public BBoxArea(int axis = -1, string fmt = "corner", string prefix = "", ParameterDict @params = null) : base(prefix, @params)
{
if (fmt.ToLower() == "corner")
{
this._pre = new BBoxCornerToCenter(split: true);
}
else if (fmt.ToLower() == "center")
{
this._pre = new BBoxSplit(axis: axis);
}
else
{
throw new Exception($"Unsupported format: {fmt}. Use 'corner' or 'center'.");
}
}
public GRUCell(int hidden_size, string i2h_weight_initializer = null, string h2h_weight_initializer = null,
string i2h_bias_initializer = "zeros", string h2h_bias_initializer = "zeros", int input_size = 0,
string prefix = null, ParameterDict @params = null) : base(prefix, @params)
{
_hidden_size = hidden_size;
_input_size = input_size;
this["i2h_weight"] = Params.Get("i2h_weight", shape: new Shape(hidden_size, input_size),
init: Initializer.Get(i2h_weight_initializer), allow_deferred_init: true);
this["h2h_weight"] = Params.Get("h2h_weight", shape: new Shape(hidden_size, hidden_size),
init: Initializer.Get(h2h_weight_initializer), allow_deferred_init: true);
this["i2h_bias"] = Params.Get("i2h_bias", shape: new Shape(hidden_size),
init: Initializer.Get(i2h_bias_initializer), allow_deferred_init: true);
this["h2h_bias"] = Params.Get("h2h_bias", shape: new Shape(hidden_size),
init: Initializer.Get(h2h_bias_initializer), allow_deferred_init: true);
}
public BasicBlockV2(int channels, int stride, bool downsample = false, int in_channels = 0,
string prefix = null, ParameterDict @params = null) : base(prefix, @params)
{
bn1 = new BatchNorm();
conv1 = ResNet.Conv3x3(channels, stride, in_channels);
bn2 = new BatchNorm();
conv2 = ResNet.Conv3x3(channels, 1, in_channels);
if (downsample)
{
ds = new Conv2D(channels, (1, 1), (stride, stride), use_bias: false, in_channels: in_channels);
}
else
{
ds = null;
}
}
public DefaultPreprocess(string prefix = "", ParameterDict @params = null) : base(prefix, @params)
{
var mean = nd.Array(new float[] {
123.675f,
116.28f,
103.53f
}).Reshape(1, 1, 1, 3);
var scale = nd.Array(new float[] {
58.395f,
57.12f,
57.375f
}).Reshape(1, 1, 1, 3);
this["init_mean"] = Params.GetConstant("init_mean", mean);
this["init_scale"] = Params.GetConstant("init_scale", scale);
}
public Dense(int units, ActivationType?activation = null, bool use_bias = true, bool flatten = true,
DType dtype = null, Initializer weight_initializer = null, string bias_initializer = "zeros",
int in_units = 0, string prefix = null, ParameterDict @params = null) : base(prefix, @params)
{
Units = units;
Act = activation != null ? new Activation(activation.Value) : null;
UseBias = use_bias;
Flatten_ = flatten;
DataType = dtype;
this["weight"] = Params.Get("weight", OpGradReq.Write, new Shape(units, in_units), dtype,
init: weight_initializer, allow_deferred_init: true);
if (UseBias)
{
this["bias"] = Params.Get("bias", OpGradReq.Write, new Shape(units), dtype,
init: Initializer.Get(bias_initializer), allow_deferred_init: true);
}
}
public LSTMCell(int hidden_size, string activation = "tanh", string recurrent_activation = "sigmoid",
string i2h_weight_initializer = null, string h2h_weight_initializer = null,
string i2h_bias_initializer = "zeros", string h2h_bias_initializer = "zeros", int input_size = 0,
