public WordEMWrapFeaturizer(string filename)
{
Txt2Vec.Decoder decoder = new Txt2Vec.Decoder();
decoder.LoadBinaryModel(filename);
string[] terms = decoder.GetAllTerms();
vectorSize = decoder.GetVectorSize();
m_WordEmbedding = new Dictionary<string, SingleVector>();
m_UnkEmbedding = new SingleVector(vectorSize);
foreach (string term in terms)
{
double[] vector = decoder.GetVector(term);
if (vector != null)
{
SingleVector spVector = new SingleVector(vectorSize, vector);
spVector.Normalize();
m_WordEmbedding.Add(term, spVector);
}
}
}
public WordEMWrapFeaturizer(string filename, bool textFormat = false)
{
var model = new Model();
model.LoadModel(filename, textFormat);
var terms = model.GetAllTerms();
vectorSize = model.VectorSize;
if (vectorSize % Vector <float> .Count != 0)
{
vectorSize += (Vector <float> .Count - (vectorSize % Vector <float> .Count));
}
m_WordEmbedding = new Dictionary <string, SingleVector>();
m_UnkEmbedding = new SingleVector(vectorSize);
foreach (var term in terms)
{
var vector = model.GetVector(term);
if (vector != null)
{
var spVector = new SingleVector(vector, vectorSize);
m_WordEmbedding.Add(term, spVector);
}
}
}
public void Append(SingleVector vector)
{
if (m_nLenPerBlock > 0)
{
if (m_nLenPerBlock != vector.Length)
{
throw new Exception("The dimension of appending vector is not the same as the previous one");
}
}
m_innerData.Add(vector);
m_nLenPerBlock = vector.Length;
m_nLen += m_nLenPerBlock;
}
public void Append(SingleVector vector)
{
if (m_nLenPerBlock > 0)
{
if (m_nLenPerBlock != vector.GetDimension())
{
throw new Exception("The dimension of appending vector is not the same as the previous one");
}
}
m_innerData.Add(vector);
m_nLenPerBlock = vector.GetDimension();
m_nLen += m_nLenPerBlock;
}
public WordEMWrapFeaturizer(string filename)
{
Txt2Vec.Model model = new Txt2Vec.Model();
model.LoadBinaryModel(filename);
string[] terms = model.GetAllTerms();
vectorSize = model.VectorSize;
m_WordEmbedding = new Dictionary <string, SingleVector>();
m_UnkEmbedding = new SingleVector(vectorSize);
foreach (string term in terms)
{
float[] vector = model.GetVector(term);
if (vector != null)
{
SingleVector spVector = new SingleVector(vectorSize, vector);
m_WordEmbedding.Add(term, spVector);
}
}
}
public WordEMWrapFeaturizer(string filename, bool textFormat = false)
{
Txt2Vec.Model model = new Txt2Vec.Model();
model.LoadModel(filename, textFormat);
string[] terms = model.GetAllTerms();
vectorSize = model.VectorSize;
m_WordEmbedding = new Dictionary<string, SingleVector>();
m_UnkEmbedding = new SingleVector(vectorSize);
foreach (string term in terms)
{
float[] vector = model.GetVector(term);
if (vector != null)
{
SingleVector spVector = new SingleVector(vectorSize, vector);
m_WordEmbedding.Add(term, spVector);
}
}
}
private void CreateCells()
{
neuFeatures = new SingleVector(DenseFeatureSize);
OutputLayer = new neuron[L2];
neuHidden = new neuron[L1];
}
public SingleVector Set(SingleVector rhs, int startOffset)
{
for (int i = 0; i < rhs.GetDimension(); i++)
{
m_innerData[i + startOffset] = rhs.m_innerData[i];
}
return this;
}
private void CreateCell(BinaryReader br)
{
neuFeatures = new SingleVector(DenseFeatureSize);
OutputLayer = new neuron[L2];
for (int a = 0; a < L2; a++)
{
OutputLayer[a].cellOutput = 0;
OutputLayer[a].er = 0;
}
neuHidden = new LSTMCell[L1];
for (int i = 0; i < L1; i++)
{
neuHidden[i] = new LSTMCell();
LSTMCellInit(neuHidden[i]);
}
if (br != null)
{
//Load weight from input file
for (int i = 0; i < L1; i++)
{
neuHidden[i].wCellIn = br.ReadSingle();
neuHidden[i].wCellForget = br.ReadSingle();
neuHidden[i].wCellOut = br.ReadSingle();
}
}
else
{
//Initialize weight by random number
for (int i = 0; i < L1; i++)
{
//internal weights, also important
neuHidden[i].wCellIn = RandInitWeight();
neuHidden[i].wCellForget = RandInitWeight();
neuHidden[i].wCellOut = RandInitWeight();
}
}
}
