| FazBrowse GitHub Viewer | Trending | | Home |
| Tools: [Download Repo ZIP] [Original HTTPS Page] |
1 parent 209c3b9 commit a4e52ab
22 files changed
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
@@ -11,9 +11,7 @@ Project("{9A19103F-16F7-4668-BE54-9A1E7A4F7556}") = "TensorFlowNET.Core", "src\T | |||
| 11 | 11 | EndProject | |
| 12 | 12 | Project("{9A19103F-16F7-4668-BE54-9A1E7A4F7556}") = "TensorFlowNET.Visualization", "TensorFlowNET.Visualization\TensorFlowNET.Visualization.csproj", "{4BB2ABD1-635E-41E4-B534-CB5B6A2D754D}" | |
| 13 | 13 | EndProject | |
| 14 | - Project("{9A19103F-16F7-4668-BE54-9A1E7A4F7556}") = "NumSharp.Core", "..\NumSharp\src\NumSharp.Core\NumSharp.Core.csproj", "{E8340C61-12C1-4BEE-A340-403E7C1ACD82}" | ||
| 15 | - EndProject | ||
| 16 | - Project("{9A19103F-16F7-4668-BE54-9A1E7A4F7556}") = "scikit-learn", "..\scikit-learn.net\src\scikit-learn\scikit-learn.csproj", "{199DDAD8-4A6F-43B3-A560-C0393619E304}" | ||
| 14 | + Project("{9A19103F-16F7-4668-BE54-9A1E7A4F7556}") = "NumSharp.Core", "..\NumSharp\src\NumSharp.Core\NumSharp.Core.csproj", "{1F4C5683-48B4-4328-8171-E9F7ABAA2E72}" | ||
| 17 | 15 | EndProject | |
| 18 | 16 | Global | |
| 19 | 17 | GlobalSection(SolutionConfigurationPlatforms) = preSolution | |
@@ -37,14 +35,10 @@ Global | |||
| 37 | 35 | {4BB2ABD1-635E-41E4-B534-CB5B6A2D754D}.Debug|Any CPU.Build.0 = Debug|Any CPU | |
| 38 | 36 | {4BB2ABD1-635E-41E4-B534-CB5B6A2D754D}.Release|Any CPU.ActiveCfg = Release|Any CPU | |
| 39 | 37 | {4BB2ABD1-635E-41E4-B534-CB5B6A2D754D}.Release|Any CPU.Build.0 = Release|Any CPU | |
| 40 | - {E8340C61-12C1-4BEE-A340-403E7C1ACD82}.Debug|Any CPU.ActiveCfg = Debug|Any CPU | ||
| 41 | - {E8340C61-12C1-4BEE-A340-403E7C1ACD82}.Debug|Any CPU.Build.0 = Debug|Any CPU | ||
| 42 | - {E8340C61-12C1-4BEE-A340-403E7C1ACD82}.Release|Any CPU.ActiveCfg = Release|Any CPU | ||
| 43 | - {E8340C61-12C1-4BEE-A340-403E7C1ACD82}.Release|Any CPU.Build.0 = Release|Any CPU | ||
| 44 | - {199DDAD8-4A6F-43B3-A560-C0393619E304}.Debug|Any CPU.ActiveCfg = Debug|Any CPU | ||
| 45 | - {199DDAD8-4A6F-43B3-A560-C0393619E304}.Debug|Any CPU.Build.0 = Debug|Any CPU | ||
| 46 | - {199DDAD8-4A6F-43B3-A560-C0393619E304}.Release|Any CPU.ActiveCfg = Release|Any CPU | ||
| 47 | - {199DDAD8-4A6F-43B3-A560-C0393619E304}.Release|Any CPU.Build.0 = Release|Any CPU | ||
| 38 | + {1F4C5683-48B4-4328-8171-E9F7ABAA2E72}.Debug|Any CPU.ActiveCfg = Debug|Any CPU | ||
| 39 | + {1F4C5683-48B4-4328-8171-E9F7ABAA2E72}.Debug|Any CPU.Build.0 = Debug|Any CPU | ||
| 40 | + {1F4C5683-48B4-4328-8171-E9F7ABAA2E72}.Release|Any CPU.ActiveCfg = Release|Any CPU | ||
| 41 | + {1F4C5683-48B4-4328-8171-E9F7ABAA2E72}.Release|Any CPU.Build.0 = Release|Any CPU | ||
| 48 | 42 | EndGlobalSection | |
| 49 | 43 | GlobalSection(SolutionProperties) = preSolution | |
| 50 | 44 | HideSolutionNode = FALSE | |
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
@@ -1,6 +1,7 @@ | |||
| 1 | 1 | using System; | |
| 2 | 2 | using System.Collections.Generic; | |
| 3 | 3 | using System.Text; | |
| 4 | + using Tensorflow.Operations.Initializers; | ||
