public class MentionTaggerIterator extends CachedSentenceIterator
| Modifier and Type | Field and Description |
|---|---|
protected Annotation.Source |
source |
docIt, sentItbatchSize, currDocument, cursor, docsInUse, documents, encoders, inputSize, labelSize, log, name, numExamples, randomize, startTime, tagset, totalExamples| Constructor and Description |
|---|
MentionTaggerIterator(Collection<Document> docs,
String name,
EncoderSet encoders,
Class tagset,
Annotation.Source source,
int numExamples,
int batchSize,
boolean randomize)
Create an Iterator for training.
|
MentionTaggerIterator(Collection<Document> docs,
String name,
EncoderSet encoders,
Class tagset,
int numExamples,
int batchSize,
boolean randomize) |
| Modifier and Type | Method and Description |
|---|---|
org.nd4j.linalg.dataset.DataSet |
generateDataSet(ArrayList<Sentence> examples,
int num,
int exampleSize) |
List<Token> |
nextTokens()
Returns the next Sentence in the Dataset as Tokens.
|
asyncSupported, clearCache, clearCachedDocument, encodeDocument, hasNext, hasNextDocument, hasNextSentence, next, nextBatchOfSentences, nextDataSet, nextDataSet, nextDocument, nextSentence, resetbatch, cursor, getDocuments, getInputSize, getLabels, getLabelSize, getPreProcessor, getTagset, inputColumns, next, numExamples, randomizeDocuments, reportProgress, resetSupported, setPreProcessor, totalExamples, totalOutcomesclone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, waitforEachRemaining, removeprotected Annotation.Source source
public MentionTaggerIterator(Collection<Document> docs, String name, EncoderSet encoders, Class tagset, Annotation.Source source, int numExamples, int batchSize, boolean randomize)
docs - the documents to use for training.name - encoders - input encoders to use. Tokens will be encoded and cached on the fly.tagset - the tagset to use for prediction.source - the Annotation source to use for predictionnumExamples - number of examples to train. -1 for complete dataset.batchSize - batch size in sentencesrandomize - randomize examples before training epochpublic MentionTaggerIterator(Collection<Document> docs, String name, EncoderSet encoders, Class tagset, int numExamples, int batchSize, boolean randomize)
public List<Token> nextTokens()
public org.nd4j.linalg.dataset.DataSet generateDataSet(ArrayList<Sentence> examples, int num, int exampleSize)
generateDataSet in class CachedSentenceIteratorCopyright © 2020. All rights reserved.