[SOLVED] Ask About The "default" Size of Vocabulary in Word2Vec in Deeplearning4j Library

Ask About The "default" Size of Vocabulary in Word2Vec in Deeplearning4j Library

I am currently learning about this library: Word2Vec from Deeplearning4j (Homepage, Github)

Following is the example usage of the method:

//build Word2Vec model
Word2Vec vec = new Word2Vec.Builder()
                .layerSize(100)
                .windowSize(5)
                .stopWords(stopList)
                .tokenizerFactory(t)
                .learningRate(0.025)
                .build();

I know that I can limit the vocabulary size with this method:

vec.limitVocabularySize(100) //limit the vocab size as 100

Above example is the command if I want to limit the vocab size into 100

My question:
Could anyone inform me what is the default size of the vocab (i.e., if I do not set the limit)?

Best,

Solution

By default there is no limit. That means it will add all words it finds to the vocabulary.

Also note, the examples you linked to are over 4 years old. I suggest you use the official examples: https://github.com/eclipse/deeplearning4j-examples