public CacheProcessor(CacheParseHypotheses cacher, LexicalizedParser parser, int dvKBest, ITreeTransformer transformer)
{
this.cacher = cacher;
this.parser = parser;
this.dvKBest = dvKBest;
this.transformer = transformer;
}
public virtual bool RunGradientCheck(IList <Tree> sentences, IdentityHashMap <Tree, byte[]> compressedParses)
{
log.Info("Gradient check: converting " + sentences.Count + " compressed trees");
IdentityHashMap <Tree, IList <Tree> > topParses = CacheParseHypotheses.ConvertToTrees(sentences, compressedParses, op.trainOptions.trainingThreads);
log.Info("Done converting trees");
DVParserCostAndGradient gcFunc = new DVParserCostAndGradient(sentences, topParses, dvModel, op);
return(gcFunc.GradientCheck(1000, 50, dvModel.ParamsToVector()));
}
public virtual void FilterRulesForBatch(IDictionary <Tree, byte[]> compressedTrees)
{
TwoDimensionalSet <string, string> binaryRules = TwoDimensionalSet.TreeSet();
ICollection <string> unaryRules = new HashSet <string>();
ICollection <string> words = new HashSet <string>();
foreach (KeyValuePair <Tree, byte[]> entry in compressedTrees)
{
SearchRulesForBatch(binaryRules, unaryRules, words, entry.Key);
foreach (Tree hypothesis in CacheParseHypotheses.ConvertToTrees(entry.Value))
{
SearchRulesForBatch(binaryRules, unaryRules, words, hypothesis);
}
}
FilterRulesForBatch(binaryRules, unaryRules, words);
}
public virtual Pair <Tree, byte[]> Process(Tree tree)
{
IList <Tree> topParses = DVParser.GetTopParsesForOneTree(parser, dvKBest, tree, transformer);
// this block is a test to make sure the conversion code is working...
IList <Tree> converted = CacheParseHypotheses.ConvertToTrees(cacher.ConvertToBytes(topParses));
IList <Tree> simplified = CollectionUtils.TransformAsList(topParses, cacher.treeBasicCategories);
simplified = CollectionUtils.FilterAsList(simplified, cacher.treeFilter);
if (simplified.Count != topParses.Count)
{
log.Info("Filtered " + (topParses.Count - simplified.Count) + " trees");
if (simplified.Count == 0)
{
log.Info(" WARNING: filtered all trees for " + tree);
}
}
if (!simplified.Equals(converted))
{
if (converted.Count != simplified.Count)
{
throw new AssertionError("horrible error: tree sizes not equal, " + converted.Count + " vs " + simplified.Count);
}
for (int i = 0; i < converted.Count; ++i)
{
if (!simplified[i].Equals(converted[i]))
{
System.Console.Out.WriteLine("=============================");
System.Console.Out.WriteLine(simplified[i]);
System.Console.Out.WriteLine("=============================");
System.Console.Out.WriteLine(converted[i]);
System.Console.Out.WriteLine("=============================");
throw new AssertionError("horrible error: tree " + i + " not equal for base tree " + tree);
}
}
}
return(Pair.MakePair(tree, cacher.ConvertToBytes(topParses)));
}
public virtual void SetRulesForTrainingSet(IList <Tree> sentences, IDictionary <Tree, byte[]> compressedTrees)
{
TwoDimensionalSet <string, string> binaryRules = TwoDimensionalSet.TreeSet();
ICollection <string> unaryRules = new HashSet <string>();
ICollection <string> words = new HashSet <string>();
foreach (Tree sentence in sentences)
{
SearchRulesForBatch(binaryRules, unaryRules, words, sentence);
foreach (Tree hypothesis in CacheParseHypotheses.ConvertToTrees(compressedTrees[sentence]))
{
SearchRulesForBatch(binaryRules, unaryRules, words, hypothesis);
}
}
foreach (Pair <string, string> binary in binaryRules)
{
AddRandomBinaryMatrix(binary.first, binary.second);
}
foreach (string unary in unaryRules)
{
AddRandomUnaryMatrix(unary);
}
FilterRulesForBatch(binaryRules, unaryRules, words);
}
/// <summary>
/// An example command line for training a new parser:
/// <br />
