About dataset and pre-trained word embedding about sequence_tagging

[guillaumegenthial]

1. Data: if you want to use CoNLL2003 data, you need to download the tag from the conll2003 webpage and the articles from reuters. Everything is explained on the conll2003 webpage. Finally, you need to make sure that your data files are in the format discribed in the README.md
2. Embeddings: you need to download the glove wordvectors (and maybe rename the file location in the config.py file)

Once you have the embeddings and the data files, you run build_data.py that produces the trimmed word glove vectors + the vocabularies (tags, words, characters)
I hope this helps.