In the paper on fasttext for supervised classification, the authors specified various quantities of hidden units by altering some parameter (h is the one on pages 3,4 - In table 1 you see "It has 10 hidden units and we evaluate it with and without bigrams.") But after reading the documentation it does not appear that there is a "hidden unit" parameter to alter. Is there a way to specify the number of hidden units? Or is this the same as specifying the -dim option?
1 Answers
Related