You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
For the dataset, you'll just need to keep the data format described in the README (one line -> one word and its tag, new sentence = empty line). Optionally, you can change the way the class Dataset works.
For the word vectors, for now only GloVe is supported, but you could easily write some code to load the Word2Vec vectors. You would only need to adapt the export_trimmed_glove_vectors functions defined in data_utils.py. This function takes the name of the file of word vectors, the vocab (word -> id) and exports a np array E such that E[id] = the word vector. There should be a lot of code available online to perform such a task. You can have a look at gensim.
And can we use a word2vec-pre-trained word vectors?
The text was updated successfully, but these errors were encountered: