private static double CalculateBleu(ThotSmtParameters parameters,
IReadOnlyList <IReadOnlyList <string> > sourceCorpus, IReadOnlyList <IReadOnlyList <string> > tuneTargetCorpus)
{
IEnumerable <IReadOnlyList <string> > translations = GenerateTranslations(parameters, sourceCorpus);
double bleu = Evaluation.CalculateBleu(translations, tuneTargetCorpus);
double penalty = 0;
for (int i = 0; i < parameters.ModelWeights.Count; i++)
{
if (i == 0 || i == 2 || i == 7)
{
continue;
}
if (parameters.ModelWeights[i] < 0)
{
penalty += parameters.ModelWeights[i] * 1000 * -1;
}
}
return((1.0 - bleu) + penalty);
}
public ThotSmtParameters Tune(ThotSmtParameters parameters,
IReadOnlyList <IReadOnlyList <string> > tuneSourceCorpus,
IReadOnlyList <IReadOnlyList <string> > tuneTargetCorpus, SmtBatchTrainStats stats,
IProgress <ProgressStatus> progress)
{
IntPtr weightUpdaterHandle = Thot.llWeightUpdater_create();
try
{
var iterQualities = new List <double>();
double bestQuality = double.MinValue;
ThotSmtParameters bestParameters = null;
int iter = 0;
HashSet <TranslationInfo>[] curNBestLists = null;
float[] curWeights = parameters.ModelWeights.ToArray();
while (true)
{
progress.Report(new ProgressStatus(iter, MaxIterations));
ThotSmtParameters newParameters = parameters.Clone();
newParameters.ModelWeights = curWeights;
newParameters.Freeze();
IList <TranslationInfo>[] nbestLists = GetNBestLists(newParameters, tuneSourceCorpus).ToArray();
double quality = Evaluation.CalculateBleu(nbestLists.Select(nbl => nbl.First().Translation),
tuneTargetCorpus);
iterQualities.Add(quality);
if (quality > bestQuality)
{
bestQuality = quality;
bestParameters = newParameters;
}
iter++;
if (iter >= MaxIterations || IsTuningConverged(iterQualities))
{
break;
}
if (curNBestLists == null)
{
curNBestLists = nbestLists.Select(nbl => new HashSet <TranslationInfo>(nbl)).ToArray();
}
else
{
for (int i = 0; i < nbestLists.Length; i++)
{
curNBestLists[i].UnionWith(nbestLists[i]);
}
}
UpdateWeights(weightUpdaterHandle, tuneTargetCorpus, curNBestLists, curWeights);
}
if (iter < MaxIterations)
{
progress.Report(new ProgressStatus(1.0));
}
stats.TranslationModelBleu = bestQuality;
return(bestParameters);
}
finally
{
Thot.llWeightUpdater_close(weightUpdaterHandle);
}
}