string prefix = null, ParameterDict @params = null) : base(prefix, @params)
{
_hidden_size = hidden_size;
_activation = activation;
_input_size = input_size;
_recurrent_activation = recurrent_activation;
this["i2h_weight"] = Params.Get("i2h_weight", shape: new Shape(hidden_size, input_size),
init: Initializer.Get(i2h_weight_initializer), allow_deferred_init: true);
this["h2h_weight"] = Params.Get("h2h_weight", shape: new Shape(hidden_size, hidden_size),
init: Initializer.Get(h2h_weight_initializer), allow_deferred_init: true);
this["i2h_bias"] = Params.Get("i2h_bias", shape: new Shape(hidden_size),
init: Initializer.Get(i2h_bias_initializer), allow_deferred_init: true);
this["h2h_bias"] = Params.Get("h2h_bias", shape: new Shape(hidden_size),
init: Initializer.Get(h2h_bias_initializer), allow_deferred_init: true);
}
public CTCLoss(string layout = "NTC", string label_layout = "NT", float?weight = null, int?batch_axis = 0,
string prefix = "", ParameterDict @params = null) : base(weight, batch_axis, prefix, @params)
{
if (layout != "NTC" && layout != "TNC")
{
throw new ArgumentException($"Only 'NTC' and 'TNC' layouts for pred are supported. Got: {layout}");
}
if (label_layout != "NT" && label_layout != "TN")
{
throw new ArgumentException(
$"Only 'NTC' and 'TNC' layouts for label are supported. Got: {label_layout}");
}
Layout = layout;
LabelLayout = label_layout;
BatchAxis = label_layout.ToCharArray().ToList().IndexOf('N');
}
public static BlockScope Create(string prefix, ParameterDict @params, string hint)
{
// Expr
var current = Getattr(BlockScope._current, "value", null);
if (IsTrue((current is null)))
{
if (IsTrue((prefix is null)))
{
if (IsTrue((!IsTrue(Hasattr(_name.NameManager._current, "value")))))
{
_name.NameManager._current.Value = _name.NameManager();
}
prefix = (_name.NameManager._current.Value.Get(null, hint) + "_");
}
if (IsTrue((@params is null)))
{
@params = new ParameterDict(prefix);
}
public ResNetV2(string block, int[] layers, int[] channels, int classes = 1000, bool thumbnail = false,
string prefix = "", ParameterDict @params = null) : base(prefix, @params)
{
if (layers.Length != channels.Length - 1)
{
throw new Exception("layers.length should be equal to channels.length - 1");
}
Features = new HybridSequential("features");
Features.Add(new BatchNorm(scale: false, center: false));
if (thumbnail)
{
Features.Add(ResNet.Conv3x3(channels[0], 1, 0));
}
else
{
Features.Add(new Conv2D(channels[0], (7, 7), (2, 2), (3, 3), use_bias: false));
Features.Add(new BatchNorm());
Features.Add(new Activation(ActivationType.Relu));
Features.Add(new MaxPool2D((3, 3), (2, 2), (1, 1)));
}
var in_channels = channels[0];
for (var i = 0; i < layers.Length; i++)
{
var stride = i == 0 ? 1 : 2;
var num_layer = layers[i];
Features.Add(MakeLayer(block, num_layer, channels[i + 1], stride, i + 1, in_channels));
in_channels = channels[i + 1];
}
Features.Add(new BatchNorm());
Features.Add(new Activation(ActivationType.Relu));
Features.Add(new GlobalAvgPool2D());
Features.Add(new Flatten());
Output = new Dense(classes, in_units: in_channels);
RegisterChild(Features, "features");
RegisterChild(Output, "output");
}
/// <summary>
/// Start Computing. Is called while rendering an animation.
