public void EvalManagedScalarTimesDualOutputTest()
{
string modelDefinition = @"deviceId = -1
precision = ""float""
traceLevel = 1
run=NDLNetworkBuilder
NDLNetworkBuilder=[
i1 = Input(1)
i2 = Input(1)
o1 = Times(Constant(3), i1, tag=""output"")
o2 = Times(Constant(5), i1, tag=""output"")
FeatureNodes = (i1)
]";
using (var model = new ModelEvaluationExtendedF())
{
model.CreateNetwork(modelDefinition);
VariableSchema outputSchema = model.GetOutputSchema();
VariableSchema inputSchema = model.GetInputSchema();
model.StartForwardEvaluation(outputSchema.Select(s => s.Name).ToList <string>());
var outputBuffer = outputSchema.CreateBuffers <float>();
var inputBuffer = inputSchema.CreateBuffers <float>();
inputBuffer[0].Buffer[0] = 2;
// We can call the evaluate method and get back the results...
model.ForwardPass(inputBuffer, outputBuffer);
float[][] expected = { new float[] { 6 }, new float[] { 10 } };
Assert.AreEqual(expected.Length, outputBuffer.Length);
for (int idx = 0; idx < expected.Length; idx++)
{
CollectionAssert.AreEqual(expected[idx], outputBuffer[idx].Buffer);
}
}
}
public void EvalManagedRNNTest()
{
string modelDefinition = @"deviceId = -1
precision = ""float""
traceLevel = 1
run=NDLNetworkBuilder
NDLNetworkBuilder = [
LSTMComponent(inputDim, outputDim, cellDim, inputx, cellDimX2, cellDimX3, cellDimX4) = [
wx = Parameter(cellDimX4, 0, init = ""uniform"", initValueScale = 1);
b = Parameter(cellDimX4, 1, init = ""fixedValue"", value = 0.0);
Wh = Parameter(cellDimX4, 0, init = ""uniform"", initValueScale = 1);
Wci = Parameter(cellDim, init = ""uniform"", initValueScale = 1);
Wcf = Parameter(cellDim, init = ""uniform"", initValueScale = 1);
Wco = Parameter(cellDim, init = ""uniform"", initValueScale = 1);
dh = PastValue(outputDim, output, timeStep = 1);
dc = PastValue(cellDim, ct, timeStep = 1);
wxx = Times(wx, inputx);
wxxpb = Plus(wxx, b);
whh = Times(wh, dh);
wxxpbpwhh = Plus(wxxpb, whh)
G1 = RowSlice(0, cellDim, wxxpbpwhh)
G2 = RowSlice(cellDim, cellDim, wxxpbpwhh)
G3 = RowSlice(cellDimX2, cellDim, wxxpbpwhh);
G4 = RowSlice(cellDimX3, cellDim, wxxpbpwhh);
Wcidc = DiagTimes(Wci, dc);
it = Sigmoid(Plus(G1, Wcidc));
bit = ElementTimes(it, Tanh(G2));
Wcfdc = DiagTimes(Wcf, dc);
ft = Sigmoid(Plus(G3, Wcfdc));
bft = ElementTimes(ft, dc);
ct = Plus(bft, bit);
Wcoct = DiagTimes(Wco, ct);
ot = Sigmoid(Plus(G4, Wcoct));
mt = ElementTimes(ot, Tanh(ct));
Wmr = Parameter(outputDim, cellDim, init = ""uniform"", initValueScale = 1);
output = Times(Wmr, mt);
]
i1 = Input(4)
o1 = LSTMComponent(4, 4, 1, i1, 2, 3, 4)
FeatureNodes = (i1)
outputNodes = (o1)
]";
using (var model = new ModelEvaluationExtendedF())
{
int featDim = 4;
int labelDim = 4;
model.CreateNetwork(modelDefinition);
VariableSchema inputSchema = model.GetInputSchema();
VariableSchema outputSchema = model.GetOutputSchema();
model.StartForwardEvaluation(outputSchema.Select(s => s.Name).ToList <string>());
// Allocate the output values layer
var outputBuffer = outputSchema.CreateBuffers <float>();
var inputBuffer = inputSchema.CreateBuffers <float>();
for (var i = 0; i < featDim; i++)
{
inputBuffer[0].Buffer[i] = (float)i;
}
int scaler = 100000;
var result = new int[labelDim];
int[] expected = { 50, 10, 54, 55 };
// the first pass with reset
model.ForwardPass(inputBuffer, outputBuffer);
for (var i = 0; i < labelDim; i++)
{
result[i] = (int)(outputBuffer[0].Buffer[i] * scaler);
}
CollectionAssert.AreEqual(expected, result);
// the second pass with reset
model.ForwardPass(inputBuffer, outputBuffer);
for (var i = 0; i < labelDim; i++)
{
result[i] = (int)(outputBuffer[0].Buffer[i] * scaler);
}
CollectionAssert.AreEqual(expected, result);
// another pass with reset
model.ForwardPass(inputBuffer, outputBuffer, true);
for (var i = 0; i < labelDim; i++)
{
result[i] = (int)(outputBuffer[0].Buffer[i] * scaler);
}
CollectionAssert.AreEqual(expected, result);
// pass w/o reset
model.ForwardPass(inputBuffer, outputBuffer, false);
for (var i = 0; i < labelDim; i++)
{
result[i] = (int)(outputBuffer[0].Buffer[i] * scaler);
}
expected = new int[] { 13, 2, 14, 14 };
CollectionAssert.AreEqual(expected, result);
// another pass w/o reset
model.ForwardPass(inputBuffer, outputBuffer, false);
for (var i = 0; i < labelDim; i++)
{
result[i] = (int)(outputBuffer[0].Buffer[i] * scaler);
}
expected = new int[] { -4, 0, -4, -4 };
CollectionAssert.AreEqual(expected, result);
}
}
public void EvalManagedScalarTimesDualOutputTest()
{
string modelDefinition = @"deviceId = -1
precision = ""float""
traceLevel = 1
run=NDLNetworkBuilder
NDLNetworkBuilder=[
i1 = Input(1)
i2 = Input(1)
o1 = Times(Constant(3), i1, tag=""output"")
o2 = Times(Constant(5), i1, tag=""output"")
FeatureNodes = (i1)
]";
using (var model = new ModelEvaluationExtendedF())
{
model.CreateNetwork(modelDefinition);
VariableSchema outputSchema = model.GetOutputSchema();
VariableSchema inputSchema = model.GetInputSchema();
model.StartForwardEvaluation(outputSchema.Select(s => s.Name).ToList<string>());
var outputBuffer = outputSchema.CreateBuffers<float>();
var inputBuffer = inputSchema.CreateBuffers<float>();
inputBuffer[0].Buffer[0] = 2;
// We can call the evaluate method and get back the results...
model.ForwardPass(inputBuffer, outputBuffer);
float[][] expected = {new float[]{6}, new float[]{10} };
Assert.AreEqual(expected.Length, outputBuffer.Length);
for(int idx=0; idx<expected.Length; idx++ )
{
CollectionAssert.AreEqual(expected[idx], outputBuffer[idx].Buffer);
}
}
}