/**
* Creates an index of the top size words based on tf-idf metrics
*
* @param size the number of words in the vocab
* @return the index of the words
* @throws IOException
*/
public Index createVocab(int size) {
Index vocab = new Index();
// bootstrapping
calcWordFrequencies();
// term frequency has every word
for (String word : tf.keySet()) {
double tfVal = MathUtils.tf((int) documentWordFrequencies.getCount(word));
double idfVal = MathUtils.idf(numFiles, idf.getCount(word));
double tfidfVal = MathUtils.tfidf(tfVal, idfVal);
java.util.regex.Matcher m = punct.matcher(word);
if (!stopWords.contains(word) && !m.matches()) tfidf.setCount(word, tfidfVal);
}
Counter<String> aggregate = tfidf;
// keep top size keys via tfidf rankings
aggregate.keepTopNKeys(size - 1);
log.info("Created vocab of size " + aggregate.size());
wordScores = aggregate;
// add words that made it via rankings
for (String word : aggregate.keySet()) {
if (vocab.indexOf(word) < 0) vocab.add(word);
}
// cache the vocab
currVocab = vocab;
return vocab;
}
protected void addForDoc(File doc) {
Set<String> encountered = new HashSet<String>();
SentenceIterator iter = new LineSentenceIterator(doc);
while (iter.hasNext()) {
String line = iter.nextSentence();
if (line == null) continue;
Tokenizer tokenizer = tokenizerFactory.create(new InputHomogenization(line).transform());
while (tokenizer.hasMoreTokens()) {
String token = tokenizer.nextToken();
java.util.regex.Matcher m = punct.matcher(token);
if (validWord(token)) {
documentWordFrequencies.incrementCount(token, doc.getAbsolutePath(), 1.0);
tf.incrementCount(token, 1.0);
if (!encountered.contains(token)) {
idf.incrementCount(token, 1.0);
encountered.add(token);
}
}
}
iter.finish();
}
}
/**
* Train on the corpus
*
* @param rdd the rdd to train
* @return the vocab and weights
*/
public Pair<VocabCache, GloveWeightLookupTable> train(JavaRDD<String> rdd) {
TextPipeline pipeline = new TextPipeline(rdd);
final Pair<VocabCache, Long> vocabAndNumWords = pipeline.process();
SparkConf conf = rdd.context().getConf();
JavaSparkContext sc = new JavaSparkContext(rdd.context());
vocabCacheBroadcast = sc.broadcast(vocabAndNumWords.getFirst());
final GloveWeightLookupTable gloveWeightLookupTable =
new GloveWeightLookupTable.Builder()
.cache(vocabAndNumWords.getFirst())
.lr(conf.getDouble(GlovePerformer.ALPHA, 0.025))
.maxCount(conf.getDouble(GlovePerformer.MAX_COUNT, 100))
.vectorLength(conf.getInt(GlovePerformer.VECTOR_LENGTH, 300))
.xMax(conf.getDouble(GlovePerformer.X_MAX, 0.75))
.build();
gloveWeightLookupTable.resetWeights();
gloveWeightLookupTable.getBiasAdaGrad().historicalGradient =
Nd4j.zeros(gloveWeightLookupTable.getSyn0().rows());
gloveWeightLookupTable.getWeightAdaGrad().historicalGradient =
Nd4j.create(gloveWeightLookupTable.getSyn0().shape());
log.info(
"Created lookup table of size "
+ Arrays.toString(gloveWeightLookupTable.getSyn0().shape()));
CounterMap<String, String> coOccurrenceCounts =
rdd.map(new TokenizerFunction(tokenizerFactoryClazz))
.map(new CoOccurrenceCalculator(symmetric, vocabCacheBroadcast, windowSize))
.fold(new CounterMap<String, String>(), new CoOccurrenceCounts());
List<Triple<String, String, Double>> counts = new ArrayList<>();
Iterator<Pair<String, String>> pairIter = coOccurrenceCounts.getPairIterator();
while (pairIter.hasNext()) {
Pair<String, String> pair = pairIter.next();
counts.add(
new Triple<>(
pair.getFirst(),
pair.getSecond(),
coOccurrenceCounts.getCount(pair.getFirst(), pair.getSecond())));
