Anwen Hu

I ran "./bin/preprocess.sh ./conf/ontonotes/bilstm-viterbi-doc.conf" to preprocess the ontonotes data set.

@strubell , thanks for your reply. These are log files of data preprocessing and training. [preprocess_ontonotes_doc.log](https://github.com/iesl/dilated-cnn-ner/files/2464086/preprocess_ontonotes_doc.log) [train_ontonotes_lstm_crf_doc.log](https://github.com/iesl/dilated-cnn-ner/files/2464087/train_ontonotes_lstm_crf_doc.log)

From my understanding, "--layers" is set for ID-CNN. It should be OK to be set null in bilstm-vertibi-doc.conf.

> I encoded each article at the word level. It's simply the average of the glove embeddings of the words in the article. so there is only an article vector...

> Yep! Just one article vector and no attention. And we were able to (slightly) beat the ones reported Goodnews. Thanks! Wonderful work! And I have one more question: Are...

Thanks!

How about the index of the "DOCSTART" ? 0?

I get it, thanks for your reply !

The detailed information about word embedding is introduced in Ma's paper(Ma X, Hovy E. End-to-end Sequence Labeling via Bi-directional LSTM-CNNs-CRF[J]. 2016.). It writes in the paper that Standford's Glove 100...