/// </summary>
public void Run(int updateSteps)
{
_currentProgress = 0;
_master.Progress(_currentProgress);
System.Diagnostics.Debug.WriteLine("PaintJob.Run " + updateSteps.ToString());
_parameters = ParameterDict.Current.Clone();
_updateSteps = updateSteps;
_currentProgressd = 100.0 / (double)(_updateSteps);
for (int i = 0; i < _updateSteps; i++)
{
if (_abort)
{
return;
}
_iterate = new Iterate(_parameters, this, false);
if (_lastIterate != null)
{
_iterate.SetOldData(_lastIterate.GraphicInfo, _lastIterate.PictureData, i);
}
if (_abort)
{
return;
}
_iterate.StartAsync();
_iterate.Wait();
if (_abort)
{
return;
}
_lastIterate = _iterate;
_currentProgress += _currentProgressd;
_master.Progress(_currentProgress);
}
Renderer renderer = PictureArtFactory.Create(_iterate.PictureData, _iterate.LastUsedFormulas, ParameterDict.Current.Clone());
renderer.Paint(_graphics);
if (_abort)
{
return;
}
_master.Progress(0);
}
public BottleneckV2(int channels, int stride, bool downsample = false, int in_channels = 0,
string prefix = null, ParameterDict @params = null) : base(prefix, @params)
{
var channel_one_fourth = Convert.ToInt32(channels / 4);
bn1 = new BatchNorm();
conv1 = new Conv2D(channel_one_fourth, (1, 1), (1, 1), use_bias: false);
bn2 = new BatchNorm();
conv2 = ResNet.Conv3x3(channel_one_fourth, stride, channel_one_fourth);
bn3 = new BatchNorm();
conv3 = new Conv2D(channels, (1, 1), (stride, stride), use_bias: false, in_channels: in_channels);
if (downsample)
{
ds = new Conv2D(channels, (1, 1), (stride, stride), use_bias: false, in_channels: in_channels);
}
else
{
ds = null;
}
}
public BasicBlockV1(int channels, int stride, bool downsample = false, int in_channels = 0,
string prefix = null, ParameterDict @params = null) : base(prefix, @params)
{
body = new HybridSequential("");
body.Add(ResNet.Conv3x3(channels, stride, in_channels));
body.Add(new BatchNorm());
body.Add(new Activation(ActivationType.Relu));
body.Add(ResNet.Conv3x3(channels, 1, in_channels));
body.Add(new BatchNorm());
if (downsample)
{
ds = new HybridSequential();
ds.Add(new Conv2D(channels, (1, 1), (stride, stride), use_bias: false, in_channels: in_channels));
ds.Add(new BatchNorm());
}
else
{
ds = null;
}
}
public BatchNorm(int axis = 1, float momentum = 0.9f, float epsilon = 1e-5f, bool center = true, bool scale = true,
bool use_global_stats = false, string beta_initializer = "zeros", string gamma_initializer = "ones",
string running_mean_initializer = "zeros", string running_variance_initializer = "ones",
int in_channels = 0, string prefix = null, ParameterDict @params = null) : base(prefix, @params)
{
Axis = axis;
Momentum = momentum;
Epsilon = epsilon;
Center = center;
Scale = scale;
Use_Global_Stats = use_global_stats;
In_Channels = in_channels;
this["gamma"] = Params.Get("gamma", scale ? OpGradReq.Write : OpGradReq.Null, new Shape(in_channels),
init: Initializer.Get(gamma_initializer), allow_deferred_init: true, differentiable: scale);
this["beta"] = Params.Get("beta", center ? OpGradReq.Write : OpGradReq.Null, new Shape(in_channels),
init: Initializer.Get(beta_initializer), allow_deferred_init: true, differentiable: center);
this["running_mean"] = Params.Get("running_mean", OpGradReq.Null, new Shape(in_channels),
init: Initializer.Get(running_mean_initializer), allow_deferred_init: true, differentiable: false);
this["running_var"] = Params.Get("running_var", OpGradReq.Null, new Shape(in_channels),
init: Initializer.Get(running_variance_initializer), allow_deferred_init: true, differentiable: false);