| 4 | 5 | ||
| 5 | 6 | namespace Tensorflow | |
| 6 | 7 | { | |
@@ -24,128 +25,5 @@ public static variable_scope variable_scope(VariableScope scope, | |||
| 24 | 25 | default_name, | |
| 25 | 26 | values, | |
| 26 | 27 | auxiliary_name_scope); | |
| 27 | - | ||
| 28 | - public class Zeros : IInitializer | ||
| 29 | - { | ||
| 30 | - private TF_DataType dtype; | ||
| 31 | - | ||
| 32 | - public Zeros(TF_DataType dtype = TF_DataType.TF_FLOAT) | ||
| 33 | - { | ||
| 34 | - this.dtype = dtype; | ||
| 35 | - } | ||
| 36 | - | ||
| 37 | - public Tensor call(TensorShape shape, TF_DataType dtype = TF_DataType.DtInvalid) | ||
| 38 | - { | ||
| 39 | - if (dtype == TF_DataType.DtInvalid) | ||
| 40 | - dtype = this.dtype; | ||
| 41 | - | ||
| 42 | - return array_ops.zeros(shape, dtype); | ||
| 43 | - } | ||
| 44 | - | ||
| 45 | - public object get_config() | ||
| 46 | - { | ||
| 47 | - return new { dtype = dtype.name() }; | ||
| 48 | - } | ||
| 49 | - } | ||
| 50 | - | ||
| 51 | - /// <summary> | ||
| 52 | - /// Initializer capable of adapting its scale to the shape of weights tensors. | ||
| 53 | - /// </summary> | ||
| 54 | - public class VarianceScaling : IInitializer | ||
| 55 | - { | ||
| 56 | - protected float _scale; | ||
| 57 | - protected string _mode; | ||
| 58 | - protected string _distribution; | ||
| 59 | - protected int? _seed; | ||
| 60 | - protected TF_DataType _dtype; | ||
| 61 | - | ||
| 62 | - public VarianceScaling(float scale = 1.0f, | ||
| 63 | - string mode = "fan_in", | ||
| 64 | - string distribution= "truncated_normal", | ||
| 65 | - int? seed = null, | ||
| 66 | - TF_DataType dtype = TF_DataType.TF_FLOAT) | ||
| 67 | - { | ||
| 68 | - if (scale < 0) | ||
| 69 | - throw new ValueError("`scale` must be positive float."); | ||
| 70 | - _scale = scale; | ||
| 71 | - _mode = mode; | ||
| 72 | - _distribution = distribution; | ||
| 73 | - _seed = seed; | ||
| 74 | - _dtype = dtype; | ||
| 75 | - } | ||
| 76 | - | ||
| 77 | - public Tensor call(TensorShape shape, TF_DataType dtype) | ||
| 78 | - { | ||
| 79 | - var (fan_in, fan_out) = _compute_fans(shape); | ||
| 80 | - if (_mode == "fan_in") | ||
| 81 | - _scale /= Math.Max(1, fan_in); | ||
| 82 | - else if (_mode == "fan_out") | ||
| 83 | - _scale /= Math.Max(1, fan_out); | ||
| 84 | - else | ||
| 85 | - _scale /= Math.Max(1, (fan_in + fan_out) / 2); | ||
| 86 | - | ||
| 87 | - if (_distribution == "normal" || _distribution == "truncated_normal") | ||
| 88 | - { | ||
| 89 | - throw new NotImplementedException("truncated_normal"); | ||
| 90 | - } | ||
| 91 | - else if(_distribution == "untruncated_normal") | ||
| 92 | - { | ||
| 93 | - throw new NotImplementedException("truncated_normal"); | ||
| 94 | - } | ||
| 95 | - else | ||
| 96 | - { | ||
| 97 | - var limit = Math.Sqrt(3.0f * _scale); | ||
| 98 | - return random_ops.random_uniform(shape, (float)-limit, (float)limit, dtype, seed: _seed); | ||
| 99 | - } | ||
| 100 | - } | ||
| 101 | - | ||
| 102 | - private (int, int) _compute_fans(int[] shape) | ||
| 103 | - { | ||
| 104 | - if (shape.Length < 1) | ||
