public void LSTM_Test_Params_Count()
{
//define values, and variables
Variable x = Variable.InputVariable(new int[] { 2 }, DataType.Float, "input");
Variable y = Variable.InputVariable(new int[] { 3 }, DataType.Float, "output");
//Number of LSTM parameters
var lstm1 = RNN.RecurrenceLSTM(x, 3, 3, DataType.Float, device, false, Activation.TanH, false, false, 1);
var ft = lstm1.Inputs.Where(l => l.Uid.StartsWith("Parameter")).ToList();
var consts = lstm1.Inputs.Where(l => l.Uid.StartsWith("Constant")).ToList();
var inp = lstm1.Inputs.Where(l => l.Uid.StartsWith("Input")).ToList();
//bias params
var bs = ft.Where(p => p.Name.Contains("_b")).ToList();
var totalBs = bs.Sum(v => v.Shape.TotalSize);
Assert.Equal(12, totalBs);
//weights
var ws = ft.Where(p => p.Name.Contains("_w")).ToList();
var totalWs = ws.Sum(v => v.Shape.TotalSize);
Assert.Equal(24, totalWs);
//update
var us = ft.Where(p => p.Name.Contains("_u")).ToList();
var totalUs = us.Sum(v => v.Shape.TotalSize);
Assert.Equal(36, totalUs);
//total number opf parameters
var totalOnly = totalBs + totalWs + totalUs;
Assert.Equal(72, totalOnly);
}
public void lstm_test01()
{
//define values, and variables
Variable x = Variable.InputVariable(new int[] { 4 }, DataType.Float, "input");
var xValues = Value.CreateBatchOfSequences <float>(new int[] { 4 }, mData, device);
//
var lstm00 = RNN.RecurrenceLSTM(x, 3, 3, DataType.Float, device, false, Activation.TanH, true, true, 1);
//
LSTMReccurentNN lstmNN = new LSTMReccurentNN(1, 1, device);
//lstm implme reference 01
var lstmCell = lstmNN.CreateLSTM(x, "output1");
var lstm01 = CNTKLib.SequenceLast(lstmCell.h);
//lstme implementation refe 02
var lstm02 = LSTMSequenceClassifier.LSTMNet(x, 1, device, "output1");
//
var wParams00 = lstm00.Inputs.Where(p => p.Uid.Contains("Parameter")).ToList();
var wParams01 = lstm00.Inputs.Where(p => p.Uid.Contains("Parameter")).ToList();
var wParams02 = lstm00.Inputs.Where(p => p.Uid.Contains("Parameter")).ToList();
//parameter count
Assert.Equal(wParams00.Count, wParams01.Count);
Assert.Equal(wParams00.Count, wParams02.Count);
//structure of parameters test
Assert.Equal(wParams00.Where(p => p.Name.Contains("_b")).Count(), wParams01.Where(p => p.Name.Contains("_b")).Count());
Assert.Equal(wParams00.Where(p => p.Name.Contains("_w")).Count(), wParams01.Where(p => p.Name.Contains("_w")).Count());
Assert.Equal(wParams00.Where(p => p.Name.Contains("_u")).Count(), wParams01.Where(p => p.Name.Contains("_u")).Count());
Assert.Equal(wParams00.Where(p => p.Name.Contains("peep")).Count(), wParams01.Where(p => p.Name.Contains("peep")).Count());
Assert.Equal(wParams00.Where(p => p.Name.Contains("stabilize")).Count(), wParams01.Where(p => p.Name.Contains("stabilize")).Count());
//check structure of parameters with originaly developed lstm
//chech for arguments
Assert.True(lstm01.Arguments.Count == lstm02.Arguments.Count);
for (int i = 0; i < lstm01.Arguments.Count; i++)
{
testVariable(lstm01.Arguments[i], lstm01.Arguments[i]);
}
///
Assert.True(lstm01.Inputs.Count == lstm02.Inputs.Count);
for (int i = 0; i < lstm01.Inputs.Count; i++)
{
testVariable(lstm01.Inputs[i], lstm02.Inputs[i]);
}
///
Assert.True(lstm01.Outputs.Count == lstm02.Outputs.Count);
for (int i = 0; i < lstm01.Outputs.Count; i++)
{
testVariable(lstm01.Outputs[i], lstm02.Outputs[i]);
}
}
/// <summary>
/// Implementation of custom NN model
/// </summary>
/// <param name="data"></param>
/// <param name="yearVar"></param>
/// <param name="montVar"></param>
/// <param name="shopVar"></param>
/// <param name="itemVar"></param>
/// <param name="cnt3Var"></param>
/// <param name="label"></param>
/// <param name="device"></param>
/// <returns></returns>
