public Tensor Enhance(Tensor input)
{
var Temp = input;
Temp = Layers.Conv2D9x9(Temp, this.Data.Conv1_Weights, this.Data.Conv1_Biases);
if (this.Step != null)
{
Step(5);
}
Temp = Layers.ReLU(Temp);
if (this.Step != null)
{
Step(10);
}
Temp = Layers.BatchNorm(
Temp,
this.Data.Conv1_Shift,
this.Data.Conv1_Scale,
this.Data.Conv1_Mean,
this.Data.Conv1_Var);
if (this.Step != null)
{
Step(15);
}
Temp = Layers.ResidualBlock(
Temp,
this.Data.Res1_Conv1_Weights,
this.Data.Res1_Conv2_Weights,
this.Data.Res1_Conv1_Biases,
this.Data.Res1_Conv2_Biases,
this.Data.Res1_Conv1_Shift,
this.Data.Res1_Conv2_Shift,
this.Data.Res1_Conv1_Scale,
this.Data.Res1_Conv2_Scale,
this.Data.Res1_Conv1_Mean,
this.Data.Res1_Conv2_Mean,
this.Data.Res1_Conv1_Var,
this.Data.Res1_Conv2_Var);
if (this.Step != null)
{
Step(20);
}
Temp = Layers.ResidualBlock(
Temp,
this.Data.Res2_Conv1_Weights,
this.Data.Res2_Conv2_Weights,
this.Data.Res2_Conv1_Biases,
this.Data.Res2_Conv2_Biases,
this.Data.Res2_Conv1_Shift,
this.Data.Res2_Conv2_Shift,
this.Data.Res2_Conv1_Scale,
this.Data.Res2_Conv2_Scale,
this.Data.Res2_Conv1_Mean,
this.Data.Res2_Conv2_Mean,
this.Data.Res2_Conv1_Var,
this.Data.Res2_Conv2_Var);
if (this.Step != null)
{
Step(25);
}
Temp = Layers.ResidualBlock(
Temp,
this.Data.Res3_Conv1_Weights,
this.Data.Res3_Conv2_Weights,
this.Data.Res3_Conv1_Biases,
this.Data.Res3_Conv2_Biases,
this.Data.Res3_Conv1_Shift,
this.Data.Res3_Conv2_Shift,
this.Data.Res3_Conv1_Scale,
this.Data.Res3_Conv2_Scale,
this.Data.Res3_Conv1_Mean,
this.Data.Res3_Conv2_Mean,
this.Data.Res3_Conv1_Var,
this.Data.Res3_Conv2_Var);
if (this.Step != null)
{
Step(30);
}
Temp = Layers.ResidualBlock(
Temp,
this.Data.Res4_Conv1_Weights,
this.Data.Res4_Conv2_Weights,
this.Data.Res4_Conv1_Biases,
this.Data.Res4_Conv2_Biases,
this.Data.Res4_Conv1_Shift,
this.Data.Res4_Conv2_Shift,
this.Data.Res4_Conv1_Scale,
this.Data.Res4_Conv2_Scale,
this.Data.Res4_Conv1_Mean,
this.Data.Res4_Conv2_Mean,
this.Data.Res4_Conv1_Var,
this.Data.Res4_Conv2_Var);
if (this.Step != null)
{
Step(35);
}
Temp = Layers.ResidualBlock(
Temp,
this.Data.Res5_Conv1_Weights,
this.Data.Res5_Conv2_Weights,
this.Data.Res5_Conv1_Biases,
this.Data.Res5_Conv2_Biases,
this.Data.Res5_Conv1_Shift,
this.Data.Res5_Conv2_Shift,
this.Data.Res5_Conv1_Scale,
this.Data.Res5_Conv2_Scale,
this.Data.Res5_Conv1_Mean,
this.Data.Res5_Conv2_Mean,
this.Data.Res5_Conv1_Var,
this.Data.Res5_Conv2_Var);
if (this.Step != null)
{
Step(40);
}
Temp = Layers.ResidualBlock(
Temp,
this.Data.Res6_Conv1_Weights,
this.Data.Res6_Conv2_Weights,
this.Data.Res6_Conv1_Biases,
this.Data.Res6_Conv2_Biases,
this.Data.Res6_Conv1_Shift,
this.Data.Res6_Conv2_Shift,
this.Data.Res6_Conv1_Scale,
this.Data.Res6_Conv2_Scale,
this.Data.Res6_Conv1_Mean,
this.Data.Res6_Conv2_Mean,
this.Data.Res6_Conv1_Var,
this.Data.Res6_Conv2_Var);
if (this.Step != null)
{
Step(45);
}
Temp = Layers.Conv2DTranspose3x3(Temp, this.Data.Deconv1_Weights, this.Data.Deconv1_Biases);
if (this.Step != null)
{
Step(50);
}
Temp = Layers.ReLU(Temp);
if (this.Step != null)
{
Step(55);
}
Temp = Layers.BatchNorm(Temp, this.Data.Deconv1_Shift, this.Data.Deconv1_Scale, this.Data.Deconv1_Mean, this.Data.Deconv1_Var);
if (this.Step != null)
{
Step(60);
}
Temp = Layers.Conv2DTranspose3x3(Temp, this.Data.Deconv2_Weights, this.Data.Deconv2_Biases);
if (this.Step != null)
{
Step(65);
}
Temp = Layers.ReLU(Temp);
if (this.Step != null)
{
Step(70);
}
Temp = Layers.BatchNorm(Temp, this.Data.Deconv2_Shift, this.Data.Deconv2_Scale, this.Data.Deconv2_Mean, this.Data.Deconv2_Var);
if (this.Step != null)
{
Step(75);
}
Temp = Layers.Conv2DTranspose9x9(Temp, this.Data.Deconv3_Weights, this.Data.Deconv3_Biases);
if (this.Step != null)
{
Step(80);
}
Temp = Layers.ReLU(Temp);
if (this.Step != null)
{
Step(85);
}
Temp = Layers.BatchNorm(Temp, this.Data.Deconv3_Shift, this.Data.Deconv3_Scale, this.Data.Deconv3_Mean, this.Data.Deconv3_Var);
if (this.Step != null)
{
Step(90);
}
Temp = Layers.Tanh(Temp);
if (this.Step != null)
{
Step(95);
}
return(Temp);
}