| 105 | - return (1, 1); | ||
| 106 | - if (shape.Length == 1) | ||
| 107 | - return (shape[0], shape[0]); | ||
| 108 | - if (shape.Length == 2) | ||
| 109 | - return (shape[0], shape[1]); | ||
| 110 | - else | ||
| 111 | - throw new NotImplementedException("VarianceScaling._compute_fans"); | ||
| 112 | - } | ||
| 113 | - | ||
| 114 | - public virtual object get_config() | ||
| 115 | - { | ||
| 116 | - return new | ||
| 117 | - { | ||
| 118 | - scale = _scale, | ||
| 119 | - mode = _mode, | ||
| 120 | - distribution = _distribution, | ||
| 121 | - seed = _seed, | ||
| 122 | - dtype = _dtype | ||
| 123 | - }; | ||
| 124 | - } | ||
| 125 | - } | ||
| 126 | - | ||
| 127 | - public class GlorotUniform : VarianceScaling | ||
| 128 | - { | ||
| 129 | - public GlorotUniform(float scale = 1.0f, | ||
| 130 | - string mode = "fan_avg", | ||
| 131 | - string distribution = "uniform", | ||
| 132 | - int? seed = null, | ||
| 133 | - TF_DataType dtype = TF_DataType.TF_FLOAT) : base(scale, mode, distribution, seed, dtype) | ||
| 134 | - { | ||
| 135 | - | ||
| 136 | - } | ||
| 137 | - | ||
| 138 | - public object get_config() | ||
| 139 | - { | ||
| 140 | - return new | ||
| 141 | - { | ||
| 142 | - scale = _scale, | ||
| 143 | - mode = _mode, | ||
| 144 | - distribution = _distribution, | ||
| 145 | - seed = _seed, | ||
| 146 | - dtype = _dtype | ||
| 147 | - }; | ||
| 148 | - } | ||
| 149 | - } | ||
| 150 | 28 | } | |
| 151 | 29 | } | |
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
@@ -22,5 +22,12 @@ public static Tensor random_normal(int[] shape, | |||
| 22 | 22 | TF_DataType dtype = TF_DataType.TF_FLOAT, | |
| 23 | 23 | int? seed = null, | |
| 24 | 24 | string name = null) => random_ops.random_normal(shape, mean, stddev, dtype, seed, name); | |
| 25 | + | ||
| 26 | + public static Tensor random_uniform(int[] shape, | ||
| 27 | + float minval = 0, | ||
| 28 | + float? maxval = null, | ||
| 29 | + TF_DataType dtype = TF_DataType.TF_FLOAT, | ||
| 30 | + int? seed = null, | ||
| 31 | + string name = null) => random_ops.random_uniform(shape, minval, maxval, dtype, seed, name); | ||
| 25 | 32 | } | |
| 26 | 33 | } | |
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
@@ -0,0 +1,20 @@ | |||
| 1 | + using System; | ||
| 2 | + using System.Collections.Generic; | ||
| 3 | + using System.Text; | ||
| 4 | + using Tensorflow.Operations.Initializers; | ||
| 5 | + | ||
| 6 | + namespace Tensorflow.Keras | ||
| 7 | + { | ||
| 8 | + public class Initializers | ||
| 9 | + { | ||
| 10 | + /// <summary> | ||
| 11 | + /// He normal initializer. | ||
| 12 | + /// </summary> | ||
| 13 | + /// <param name="seed"></param> | ||
| 14 | + /// <returns></returns> | ||
| 15 | + public IInitializer he_normal(int? seed = null) | ||
| 16 | + { | ||
| 17 | + return new VarianceScaling(scale: 20f, mode: "fan_in", distribution: "truncated_normal", seed: seed); | ||
| 18 | + } | ||
| 19 | + } | ||
| 20 | + } | ||
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
@@ -0,0 +1,15 @@ | |||
| 1 | + using System; | ||
| 2 | + using System.Collections.Generic; | ||
