public BottleneckV2(int channels, int stride, bool downsample = false, int in_channels = 0,
string prefix = "", 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), (1, 1), use_bias: false);
RegisterChild(bn1, "bn1");
RegisterChild(conv1, "conv1");
RegisterChild(bn2, "bn2");
RegisterChild(conv2, "conv2");
RegisterChild(bn3, "bn3");
RegisterChild(conv3, "conv3");
if (downsample)
{
ds = new Conv2D(channels, (1, 1), (stride, stride), use_bias: false, in_channels: in_channels);
RegisterChild(ds, "downsample");
}
else
{
ds = null;
}
}
public ResNetV1(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();
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)));
}
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, channels[i]));
}
Features.Add(new GlobalAvgPool2D());
Output = new Dense(classes, in_units: channels.Last());
RegisterChild(Features, "features");
RegisterChild(Output, "output");
}
private void Btn_runML_Click(object sender, RoutedEventArgs e)
{
MainWindow.mMainWindow.listBox.Items.Clear();
switch (MLCore.MLModelSelected)
{
case MLModel.ResNet:
ResNet.EvaluationSingleImage(GV.imgOriginal);
for (int i = 0; i < ResNet.resultList.Count; i++)
{
MainWindow.mMainWindow.listBox.Items.Add(string.Format("{0}: {1}", MLCore.MLSelectedLabels[i], ResNet.resultList[i]));
}
BindMngr.GMessage.value = string.Format("This must be a {0}!", ResNet.OutputString, ResNet.OutputProbablility);
break;
case MLModel.FastRCNN:
FastRCNN.EvaluateObjectDetectionModel();
break;
case MLModel.Yolo:
GV.imgProcessed = new Image <Bgr, byte>(mYolo.Detect(GV.imgOriginal.ToBitmap()));
MainWindow.mMainWindow.ibOriginal.Source = ImgConverter.ToBitmapSource(GV.imgProcessed);
break;
}
}
public BottleneckV1(int channels, int stride, bool downsample = false, int in_channels = 0,
string prefix = "", ParameterDict @params = null) : base(prefix, @params)
{
var channel_one_fourth = Convert.ToInt32(channels / 4);
body = new HybridSequential("");
body.Add(new Conv2D(channel_one_fourth, (1, 1), (stride, stride)));
body.Add(new BatchNorm());
body.Add(new Activation(ActivationType.Relu));
body.Add(ResNet.Conv3x3(channel_one_fourth, 1, channel_one_fourth));
body.Add(new BatchNorm());
body.Add(new Activation(ActivationType.Relu));
body.Add(new Conv2D(channels, (1, 1), (1, 1)));
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());
RegisterChild(ds, "downsample");
}
else
{
ds = null;
}
RegisterChild(body, "body");
}
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;
}
}
private void Btn_ML_modelLoad_Click(object sender, RoutedEventArgs e)
{
switch (MLCore.MLModelSelected)
{
case MLModel.ResNet:
ResNet.LoadModel(TB_ML_modelName.Text);
break;
case MLModel.FastRCNN:
break;
case MLModel.Yolo:
mYolo.LoadModel(GV.ML_Folders[(int)MLFolders.ML_YOLO_model]);
break;
}
GB_ML_operation.IsEnabled = true;
}
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 MainWindow()
{
//Static MainWindow
mMainWindow = this;
//Windows.main = this;
DataContext = BindMngr;
loadingScreen.Show();
InitializeComponent();
mTrainingWindow = new TrainingWindow();
// Setup background worker
PreviewRoutine.previewSetup();
ConnectRoutine.connectionSetup();
ImageResizing.ImageResizingSetup();
ResNet.CNTK_ResNetSetup();
ZxingDecoder.DecoderSetup();
mTrainingWindow.TrainModelRoutineSetup();
OCR.OCRSetup(OCRMode.NUMBERS);
// Create Directories
GV.ML_Folders[(int)MLFolders.MLRoot] = Environment.CurrentDirectory + MLFolders.MLRoot.GetDescription();
GV.ML_Folders[(int)MLFolders.ML_CNTK] = Environment.CurrentDirectory + MLFolders.ML_CNTK.GetDescription();
GV.ML_Folders[(int)MLFolders.ML_CNTK_model] = Environment.CurrentDirectory + MLFolders.ML_CNTK_model.GetDescription();
GV.ML_Folders[(int)MLFolders.ML_YOLO] = Environment.CurrentDirectory + MLFolders.ML_YOLO.GetDescription();
GV.ML_Folders[(int)MLFolders.ML_YOLO_backup] = Environment.CurrentDirectory + MLFolders.ML_YOLO_backup.GetDescription();
GV.ML_Folders[(int)MLFolders.ML_YOLO_model] = Environment.CurrentDirectory + MLFolders.ML_YOLO_model.GetDescription();
GV.ML_Folders[(int)MLFolders.ML_YOLO_data] = Environment.CurrentDirectory + MLFolders.ML_YOLO_data.GetDescription();
GV.ML_Folders[(int)MLFolders.ML_YOLO_data_img] = Environment.CurrentDirectory + MLFolders.ML_YOLO_data_img.GetDescription();
foreach (string str in GV.ML_Folders)
{
if (str != null && !Directory.Exists(str))
{
Directory.CreateDirectory(str);
}
}
}
public static void Run()
{
var net = ResNet.ResNet18_v2(true);
var image = Img.ImRead("goldfish.jpg");
image = Img.ResizeShort(image, 227);
image = image.AsType(DType.Float32) / 255;
var normalized = Img.ColorNormalize(image, new NDArray(new[] { 0.485f, 0.456f, 0.406f }),
new NDArray(new[] { 0.229f, 0.224f, 0.225f }));
normalized = normalized.Transpose(new Shape(2, 0, 1));
normalized = normalized.ExpandDims(axis: 0);
var pred = net.Call(normalized);
NDArray prob = nd.Softmax(pred).Topk(k: 5);
var label_index = prob.ArrayData.OfType <float>().ToList();
var imagenet_labels = TestUtils.GetImagenetLabels();
foreach (int i in label_index)
{
Console.WriteLine(imagenet_labels[i]);
}
}
