public AttentionSeq2Seq(string modelFilePath, int batchSize, ArchTypeEnums archType, int[] deviceIds)
{
CheckParameters(batchSize, archType, deviceIds);
if (archType == ArchTypeEnums.GPU_CUDA)
{
TensorAllocator.InitDevices(deviceIds);
SetDefaultDeviceIds(deviceIds.Length);
}
m_archType = archType;
m_deviceIds = deviceIds;
Load(modelFilePath);
InitWeightsFactory();
SetBatchSize(batchSize);
}
public void UpdateWeights(List <IWeightTensor> model, int batchSize, float step_size, float regc, int iter)
{
Dictionary <int, List <IWeightTensor> > id2Models = new Dictionary <int, List <IWeightTensor> >();
HashSet <string> setWeightsName = new HashSet <string>();
foreach (var item in model)
{
if (setWeightsName.Contains(item.Name))
{
throw new ArgumentException($"Found duplicated weights name '{item.Name}'");
}
setWeightsName.Add(item.Name);
if (id2Models.ContainsKey(item.DeviceId) == false)
{
id2Models.Add(item.DeviceId, new List <IWeightTensor>());
}
id2Models[item.DeviceId].Add(item);
if (m_cacheName2V.ContainsKey(item.Name) == false)
{
var allocator = TensorAllocator.Allocator(item.DeviceId);
m_cacheName2V[item.Name] = new Tensor(allocator, DType.Float32, item.Sizes);
Ops.Fill(m_cacheName2V[item.Name], 0.0f);
m_cacheName2M[item.Name] = new Tensor(allocator, DType.Float32, item.Sizes);
Ops.Fill(m_cacheName2M[item.Name], 0.0f);
Logger.WriteLine($"Added weight '{item.Name}' to optimizer.");
}
}
Parallel.ForEach(id2Models, kv =>
{
foreach (var item in kv.Value)
{
var m = item as WeightTensor;
UpdateWeightsTensor(m, batchSize, step_size, m_clipval, regc, iter);
}
});
}
private void TrainEp(int ep, float learningRate)
{
int processedLine = 0;
DateTime startDateTime = DateTime.Now;
double costInTotal = 0.0;
long srcWordCnts = 0;
long tgtWordCnts = 0;
double avgCostPerWordInTotal = 0.0;
double lastAvgCostPerWordInTotal = 100000.0;
List <SntPair> sntPairs = new List <SntPair>();
TensorAllocator.FreeMemoryAllDevices();
Logger.WriteLine($"Base learning rate is '{learningRate}' at epoch '{ep}'");
//Clean caches of parameter optmization
Logger.WriteLine($"Cleaning cache of weights optmiazation.'");
CleanWeightCache();
Logger.WriteLine($"Start to process training corpus.");
foreach (var sntPair in TrainCorpus)
{
sntPairs.Add(sntPair);
if (sntPairs.Count == TrainCorpus.BatchSize)
{
List <IWeightMatrix> encoded = new List <IWeightMatrix>();
List <List <string> > srcSnts = new List <List <string> >();
List <List <string> > tgtSnts = new List <List <string> >();
var slen = 0;
var tlen = 0;
for (int j = 0; j < TrainCorpus.BatchSize; j++)
{
List <string> srcSnt = new List <string>();
//Add BOS and EOS tags to source sentences
srcSnt.Add(m_START);
srcSnt.AddRange(sntPairs[j].SrcSnt);
srcSnt.Add(m_END);
srcSnts.Add(srcSnt);
tgtSnts.Add(sntPairs[j].TgtSnt.ToList());
slen += srcSnt.Count;
tlen += sntPairs[j].TgtSnt.Length;
}
srcWordCnts += slen;
tgtWordCnts += tlen;
Reset();
//Copy weights from weights kept in default device to all other devices
SyncWeights();
float cost = 0.0f;
Parallel.For(0, m_deviceIds.Length, i =>
{
IComputeGraph computeGraph = CreateComputGraph(i);
//Bi-directional encoding input source sentences
IWeightMatrix encodedWeightMatrix = Encode(computeGraph, srcSnts.GetRange(i * m_batchSize, m_batchSize), m_biEncoder[i], m_srcEmbedding[i]);
//Generate output decoder sentences
List <List <string> > predictSentence;
float lcost = Decode(tgtSnts.GetRange(i * m_batchSize, m_batchSize), computeGraph, encodedWeightMatrix, m_decoder[i], m_decoderFFLayer[i],
m_tgtEmbedding[i], out predictSentence);
lock (locker)
{
cost += lcost;
}
//Calculate gradients
computeGraph.Backward();
});
//Sum up gradients in all devices, and kept it in default device for parameters optmization
SyncGradient();
if (float.IsInfinity(cost) == false && float.IsNaN(cost) == false)
{
processedLine += TrainCorpus.BatchSize;
double costPerWord = (cost / tlen);
costInTotal += cost;
avgCostPerWordInTotal = costInTotal / tgtWordCnts;
lastAvgCostPerWordInTotal = avgCostPerWordInTotal;
}
else
{
Logger.WriteLine($"Invalid cost value.");
}
//Optmize parameters
float avgAllLR = UpdateParameters(learningRate, TrainCorpus.BatchSize);
//Clear gradient over all devices
ClearGradient();
if (IterationDone != null && processedLine % (100 * TrainCorpus.BatchSize) == 0)
{
IterationDone(this, new CostEventArg()
{
AvgLearningRate = avgAllLR,
CostPerWord = cost / tlen,
avgCostInTotal = avgCostPerWordInTotal,
Epoch = ep,
ProcessedSentencesInTotal = processedLine,
ProcessedWordsInTotal = srcWordCnts * 2 + tgtWordCnts,
StartDateTime = startDateTime
});
}
//Save model for each 10000 steps
if (processedLine % (TrainCorpus.BatchSize * 1000) == 0)
{
Save();
TensorAllocator.FreeMemoryAllDevices();
}
sntPairs.Clear();
}
}
Logger.WriteLine($"Epoch '{ep}' took '{DateTime.Now - startDateTime}' time to finish.");
Save();
}