/// nohup java -mx6g edu.stanford.nlp.parser.dvparser.DVParser -cachedTrees /scr/nlp/data/dvparser/wsj/cached.wsj.train.simple.ser.gz -train -testTreebank /afs/ir/data/linguistic-data/Treebank/3/parsed/mrg/wsj/22 2200-2219 -debugOutputFrequency 400 -nofilter -trainingThreads 5 -parser /u/nlp/data/lexparser/wsjPCFG.nocompact.simple.ser.gz -trainingIterations 40 -batchSize 25 -model /scr/nlp/data/dvparser/wsj/wsj.combine.v2.ser.gz -unkWord "*UNK*" -dvCombineCategories > /scr/nlp/data/dvparser/wsj/wsj.combine.v2.out 2>&1 &
/// </summary>
/// <exception cref="System.IO.IOException"/>
/// <exception cref="System.TypeLoadException"/>
public static void Main(string[] args)
{
if (args.Length == 0)
{
Help();
System.Environment.Exit(2);
}
log.Info("Running DVParser with arguments:");
foreach (string arg in args)
{
log.Info(" " + arg);
}
log.Info();
string parserPath = null;
string trainTreebankPath = null;
IFileFilter trainTreebankFilter = null;
string cachedTrainTreesPath = null;
bool runGradientCheck = false;
bool runTraining = false;
string testTreebankPath = null;
IFileFilter testTreebankFilter = null;
string initialModelPath = null;
string modelPath = null;
bool filter = true;
string resultsRecordPath = null;
IList <string> unusedArgs = new List <string>();
// These parameters can be null or 0 if the model was not
// serialized with the new parameters. Setting the options at the
// command line will override these defaults.
// TODO: if/when we integrate back into the main branch and
// rebuild models, we can get rid of this
IList <string> argsWithDefaults = new List <string>(Arrays.AsList(new string[] { "-wordVectorFile", Options.LexOptions.DefaultWordVectorFile, "-dvKBest", int.ToString(TrainOptions.DefaultKBest), "-batchSize", int.ToString(TrainOptions.DefaultBatchSize
), "-trainingIterations", int.ToString(TrainOptions.DefaultTrainingIterations), "-qnIterationsPerBatch", int.ToString(TrainOptions.DefaultQnIterationsPerBatch), "-regCost", double.ToString(TrainOptions.DefaultRegcost), "-learningRate", double
.ToString(TrainOptions.DefaultLearningRate), "-deltaMargin", double.ToString(TrainOptions.DefaultDeltaMargin), "-unknownNumberVector", "-unknownDashedWordVectors", "-unknownCapsVector", "-unknownchinesepercentvector", "-unknownchinesenumbervector"
, "-unknownchineseyearvector", "-unkWord", "*UNK*", "-transformMatrixType", "DIAGONAL", "-scalingForInit", double.ToString(TrainOptions.DefaultScalingForInit), "-trainWordVectors" }));
Sharpen.Collections.AddAll(argsWithDefaults, Arrays.AsList(args));
args = Sharpen.Collections.ToArray(argsWithDefaults, new string[argsWithDefaults.Count]);
for (int argIndex = 0; argIndex < args.Length;)
{
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-parser"))
{
parserPath = args[argIndex + 1];
argIndex += 2;
}
else
{
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-testTreebank"))
{
Pair <string, IFileFilter> treebankDescription = ArgUtils.GetTreebankDescription(args, argIndex, "-testTreebank");
argIndex = argIndex + ArgUtils.NumSubArgs(args, argIndex) + 1;
testTreebankPath = treebankDescription.First();
testTreebankFilter = treebankDescription.Second();
}
else
{
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-treebank"))
{
Pair <string, IFileFilter> treebankDescription = ArgUtils.GetTreebankDescription(args, argIndex, "-treebank");
argIndex = argIndex + ArgUtils.NumSubArgs(args, argIndex) + 1;
trainTreebankPath = treebankDescription.First();
trainTreebankFilter = treebankDescription.Second();
}
else
{
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-cachedTrees"))
{
cachedTrainTreesPath = args[argIndex + 1];
argIndex += 2;
}
else
{
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-runGradientCheck"))
{
runGradientCheck = true;
argIndex++;
}
else
{
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-train"))
{
runTraining = true;