| 3 | + using System.Text; | ||
| 4 | + using Tensorflow.Keras; | ||
| 5 | + | ||
| 6 | + namespace Tensorflow | ||
| 7 | + { | ||
| 8 | + public static partial class tf | ||
| 9 | + { | ||
| 10 | + public static class keras | ||
| 11 | + { | ||
| 12 | + public static Initializers initializers => new Initializers(); | ||
| 13 | + } | ||
| 14 | + } | ||
| 15 | + } | ||
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
@@ -0,0 +1,30 @@ | |||
| 1 | + using System; | ||
| 2 | + using System.Collections.Generic; | ||
| 3 | + using System.Text; | ||
| 4 | + | ||
| 5 | + namespace Tensorflow.Operations.Initializers | ||
| 6 | + { | ||
| 7 | + public class GlorotUniform : VarianceScaling | ||
| 8 | + { | ||
| 9 | + public GlorotUniform(float scale = 1.0f, | ||
| 10 | + string mode = "fan_avg", | ||
| 11 | + string distribution = "uniform", | ||
| 12 | + int? seed = null, | ||
| 13 | + TF_DataType dtype = TF_DataType.TF_FLOAT) : base(scale, mode, distribution, seed, dtype) | ||
| 14 | + { | ||
| 15 | + | ||
| 16 | + } | ||
| 17 | + | ||
| 18 | + public object get_config() | ||
| 19 | + { | ||
| 20 | + return new | ||
| 21 | + { | ||
| 22 | + scale = _scale, | ||
| 23 | + mode = _mode, | ||
| 24 | + distribution = _distribution, | ||
| 25 | + seed = _seed, | ||
| 26 | + dtype = _dtype | ||
| 27 | + }; | ||
| 28 | + } | ||
| 29 | + } | ||
| 30 | + } | ||
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
@@ -6,7 +6,7 @@ namespace Tensorflow | |||
| 6 | 6 | { | |
| 7 | 7 | public interface IInitializer | |
| 8 | 8 | { | |
| 9 | - Tensor call(TensorShape shape, TF_DataType dtype); | ||
| 9 | + Tensor call(TensorShape shape, TF_DataType dtype = TF_DataType.DtInvalid); | ||
| 10 | 10 | object get_config(); | |
| 11 | 11 | } | |
| 12 | 12 | } | |
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
@@ -0,0 +1,41 @@ | |||
| 1 | + using System; | ||
| 2 | + using System.Collections.Generic; | ||
| 3 | + using System.Text; | ||
| 4 | + | ||
| 5 | + namespace Tensorflow.Operations.Initializers | ||
| 6 | + { | ||
| 7 | + public class TruncatedNormal : IInitializer | ||
| 8 | + { | ||
| 9 | + private float mean; | ||
| 10 | + private float stddev; | ||
| 11 | + private int? seed; | ||
| 12 | + private TF_DataType dtype; | ||
| 13 | + | ||
| 14 | + public TruncatedNormal(float mean = 0.0f, | ||
| 15 | + float stddev = 1.0f, | ||
| 16 | + int? seed = null, | ||
| 17 | + TF_DataType dtype = TF_DataType.TF_FLOAT) | ||
| 18 | + { | ||
| 19 | + this.mean = mean; | ||
| 20 | + this.stddev = stddev; | ||
| 21 | + this.seed = seed; | ||
| 22 | + this.dtype = dtype; | ||
| 23 | + } | ||
| 24 | + | ||
| 25 | + public Tensor call(TensorShape shape, TF_DataType dtype) | ||
| 26 | + { | ||
| 27 | + throw new NotImplementedException(""); | ||
| 28 | + } | ||
| 29 | + | ||
| 30 | + public object get_config() | ||
| 31 | + { | ||
| 32 | + return new | ||
| 33 | + { | ||
| 34 | + mean = mean, | ||
| 35 | + stddev = stddev, | ||
| 36 | + seed = seed, | ||
| 37 | + dtype = dtype.name() | ||
| 38 | + }; | ||
| 39 | + } | ||
| 40 | + } | ||