private static Function PredictFutureSalesModel(List <Variable> variables, DeviceDescriptor device)
{
//define features and label vars
Variable yearVar = variables[0];
Variable montVar = variables[1];
Variable shopVar = variables[2];
Variable itemVar = variables[3];
Variable cnt3Var = variables[4];
Variable label = variables[5];
//create rnn object
var ffNet = new FeedForwaredNN(device);
//predefined parameters
var H_DIMS = 11;
var CELL_DIMS = 3;
var DROPRATRE = 0.2f;
var outDim = label.Shape.Dimensions.Last();
//embedding layer and dimensionality reduction
var yearEmb = Embedding.Create(yearVar, yearVar.Shape.Dimensions[0] - 1, DataType.Float, device, 1, yearVar.Name + "_emb");
var monthEmb = Embedding.Create(montVar, montVar.Shape.Dimensions[0] / 2, DataType.Float, device, 1, montVar.Name + "_emb");
var varshopEmb = Embedding.Create(shopVar, shopVar.Shape.Dimensions[0] / 2, DataType.Float, device, 1, shopVar.Name + "_emb");
var itemEmb = Embedding.Create(itemVar, itemVar.Shape.Dimensions[0] / 2, DataType.Float, device, 1, itemVar.Name + "_emb");
var itemEmb2 = Embedding.Create(itemEmb, itemEmb.Output.Shape.Dimensions[0] / 4, DataType.Float, device, 1, itemEmb.Name + "_emb");
//join all embedding layers with input variable of previous product sales
var emb = CNTKLib.Splice(new VariableVector()
{
yearEmb, monthEmb, varshopEmb, itemEmb2, cnt3Var
}, new Axis(0));
//create recurrence for time series on top of joined layer
var lstmLayer = RNN.RecurrenceLSTM(emb, H_DIMS, CELL_DIMS, DataType.Float, device, false, Activation.TanH, true, true);
//create dense on top of LSTM recurrence layers
var denseLayer = ffNet.Dense(lstmLayer, 33, Activation.TanH);
//create dropout layer on top of dense layer
var dropoutLay = CNTKLib.Dropout(denseLayer, DROPRATRE);
//create dense layer without activation function
var outLayer = ffNet.Dense(dropoutLay, outDim, Activation.None, label.Name);
//
return(outLayer);
}
/// <summary>
/// Create cntk model function by providing parameters. The method is able for create:
/// - feedforward with one hidden layer and any number of neurons
/// - deep neural network with any number of hidden layers and any number of neurons. Each hidden number has the same number of neurons
/// - LSTM NN with any number of hidden layers of LSTM , and any number of LSTM Cells in each layer. Also at the top of the network you can define
/// one dense layer and one dropout layer.
/// </summary>
/// <param name="nnParams"></param>
/// <returns></returns>
public static Function CreateNetwrok(List <NNLayer> layers, List <Variable> inputVars, Variable outpuVar, DeviceDescriptor device)
{
DataType type = DataType.Float;
Variable inputLayer = null;
if (inputVars.Count > 1)
{
var vv = new VariableVector();
foreach (var v in inputVars)
{
//check if variable is stores as Sparse then we should create one embedding layer before slice
//since mixing sparse and dense data is not supported
if (v.IsSparse)
{
var v1 = Embedding.Create(v, v.Shape.Dimensions.Last(), type, device, 1, v.Name + "_sp_emb");
vv.Add(v1);
}
else
{
vv.Add(v);
}
}
//
inputLayer = (Variable)CNTKLib.Splice(vv, new Axis(0));
}
else //define input layer
{
inputLayer = inputVars.First();
}
//Create network
var net = inputLayer;
var ff = new FeedForwaredNN(device, type);
//set last layer name to label name
layers.Last().Name = outpuVar.Name;
//get last LSTM layer
var lastLSTM = layers.Where(x => x.Type == LayerType.LSTM).LastOrDefault();
//
foreach (var layer in layers)
{
if (layer.Type == LayerType.Dense)
{
net = ff.Dense(net, layer.HDimension, layer.Activation, layer.Name);
}
else if (layer.Type == LayerType.Drop)
{
net = CNTKLib.Dropout(net, layer.Value / 100.0f);
}
else if (layer.Type == LayerType.Embedding)
{