argIndex++;
}
else
{
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-model"))
{
modelPath = args[argIndex + 1];
argIndex += 2;
}
else
{
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-nofilter"))
{
filter = false;
argIndex++;
}
else
{
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-continueTraining"))
{
runTraining = true;
filter = false;
initialModelPath = args[argIndex + 1];
argIndex += 2;
}
else
{
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-resultsRecord"))
{
resultsRecordPath = args[argIndex + 1];
argIndex += 2;
}
else
{
unusedArgs.Add(args[argIndex++]);
}
}
}
}
}
}
}
}
}
}
}
if (parserPath == null && modelPath == null)
{
throw new ArgumentException("Must supply either a base parser model with -parser or a serialized DVParser with -model");
}
if (!runTraining && modelPath == null && !runGradientCheck)
{
throw new ArgumentException("Need to either train a new model, run the gradient check or specify a model to load with -model");
}
string[] newArgs = Sharpen.Collections.ToArray(unusedArgs, new string[unusedArgs.Count]);
Edu.Stanford.Nlp.Parser.Dvparser.DVParser dvparser = null;
LexicalizedParser lexparser = null;
if (initialModelPath != null)
{
lexparser = ((LexicalizedParser)LexicalizedParser.LoadModel(initialModelPath, newArgs));
DVModel model = GetModelFromLexicalizedParser(lexparser);
dvparser = new Edu.Stanford.Nlp.Parser.Dvparser.DVParser(model, lexparser);
}
else
{
if (runTraining || runGradientCheck)
{
lexparser = ((LexicalizedParser)LexicalizedParser.LoadModel(parserPath, newArgs));
dvparser = new Edu.Stanford.Nlp.Parser.Dvparser.DVParser(lexparser);
}
else
{
if (modelPath != null)
{
lexparser = ((LexicalizedParser)LexicalizedParser.LoadModel(modelPath, newArgs));
DVModel model = GetModelFromLexicalizedParser(lexparser);
dvparser = new Edu.Stanford.Nlp.Parser.Dvparser.DVParser(model, lexparser);
}
}
}
IList <Tree> trainSentences = new List <Tree>();
IdentityHashMap <Tree, byte[]> trainCompressedParses = Generics.NewIdentityHashMap();
if (cachedTrainTreesPath != null)
{
foreach (string path in cachedTrainTreesPath.Split(","))
{
IList <Pair <Tree, byte[]> > cache = IOUtils.ReadObjectFromFile(path);
foreach (Pair <Tree, byte[]> pair in cache)
{
trainSentences.Add(pair.First());
trainCompressedParses[pair.First()] = pair.Second();
}
log.Info("Read in " + cache.Count + " trees from " + path);
}
}
if (trainTreebankPath != null)
{
// TODO: make the transformer a member of the model?
ITreeTransformer transformer = BuildTrainTransformer(dvparser.GetOp());
Treebank treebank = dvparser.GetOp().tlpParams.MemoryTreebank();
treebank.LoadPath(trainTreebankPath, trainTreebankFilter);
treebank = treebank.Transform(transformer);
log.Info("Read in " + treebank.Count + " trees from " + trainTreebankPath);
CacheParseHypotheses cacher = new CacheParseHypotheses(dvparser.parser);
CacheParseHypotheses.CacheProcessor processor = new CacheParseHypotheses.CacheProcessor(cacher, lexparser, dvparser.op.trainOptions.dvKBest, transformer);
foreach (Tree tree in treebank)
{
trainSentences.Add(tree);
trainCompressedParses[tree] = processor.Process(tree).second;
}
//System.out.println(tree);
log.Info("Finished parsing " + treebank.Count + " trees, getting " + dvparser.op.trainOptions.dvKBest + " hypotheses each");
}
if ((runTraining || runGradientCheck) && filter)
{
log.Info("Filtering rules for the given training set");
dvparser.dvModel.SetRulesForTrainingSet(trainSentences, trainCompressedParses);
log.Info("Done filtering rules; " + dvparser.dvModel.numBinaryMatrices + " binary matrices, " + dvparser.dvModel.numUnaryMatrices + " unary matrices, " + dvparser.dvModel.wordVectors.Count + " word vectors");
}
//dvparser.dvModel.printAllMatrices();
Treebank testTreebank = null;
