Github user ygcao commented on a diff in the pull request: https://github.com/apache/spark/pull/10152#discussion_r47552249 --- Diff: mllib/src/main/scala/org/apache/spark/mllib/feature/Word2Vec.scala --- @@ -281,17 +294,28 @@ class Word2Vec extends Serializable with Logging { val expTable = sc.broadcast(createExpTable()) val bcVocab = sc.broadcast(vocab) val bcVocabHash = sc.broadcast(vocabHash) - - val sentences: RDD[Array[Int]] = words.mapPartitions { iter => + // each partition is a collection of sentences, will be translated into arrays of Index integer + val sentences: RDD[Array[Int]] = dataset.mapPartitions { sentenceIter => --- End diff -- This is very close to my original version exception for will throw out all words after MAX_SENTENCE_LENGTH, and are you preferring to make the maxSentenceLength static config? The latest version of mine will still try to take use of the rest of sentences for training after cutting by maxSentenceLength. e.g. for a 2200 word long sentence, it will be used as three cut sentences just like the old version except for the last/third sentence from the cut will be 200 words long without words padded from the next sentence. This way, we can maximize the usage of our data with both respecting sentence boundary and sentence length restriction.
--- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org