net = Embedding.Create(net, layer.HDimension, type, device, 1, layer.Name);
}
else if (layer.Type == LayerType.LSTM)
{
var returnSequence = true;
if (layers.IndexOf(lastLSTM) == layers.IndexOf(layer))
{
returnSequence = false;
}
net = RNN.RecurrenceLSTM(net, layer.HDimension, layer.CDimension, type, device, returnSequence, layer.Activation,
layer.Peephole, layer.SelfStabilization, 1);
}
}
//check if last layer is compatible with the output
if (net.Shape.Dimensions.Last() != outpuVar.Shape.Dimensions.Last())
{
ff.CreateOutputLayer(net, outpuVar, Activation.None);
}
return(net);
}
public void LSTM_Test_Params_Count_with_peep_selfstabilize()
{
//define values, and variables
Variable x = Variable.InputVariable(new int[] { 2 }, DataType.Float, "input");
Variable y = Variable.InputVariable(new int[] { 3 }, DataType.Float, "output");
#region lstm org implemented in cntk for reference
//lstme implementation refe 02
var lstmTest02 = LSTMSequenceClassifier.LSTMNet(x, 3, device, "output1");
var ft2 = lstmTest02.Inputs.Where(l => l.Uid.StartsWith("Parameter")).ToList();
var totalSize = ft2.Sum(p => p.Shape.TotalSize);
//bias params
var bs2 = ft2.Where(p => p.Name.Contains("_b")).ToList();
var totalBs2 = bs2.Sum(v => v.Shape.TotalSize);
//weights
var ws2 = ft2.Where(p => p.Name.Contains("_w")).ToList();
var totalWs2 = ws2.Sum(v => v.Shape.TotalSize);
//update
var us2 = ft2.Where(p => p.Name.Contains("_u")).ToList();
var totalUs2 = us2.Sum(v => v.Shape.TotalSize);
//peephole
var ph2 = ft2.Where(p => p.Name.Contains("_peep")).ToList();
var totalph2 = ph2.Sum(v => v.Shape.TotalSize);
//stabilize
var st2 = ft2.Where(p => p.Name.Contains("_stabilize")).ToList();
var totalst2 = st2.Sum(v => v.Shape.TotalSize);
#endregion
#region anndotnet old implementation
//
//LSTMReccurentNN lstmNN = new LSTMReccurentNN(3, 3, device);
////lstm implme reference 01
//var lstmCell11 = lstmNN.CreateLSTM(x, "output1");
//var lstmTest01 = CNTKLib.SequenceLast(lstmCell11.h);
//var ft1 = lstmTest01.Inputs.Where(l => l.Uid.StartsWith("Parameter")).ToList();
//var consts1 = lstmTest01.Inputs.Where(l => l.Uid.StartsWith("Constant")).ToList();
//var inp1 = lstmTest01.Inputs.Where(l => l.Uid.StartsWith("Input")).ToList();
//var pparams1 = ft1.Sum(v => v.Shape.TotalSize);
#endregion
//Number of LSTM parameters
var lstm1 = RNN.RecurrenceLSTM(x, 3, 3, DataType.Float, device, false, Activation.TanH, true, true, 1);
var ft = lstm1.Inputs.Where(l => l.Uid.StartsWith("Parameter")).ToList();
var consts = lstm1.Inputs.Where(l => l.Uid.StartsWith("Constant")).ToList();
var inp = lstm1.Inputs.Where(l => l.Uid.StartsWith("Input")).ToList();
//bias params
var bs = ft.Where(p => p.Name.Contains("_b")).ToList();
var totalBs = bs.Sum(v => v.Shape.TotalSize);
Assert.Equal(12, totalBs);
//weights
var ws = ft.Where(p => p.Name.Contains("_w")).ToList();
var totalWs = ws.Sum(v => v.Shape.TotalSize);
Assert.Equal(24, totalWs);
//update
var us = ft.Where(p => p.Name.Contains("_u")).ToList();
var totalUs = us.Sum(v => v.Shape.TotalSize);
Assert.Equal(36, totalUs);
//peephole
var ph = ft.Where(p => p.Name.Contains("_peep")).ToList();
var totalPh = ph.Sum(v => v.Shape.TotalSize);
Assert.Equal(9, totalPh);
//stabilize
var st = ft.Where(p => p.Name.Contains("_stabilize")).ToList();
var totalst = st.Sum(v => v.Shape.TotalSize);
Assert.Equal(6, totalst);
var totalOnly = totalBs + totalWs + totalUs;
var totalWithSTabilize = totalOnly + totalst;
var totalWithPeep = totalOnly + totalPh;
var totalP = totalOnly + totalst + totalPh;
var totalParams = ft.Sum(v => v.Shape.TotalSize);
Assert.Equal(totalP, totalParams);
//72- without peep and stab
//75 - witout peep with stabil +3xm =
//81 - with peephole and without stabil
//87 - with peep ans stab 3+9
}