| 41 | + } | ||
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
@@ -0,0 +1,82 @@ | |||
| 1 | + using System; | ||
| 2 | + using System.Collections.Generic; | ||
| 3 | + using System.Text; | ||
| 4 | + | ||
| 5 | + namespace Tensorflow.Operations.Initializers | ||
| 6 | + { | ||
| 7 | + /// <summary> | ||
| 8 | + /// Initializer capable of adapting its scale to the shape of weights tensors. | ||
| 9 | + /// </summary> | ||
| 10 | + public class VarianceScaling : IInitializer | ||
| 11 | + { | ||
| 12 | + protected float _scale; | ||
| 13 | + protected string _mode; | ||
| 14 | + protected string _distribution; | ||
| 15 | + protected int? _seed; | ||
| 16 | + protected TF_DataType _dtype; | ||
| 17 | + | ||
| 18 | + public VarianceScaling(float scale = 1.0f, | ||
| 19 | + string mode = "fan_in", | ||
| 20 | + string distribution = "truncated_normal", | ||
| 21 | + int? seed = null, | ||
| 22 | + TF_DataType dtype = TF_DataType.TF_FLOAT) | ||
| 23 | + { | ||
| 24 | + if (scale < 0) | ||
| 25 | + throw new ValueError("`scale` must be positive float."); | ||
| 26 | + _scale = scale; | ||
| 27 | + _mode = mode; | ||
| 28 | + _distribution = distribution; | ||
| 29 | + _seed = seed; | ||
| 30 | + _dtype = dtype; | ||
| 31 | + } | ||
| 32 | + | ||
| 33 | + public Tensor call(TensorShape shape, TF_DataType dtype) | ||
| 34 | + { | ||
| 35 | + var (fan_in, fan_out) = _compute_fans(shape); | ||
| 36 | + if (_mode == "fan_in") | ||
| 37 | + _scale /= Math.Max(1, fan_in); | ||
| 38 | + else if (_mode == "fan_out") | ||
| 39 | + _scale /= Math.Max(1, fan_out); | ||
| 40 | + else | ||
| 41 | + _scale /= Math.Max(1, (fan_in + fan_out) / 2); | ||
| 42 | + | ||
| 43 | + if (_distribution == "normal" || _distribution == "truncated_normal") | ||
| 44 | + { | ||
| 45 | + throw new NotImplementedException("truncated_normal"); | ||
| 46 | + } | ||
| 47 | + else if (_distribution == "untruncated_normal") | ||
| 48 | + { | ||
| 49 | + throw new NotImplementedException("truncated_normal"); | ||
| 50 | + } | ||
| 51 | + else | ||
| 52 | + { | ||
| 53 | + var limit = Math.Sqrt(3.0f * _scale); | ||
| 54 | + return random_ops.random_uniform(shape, (float)-limit, (float)limit, dtype, seed: _seed); | ||
| 55 | + } | ||
| 56 | + } | ||
| 57 | + | ||
| 58 | + private (int, int) _compute_fans(int[] shape) | ||
| 59 | + { | ||
| 60 | + if (shape.Length < 1) | ||
| 61 | + return (1, 1); | ||
| 62 | + if (shape.Length == 1) | ||
| 63 | + return (shape[0], shape[0]); | ||
| 64 | + if (shape.Length == 2) | ||
| 65 | + return (shape[0], shape[1]); | ||
| 66 | + else | ||
| 67 | + throw new NotImplementedException("VarianceScaling._compute_fans"); | ||
| 68 | + } | ||
| 69 | + | ||
| 70 | + public virtual object get_config() | ||
| 71 | + { | ||
| 72 | + return new | ||
| 73 | + { | ||
| 74 | + scale = _scale, | ||
| 75 | + mode = _mode, | ||
| 76 | + distribution = _distribution, | ||
| 77 | + seed = _seed, | ||
| 78 | + dtype = _dtype | ||
| 79 | + }; | ||
| 80 | + } | ||
| 81 | + } | ||
| 82 | + } | ||
| Back | FazBrowse Home | New Git URL |
0 commit comments