public void OverflowTest()
{
foreach (bool gradmode in new bool[] { false, true })
{
foreach (int batch in new int[] { 1, 2, 3 })
{
foreach (int inchannels in new int[] { 3, 6, 9, 12 })
{
foreach (int outchannels in new int[] { 3, 6, 9, 12 })
{
float[] yval = (new float[outchannels * batch]).Select((_, idx) => idx * 1e-3f).ToArray();
float[] wval = (new float[inchannels * outchannels / 9 * 4]).Select((_, idx) => idx * 1e-3f).Reverse().ToArray();
OverflowCheckedTensor y_tensor = new OverflowCheckedTensor(Shape.Map0D(outchannels, batch), yval);
OverflowCheckedTensor w_tensor = new OverflowCheckedTensor(Shape.Kernel0D(inchannels / 3 * 4, outchannels / 3), wval);
OverflowCheckedTensor x_tensor = new OverflowCheckedTensor(Shape.Map0D(inchannels, batch));
TrivectorTransposeDense ope = new TrivectorTransposeDense(outchannels, inchannels, gradmode, batch);
ope.Execute(y_tensor, w_tensor, x_tensor);
CollectionAssert.AreEqual(yval, y_tensor.State);
CollectionAssert.AreEqual(wval, w_tensor.State);
x_tensor.CheckOverflow();
Console.WriteLine($"pass: {inchannels},{outchannels},{batch},{gradmode}");
}
}
}
}
}
public void SpeedTest()
{
int inchannels = 33, outchannels = 33;
OverflowCheckedTensor y_tensor = new OverflowCheckedTensor(Shape.Map0D(outchannels));
OverflowCheckedTensor w_tensor = new OverflowCheckedTensor(Shape.Kernel0D(inchannels / 3 * 4, outchannels / 3));
OverflowCheckedTensor x_tensor = new OverflowCheckedTensor(Shape.Map0D(inchannels));
TrivectorTransposeDense ope = new TrivectorTransposeDense(outchannels, inchannels);
ope.Execute(y_tensor, w_tensor, x_tensor);
Stopwatch sw = new Stopwatch();
sw.Start();
ope.Execute(y_tensor, w_tensor, x_tensor);
ope.Execute(y_tensor, w_tensor, x_tensor);
ope.Execute(y_tensor, w_tensor, x_tensor);
ope.Execute(y_tensor, w_tensor, x_tensor);
sw.Stop();
Console.WriteLine($"{sw.ElapsedMilliseconds / 4} msec");
}
public void ExecuteTest()
{
float max_err = 0;
foreach (int batch in new int[] { 1, 2, 3 })
{
foreach (int inchannels in new int[] { 3, 6, 9, 12 })
{
foreach (int outchannels in new int[] { 3, 6, 9, 12 })
{
float[] yval = (new float[outchannels * batch]).Select((_, idx) => idx * 1e-3f).ToArray();
float[] wval = (new float[inchannels * outchannels / 9 * 4]).Select((_, idx) => idx * 1e-3f).Reverse().ToArray();
Trivector[] ycval = (new Trivector[yval.Length / 3])
.Select((_, idx) => new Trivector(yval[idx * 3], yval[idx * 3 + 1], yval[idx * 3 + 2])).ToArray();
Quaternion.Quaternion[] wcval = (new Quaternion.Quaternion[wval.Length / 4])
.Select((_, idx) => new Quaternion.Quaternion(wval[idx * 4], wval[idx * 4 + 1], wval[idx * 4 + 2], wval[idx * 4 + 3])).ToArray();
TrivectorMap0D y = new TrivectorMap0D(outchannels / 3, batch, ycval);
Quaternion.QuaternionFilter0D w = new Quaternion.QuaternionFilter0D(inchannels / 3, outchannels / 3, wcval);
TrivectorMap0D x = Reference(y, w);
OverflowCheckedTensor y_tensor = new OverflowCheckedTensor(Shape.Map0D(outchannels, batch), yval);
OverflowCheckedTensor w_tensor = new OverflowCheckedTensor(Shape.Kernel0D(inchannels / 3 * 4, outchannels / 3), wval);
OverflowCheckedTensor x_tensor = new OverflowCheckedTensor(Shape.Map0D(inchannels, batch));
TrivectorTransposeDense ope = new TrivectorTransposeDense(outchannels, inchannels, gradmode: false, batch);
ope.Execute(y_tensor, w_tensor, x_tensor);
float[] x_expect = x.ToArray();
float[] x_actual = x_tensor.State;
CollectionAssert.AreEqual(yval, y_tensor.State);
CollectionAssert.AreEqual(wval, w_tensor.State);
AssertError.Tolerance(x_expect, x_actual, 1e-7f, 1e-5f, ref max_err, $"mismatch value {inchannels},{outchannels},{batch}");
Console.WriteLine($"pass: {inchannels},{outchannels},{batch}");
}
}
}
Console.WriteLine($"maxerr:{max_err}");
}
