public BinLeNetTest()
{
conv1 = new Conv2d(1, 20, 5, 1, 0);
pool1 = new MaxPool2d(2, 2);
conv2 = new BinConv2d(20, 50, 5, 1, 0);
pool2 = new MaxPool2d(2, 2);
fc1 = new BinLinear(50 * 4 * 4, 500, false);
fc2 = new Linear(500, 10, true);
bn1 = new BatchNorm2d(20);
bn2 = new BatchNorm2d(50);
bn3 = new BatchNorm1d(500);
conv1.Name = "conv1";
pool1.Name = "pool1";
pool2.Name = "pool2";
conv2.Name = "conv2";
fc1.Name = "fc1";
fc2.Name = "fc2";
bn1.Name = "bn1";
bn2.Name = "bn2";
bn3.Name = "bn3";
items.Add(bn1);
items.Add(bn2);
items.Add(bn3);
items.Add(fc2);
items.Add(fc1);
items.Add(conv1);
items.Add(conv2);
items.Add(pool1);
items.Add(pool2);
}
public BinConv2d(int inChannels, int outChannels, int kSize, int stride, int padding, bool bias = false, int dilation = 1, bool fpBn = false)
{
this.inChannels = inChannels;
this.outChannels = outChannels;
Weight = new InternalArray(new int[] { outChannels, inChannels, kSize, kSize });
this.padding = new int[] { padding, padding };
this.stride = new[] { stride, stride };
this.kernelSize = new[] { kSize, kSize };
this.dilation = new[] { dilation, dilation };
if (fpBn)
{
bn = new FPBatchNorm2d(inChannels);
}
else
{
bn = new BatchNorm2d(inChannels);
}
}