if (testTreebankPath != null)
{
log.Info("Reading in trees from " + testTreebankPath);
if (testTreebankFilter != null)
{
log.Info("Filtering on " + testTreebankFilter);
}
testTreebank = dvparser.GetOp().tlpParams.MemoryTreebank();
testTreebank.LoadPath(testTreebankPath, testTreebankFilter);
log.Info("Read in " + testTreebank.Count + " trees for testing");
}
// runGradientCheck= true;
if (runGradientCheck)
{
log.Info("Running gradient check on " + trainSentences.Count + " trees");
dvparser.RunGradientCheck(trainSentences, trainCompressedParses);
}
if (runTraining)
{
log.Info("Training the RNN parser");
log.Info("Current train options: " + dvparser.GetOp().trainOptions);
dvparser.Train(trainSentences, trainCompressedParses, testTreebank, modelPath, resultsRecordPath);
if (modelPath != null)
{
dvparser.SaveModel(modelPath);
}
}
if (testTreebankPath != null)
{
EvaluateTreebank evaluator = new EvaluateTreebank(dvparser.AttachModelToLexicalizedParser());
evaluator.TestOnTreebank(testTreebank);
}
log.Info("Successfully ran DVParser");
}
public virtual void ExecuteOneTrainingBatch(IList <Tree> trainingBatch, IdentityHashMap <Tree, byte[]> compressedParses, double[] sumGradSquare)
{
Timing convertTiming = new Timing();
convertTiming.Doing("Converting trees");
IdentityHashMap <Tree, IList <Tree> > topParses = CacheParseHypotheses.ConvertToTrees(trainingBatch, compressedParses, op.trainOptions.trainingThreads);
convertTiming.Done();
DVParserCostAndGradient gcFunc = new DVParserCostAndGradient(trainingBatch, topParses, dvModel, op);
double[] theta = dvModel.ParamsToVector();
switch (Minimizer)
{
case (1):
{
//maxFuncIter = 10;
// 1: QNMinimizer, 2: SGD
QNMinimizer qn = new QNMinimizer(op.trainOptions.qnEstimates, true);
qn.UseMinPackSearch();
qn.UseDiagonalScaling();
qn.TerminateOnAverageImprovement(true);
qn.TerminateOnNumericalZero(true);
qn.TerminateOnRelativeNorm(true);
theta = qn.Minimize(gcFunc, op.trainOptions.qnTolerance, theta, op.trainOptions.qnIterationsPerBatch);
break;
}
case 2:
{
//Minimizer smd = new SGDMinimizer(); double tol = 1e-4; theta = smd.minimize(gcFunc,tol,theta,op.trainOptions.qnIterationsPerBatch);
double lastCost = 0;
double currCost = 0;
bool firstTime = true;
for (int i = 0; i < op.trainOptions.qnIterationsPerBatch; i++)
{
//gcFunc.calculate(theta);
double[] grad = gcFunc.DerivativeAt(theta);
currCost = gcFunc.ValueAt(theta);
log.Info("batch cost: " + currCost);
// if(!firstTime){
// if(currCost > lastCost){
// System.out.println("HOW IS FUNCTION VALUE INCREASING????!!! ... still updating theta");
// }
// if(Math.abs(currCost - lastCost) < 0.0001){
// System.out.println("function value is not decreasing. stop");
// }
// }else{
// firstTime = false;
// }
lastCost = currCost;
ArrayMath.AddMultInPlace(theta, grad, -1 * op.trainOptions.learningRate);
}
break;
}
case 3:
{
// AdaGrad
double eps = 1e-3;
double currCost = 0;
for (int i = 0; i < op.trainOptions.qnIterationsPerBatch; i++)
{
double[] gradf = gcFunc.DerivativeAt(theta);
currCost = gcFunc.ValueAt(theta);
log.Info("batch cost: " + currCost);
for (int feature = 0; feature < gradf.Length; feature++)
{
sumGradSquare[feature] = sumGradSquare[feature] + gradf[feature] * gradf[feature];
theta[feature] = theta[feature] - (op.trainOptions.learningRate * gradf[feature] / (System.Math.Sqrt(sumGradSquare[feature]) + eps));
}
}
break;
}
default:
{
throw new ArgumentException("Unsupported minimizer " + Minimizer);
}
}
dvModel.VectorToParams(theta);
}
/// <summary>
/// An example of a command line is
/// <br />
/// java -mx1g edu.stanford.nlp.parser.dvparser.CacheParseHypotheses -model /scr/horatio/dvparser/wsjPCFG.nocompact.simple.ser.gz -output cached9.simple.ser.gz -treebank /afs/ir/data/linguistic-data/Treebank/3/parsed/mrg/wsj 200-202