}
public _Conv(int channels, int[] kernel_size, int[] strides, int[] padding, int[] dilation,
int groups, string layout, int in_channels = 0, ActivationType?activation = null, bool use_bias = true,
Initializer weight_initializer = null, string bias_initializer = "zeros", int[] adj = null,
string op_name = "Convolution", string prefix = "", ParameterDict @params = null) : base(prefix, @params)
{
NumFilter = channels;
InChannels = in_channels;
Strides = strides;
Padding = padding;
Dialation = dilation;
_op_name = op_name;
KernalSize = kernel_size;
NumGroup = groups;
Layout = layout;
Activation = activation.HasValue ? new Activation(activation.Value) : null;
UseBias = use_bias;
WeightInitializer = weight_initializer;
BiasInitializer = bias_initializer;
Adj = adj;
var dshape = new int[kernel_size.Length + 2];
dshape[layout.ToCharArray().ToList().IndexOf('N')] = 1;
dshape[layout.ToCharArray().ToList().IndexOf('C')] = in_channels;
var wshapes = _infer_weight_shape(op_name, new Shape(dshape));
this["weight"] = Params.Get("weight", OpGradReq.Write, wshapes[1], init: weight_initializer,
allow_deferred_init: true);
if (UseBias)
{
this["bias"] = Params.Get("bias", OpGradReq.Write, wshapes[2], init: Initializer.Get(bias_initializer),
allow_deferred_init: true);
}
else
{
this["bias"] = null;
}
}
void RemoveUnusedParameters(string tomoSource)
{
lock (lockRemoveUnusedParameters)
{
try
{
if (tomoSource.Contains("SetParameterBulk"))
{
return;
}
Dictionary <string, bool> usedParameters = UsedParameters(tomoSource);
List <string> entriesToDelete = new List <string>();
foreach (KeyValuePair <string, string> entry in ParameterDict.Current.Entries)
{
if (entry.Key.StartsWith("Formula.Parameters."))
{
if (!ParameterDict.IsAdditionalInfo(entry.Key))
{
string parameterName = entry.Key.Substring("Formula.Parameters.".Length);
if (!usedParameters.ContainsKey(parameterName))
{
entriesToDelete.Add(parameterName);
}
}
}
}
foreach (string entryToDelete in entriesToDelete)
{
ParameterDict.Current.RemoveProperty("Formula.Parameters." + entryToDelete);
}
} catch (System.InvalidOperationException)
{
// throwed if enumeration is not possible because ParameterDict.Current.Entries has changed in other thread.
}
}
}
public AlexNet(int classes = 1000, string prefix = "", ParameterDict @params = null) : base(prefix, @params)
{
Features = new HybridSequential(prefix);
Features.Add(new Conv2D(64, (11, 11), (4, 4), (2, 2), activation: ActivationType.Relu));
Features.Add(new MaxPool2D((3, 3), (2, 2)));
Features.Add(new Conv2D(192, (5, 5), padding: (2, 2), activation: ActivationType.Relu));
Features.Add(new MaxPool2D((3, 3), (2, 2)));
Features.Add(new Conv2D(384, (3, 3), padding: (1, 1), activation: ActivationType.Relu));
Features.Add(new Conv2D(256, (3, 3), padding: (1, 1), activation: ActivationType.Relu));
Features.Add(new Conv2D(256, (3, 3), padding: (1, 1), activation: ActivationType.Relu));
Features.Add(new MaxPool2D((3, 3), (2, 2)));
Features.Add(new Flatten());
Features.Add(new Dense(4096, ActivationType.Relu));
Features.Add(new Dropout(0.5f));
Features.Add(new Dense(4096, ActivationType.Relu));
Features.Add(new Dropout(0.5f));
Output = new Dense(classes);
RegisterChild(Features);
RegisterChild(Output);
}
public Dropout(float rate, Shape axes = null, string prefix = null, ParameterDict @params = null) : base(prefix,
@params)
{
Rate = rate;
Axes = axes;
}
public AvgPoolPad(string prefix = null, ParameterDict @params = null) : base(prefix, @params)