}
log.info("Calculated co occurrences");
JavaRDD<Triple<String, String, Double>> parallel = sc.parallelize(counts);
JavaPairRDD<String, Tuple2<String, Double>> pairs =
parallel.mapToPair(
new PairFunction<Triple<String, String, Double>, String, Tuple2<String, Double>>() {
@Override
public Tuple2<String, Tuple2<String, Double>> call(
Triple<String, String, Double> stringStringDoubleTriple) throws Exception {
return new Tuple2<>(
stringStringDoubleTriple.getFirst(),
new Tuple2<>(
stringStringDoubleTriple.getFirst(), stringStringDoubleTriple.getThird()));
}
});
JavaPairRDD<VocabWord, Tuple2<VocabWord, Double>> pairsVocab =
pairs.mapToPair(
new PairFunction<
Tuple2<String, Tuple2<String, Double>>, VocabWord, Tuple2<VocabWord, Double>>() {
@Override
public Tuple2<VocabWord, Tuple2<VocabWord, Double>> call(
Tuple2<String, Tuple2<String, Double>> stringTuple2Tuple2) throws Exception {
return new Tuple2<>(
vocabCacheBroadcast.getValue().wordFor(stringTuple2Tuple2._1()),
new Tuple2<>(
vocabCacheBroadcast.getValue().wordFor(stringTuple2Tuple2._2()._1()),
stringTuple2Tuple2._2()._2()));
}
});
for (int i = 0; i < iterations; i++) {
JavaRDD<GloveChange> change =
pairsVocab.map(
new Function<Tuple2<VocabWord, Tuple2<VocabWord, Double>>, GloveChange>() {
@Override
public GloveChange call(
Tuple2<VocabWord, Tuple2<VocabWord, Double>> vocabWordTuple2Tuple2)
throws Exception {
VocabWord w1 = vocabWordTuple2Tuple2._1();
VocabWord w2 = vocabWordTuple2Tuple2._2()._1();
INDArray w1Vector = gloveWeightLookupTable.getSyn0().slice(w1.getIndex());
INDArray w2Vector = gloveWeightLookupTable.getSyn0().slice(w2.getIndex());
INDArray bias = gloveWeightLookupTable.getBias();
double score = vocabWordTuple2Tuple2._2()._2();
double xMax = gloveWeightLookupTable.getxMax();
double maxCount = gloveWeightLookupTable.getMaxCount();
// w1 * w2 + bias
double prediction = Nd4j.getBlasWrapper().dot(w1Vector, w2Vector);
prediction += bias.getDouble(w1.getIndex()) + bias.getDouble(w2.getIndex());
double weight = Math.pow(Math.min(1.0, (score / maxCount)), xMax);
double fDiff =
score > xMax ? prediction : weight * (prediction - Math.log(score));
if (Double.isNaN(fDiff)) fDiff = Nd4j.EPS_THRESHOLD;
// amount of change
double gradient = fDiff;
// update(w1,w1Vector,w2Vector,gradient);
// update(w2,w2Vector,w1Vector,gradient);
Pair<INDArray, Double> w1Update =
update(
gloveWeightLookupTable.getWeightAdaGrad(),
gloveWeightLookupTable.getBiasAdaGrad(),
gloveWeightLookupTable.getSyn0(),
gloveWeightLookupTable.getBias(),
w1,
w1Vector,
w2Vector,
gradient);
Pair<INDArray, Double> w2Update =
update(
gloveWeightLookupTable.getWeightAdaGrad(),
gloveWeightLookupTable.getBiasAdaGrad(),
gloveWeightLookupTable.getSyn0(),
gloveWeightLookupTable.getBias(),
w2,
w2Vector,
w1Vector,
gradient);
return new GloveChange(
w1,
w2,
w1Update.getFirst(),
w2Update.getFirst(),
w1Update.getSecond(),
w2Update.getSecond(),
fDiff);
}
});
JavaRDD<Double> error =
change.map(
new Function<GloveChange, Double>() {
@Override
public Double call(GloveChange gloveChange) throws Exception {
gloveChange.apply(gloveWeightLookupTable);
return gloveChange.getError();
}
});
final Accumulator<Double> d = sc.accumulator(0.0);
error.foreach(
new VoidFunction<Double>() {
@Override
public void call(Double aDouble) throws Exception {
d.$plus$eq(aDouble);
}
});
log.info("Error at iteration " + i + " was " + d.value());
}
return new Pair<>(vocabAndNumWords.getFirst(), gloveWeightLookupTable);
}