tensor<fp16, [1, 3000, 250]> input_29_cast_fp16 = relu(x = linear_9_cast_fp16)[name = tensor<string, []>("input_29_cast_fp16")];
-            tensor<fp16, [140, 250]> net_out_linear1_linear_weight_to_fp16 = const()[name = tensor<string, []>("net_out_linear1_linear_weight_to_fp16"), val = tensor<fp16, [140, 250]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(723968)))];
-            tensor<fp16, [140]> net_out_linear1_linear_bias_to_fp16 = const()[name = tensor<string, []>("net_out_linear1_linear_bias_to_fp16"), val = tensor<fp16, [140]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(794048)))];
-            tensor<fp16, [1, 3000, 140]> linear_10_cast_fp16 = linear(bias = net_out_linear1_linear_bias_to_fp16, weight = net_out_linear1_linear_weight_to_fp16, x = input_29_cast_fp16)[name = tensor<string, []>("linear_10_cast_fp16")];
-            tensor<fp16, [248, 140]> net_out_linear2_linear_weight_to_fp16 = const()[name = tensor<string, []>("net_out_linear2_linear_weight_to_fp16"), val = tensor<fp16, [248, 140]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(794432)))];
-            tensor<fp16, [248]> net_out_linear2_linear_bias_to_fp16 = const()[name = tensor<string, []>("net_out_linear2_linear_bias_to_fp16"), val = tensor<fp16, [248]>(BLOBFILE(path = tensor<string, []>("@model_path/weights/weight.bin"), offset = tensor<uint64, []>(863936)))];
-            tensor<fp16, [1, 3000, 248]> linear_11_cast_fp16 = linear(bias = net_out_linear2_linear_bias_to_fp16, weight = net_out_linear2_linear_weight_to_fp16, x = linear_10_cast_fp16)[name = tensor<string, []>("linear_11_cast_fp16")];
-            tensor<fp16, [1, 3000, 248]> scores = softmax(axis = var_3, x = linear_11_cast_fp16)[name = tensor<string, []>("op_169_cast_fp16")];
-        } -> (scores);
-}

FsmnVad.mlmodelc/weights/weight.bin

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:8b115e3a6ff7b89778fe9c2cd6d65f739ab64438b8ce1eb18de056707898b3ec
-size 864496