{
}
public ActionRecInceptionV3(int nclass, bool pretrained_base = true, bool partial_bn = true, float dropout_ratio = 0.8f, float init_std = 0.001f,
int feat_dim = 2048, int num_segments = 1, int num_crop = 1, string prefix = null, ParameterDict @params = null) : base(prefix, @params)
{
throw new NotImplementedException();
}
private static ISessionApiRequest<GetMessagesReply> GenericGetMessages(
string methodName, Method httpMethod, ParameterDict parameters)
{
// {"response":[4680,{"body":"thank you...<br>yeah...it'll pass....","title":"Re: ...",
// "date":1268213453,"uid":28672529,"mid":10836,"read_state":0},{"body":"nothing for)",
// "title":"Re: ...","date":1268213443,"uid":58416643,"mid":10835,"read_state":0}]}
return new SessionApiRequest<GetMessagesReply>(
methodName,
httpMethod,
parameters,
json => new GetMessagesReply
{
TotalCount = json.First().ToObject<int>(),
Messages = from message in json.Skip(1) select message.ToObject<VkMessage>()
}
);
}
/// <summary>
/// Generates the compressed formula.
/// The formula parameters and the formula itself are combined in a compact text,
/// which can later be copied into the formula text window to get the current
/// formula configuration.
/// </summary>
public static string GenerateCompressedFormulaAndViewSettings()
{
string formula = ParameterDict.Current["Intern.Formula.Source"];
List <string> formulaSettingCategories = new List <string>();
formulaSettingCategories.Add("Scene");
formulaSettingCategories.Add("Transformation.Camera");
formulaSettingCategories.Add("Transformation.Perspectice");
formulaSettingCategories.Add("Formula");
formulaSettingCategories.Add("Renderer");
formulaSettingCategories.Add("Renderer.BackColor");
formulaSettingCategories.Add("Renderer.ColorFactor");
formulaSettingCategories.Add("Renderer.Light");
// To make the new settings unique
ParameterDict.Current["intern.Formula.TempUpdateVal"] = "vv";
string testHash = ParameterDict.Current.GetHash("");
string insertSettingsStringHere = "base.Init();";
if (!formula.Contains(insertSettingsStringHere))
{
formula = @"
public override void Init()
{
base.Init();
}
" + formula;
}
StringBuilder settingsString = new StringBuilder();
settingsString.Append("if(GetString(\"intern.Formula.TempUpdateVal\")!=\"" + testHash + "\"){");
settingsString.Append("SetParameterBulk(\"");
foreach (KeyValuePair <string, string> entry in ParameterDict.Current.SortedEntries)
{
bool isInCategorie = false;
foreach (string testCategorie in formulaSettingCategories)
{
if (entry.Key.StartsWith(testCategorie))
{
isInCategorie = true;
break;
}
}
if (isInCategorie)
{
if (!ParameterDict.IsAdditionalInfo(entry.Key))
{
settingsString.Append("<Entry Key='" + entry.Key + "' Value='" + entry.Value + "' />");
}
}
}
// fix this formula to testHash
settingsString.Append("<Entry Key='intern.Formula.TempUpdateVal' Value='" + testHash + "' />");
settingsString.Append("\");");
settingsString.Append("}");
formula = formula.Replace(insertSettingsStringHere, insertSettingsStringHere + settingsString.ToString());
StringBuilder retVal = new StringBuilder();
retVal.Append(CompressFormula(formula));
return(retVal.ToString());
}
public FastSEResNet(string architecture, string norm_layer = "BatchNorm", FuncArgs kwargs = null, string prefix = "", ParameterDict @params = null) : base(prefix, @params)
{
throw new NotImplementedException();
}
public DarknetV3(int[] layers, int[] channels, int classes = 1000, string norm_layer = "BatchNorm", FuncArgs norm_kwargs = null, string prefix = null, ParameterDict @params = null) : base(prefix, @params)