/// <br />
/// java -mx4g edu.stanford.nlp.parser.dvparser.CacheParseHypotheses -model ~/scr/dvparser/wsjPCFG.nocompact.simple.ser.gz -output cached.train.simple.ser.gz -treebank /afs/ir/data/linguistic-data/Treebank/3/parsed/mrg/wsj 200-2199 -numThreads 6
/// <br />
/// java -mx4g edu.stanford.nlp.parser.dvparser.CacheParseHypotheses -model ~/scr/dvparser/chinese/xinhuaPCFG.ser.gz -output cached.xinhua.train.ser.gz -treebank /afs/ir/data/linguistic-data/Chinese-Treebank/6/data/utf8/bracketed 026-270,301-499,600-999
/// </summary>
/// <exception cref="System.IO.IOException"/>
public static void Main(string[] args)
{
string parserModel = null;
string output = null;
IList <Pair <string, IFileFilter> > treebanks = Generics.NewArrayList();
int dvKBest = 200;
int numThreads = 1;
for (int argIndex = 0; argIndex < args.Length;)
{
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-dvKBest"))
{
dvKBest = System.Convert.ToInt32(args[argIndex + 1]);
argIndex += 2;
continue;
}
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-parser") || args[argIndex].Equals("-model"))
{
parserModel = args[argIndex + 1];
argIndex += 2;
continue;
}
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-output"))
{
output = args[argIndex + 1];
argIndex += 2;
continue;
}
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-treebank"))
{
Pair <string, IFileFilter> treebankDescription = ArgUtils.GetTreebankDescription(args, argIndex, "-treebank");
argIndex = argIndex + ArgUtils.NumSubArgs(args, argIndex) + 1;
treebanks.Add(treebankDescription);
continue;
}
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-numThreads"))
{
numThreads = System.Convert.ToInt32(args[argIndex + 1]);
argIndex += 2;
continue;
}
throw new ArgumentException("Unknown argument " + args[argIndex]);
}
if (parserModel == null)
{
throw new ArgumentException("Need to supply a parser model with -model");
}
if (output == null)
{
throw new ArgumentException("Need to supply an output filename with -output");
}
if (treebanks.IsEmpty())
{
throw new ArgumentException("Need to supply a treebank with -treebank");
}
log.Info("Writing output to " + output);
log.Info("Loading parser model " + parserModel);
log.Info("Writing " + dvKBest + " hypothesis trees for each tree");
LexicalizedParser parser = ((LexicalizedParser)LexicalizedParser.LoadModel(parserModel, "-dvKBest", int.ToString(dvKBest)));
CacheParseHypotheses cacher = new CacheParseHypotheses(parser);
ITreeTransformer transformer = DVParser.BuildTrainTransformer(parser.GetOp());
IList <Tree> sentences = new List <Tree>();
foreach (Pair <string, IFileFilter> description in treebanks)
{
log.Info("Reading trees from " + description.first);
Treebank treebank = parser.GetOp().tlpParams.MemoryTreebank();
treebank.LoadPath(description.first, description.second);
treebank = treebank.Transform(transformer);
Sharpen.Collections.AddAll(sentences, treebank);
}
log.Info("Processing " + sentences.Count + " trees");
IList <Pair <Tree, byte[]> > cache = Generics.NewArrayList();
transformer = new SynchronizedTreeTransformer(transformer);
MulticoreWrapper <Tree, Pair <Tree, byte[]> > wrapper = new MulticoreWrapper <Tree, Pair <Tree, byte[]> >(numThreads, new CacheParseHypotheses.CacheProcessor(cacher, parser, dvKBest, transformer));
foreach (Tree tree in sentences)
{
wrapper.Put(tree);
while (wrapper.Peek())
{
cache.Add(wrapper.Poll());
if (cache.Count % 10 == 0)
{
System.Console.Out.WriteLine("Processed " + cache.Count + " trees");
}
}
}
wrapper.Join();
while (wrapper.Peek())
{
cache.Add(wrapper.Poll());
if (cache.Count % 10 == 0)
{
System.Console.Out.WriteLine("Processed " + cache.Count + " trees");
}
}
System.Console.Out.WriteLine("Finished processing " + cache.Count + " trees");
IOUtils.WriteObjectToFile(cache, output);
}