{
Debug.Assert(layers.Length == channels.Length - 1, $"len(channels) should equal to len(layers) + 1, given {channels.Length} vs {layers.Length}");
this.features = new HybridSequential();
// first 3x3 conv
this.features.Add(Conv2d(channels[0], 3, 1, 1, norm_layer: norm_layer, norm_kwargs: norm_kwargs));
for (int i = 0; i < layers.Length; i++)
{
int nlayer = layers[i];
int channel = channels[i];
Debug.Assert(channel % 2 == 0, $"channel {channel} cannot be divided by 2");
// add downsample conv with stride=2
this.features.Add(Conv2d(channel, 3, 1, 2, norm_layer: norm_layer, norm_kwargs: norm_kwargs));
// add nlayer basic blocks
foreach (var _ in Enumerable.Range(0, nlayer))
{
this.features.Add(new DarknetBasicBlockV3(channel / 2, norm_layer: norm_layer, norm_kwargs: norm_kwargs));
}
}
// output
this.output = new Dense(classes);
RegisterChild(features);
RegisterChild(output);
}
public L2Loss(float?weight = 1, int?batch_axis = 0, string prefix = null, ParameterDict @params = null) :
base(weight, batch_axis, prefix, @params)
{
}
public MaskRCNN(HybridBlock features, HybridBlock top_features, string[] classes, int mask_channels = 256, int rcnn_max_dets = 1000, int rpn_test_pre_nms = 6000, int rpn_test_post_nms = 1000, int target_roi_scale = 1, int num_fcn_convs = 0, string norm_layer = "", FuncArgs norm_kwargs = null, string prefix = null, ParameterDict @params = null) : base(features: features, top_features: top_features, classes: classes, rpn_test_pre_nms: rpn_test_pre_nms, rpn_test_post_nms: rpn_test_post_nms, additional_output: true, prefix: prefix, @params: @params)
{
throw new NotImplementedException();
}
public SparseEmbedding(int input_dim, int output_dim, DType dtype = null,
string weight_initializer = null, string prefix = "", ParameterDict @params = null) : base(prefix,
@params)
{
throw new NotImplementedException();
}
public IDAUp(int out_channels, int in_channels, float[] up_f, bool use_dcnv2 = false, string norm_layer = "BatchNorm", FuncArgs norm_kwargs = null, string prefix = "", ParameterDict @params = null) : base(prefix, @params)
{
throw new NotImplementedException();
}
public GoogLeNet(int classes = 1000, string norm_layer = "BatchNorm", float dropout_ratio = 0.4f, bool aux_logits = false, FuncArgs norm_kwargs = null, bool partial_bn = false, bool pretrained_base = true, Context ctx = null, string prefix = null, ParameterDict @params = null) : base(prefix, @params)
{
throw new NotImplementedException();
}
public SoftmaxCELoss(int axis = -1, bool sparse_label = true, bool from_logits = false,
float?weight = null, int?batch_axis = 0, string prefix = null, ParameterDict @params = null) : base(
axis, sparse_label, from_logits, weight, batch_axis, prefix, @params)
{
}
public SoftmaxCrossEntropyLoss(int axis = -1, bool sparse_label = true, bool from_logits = false,
float?weight = null, int?batch_axis = 0, string prefix = null, ParameterDict @params = null) : base(
weight, batch_axis, prefix, @params)
{
_axis = axis;
_sparse_label = sparse_label;
_from_logits = from_logits;
}
public Tree(int levels, HybridBlock block, int in_channels, int out_channels, int stride = 1, bool level_root = false, int root_dim = 0, int root_kernel_size = 1,
int dilation = 1, bool root_residual = false, string norm_layer = "BatchNorm", FuncArgs norm_kwargs = null, string prefix = null, ParameterDict @params = null) : base(prefix, @params)
{
throw new NotImplementedException();
}
