How to preprocess data to get .npy file?

[trungdvq]

Dear Lee

I was see your MTA-LSTM in old git. But i can't not understant how we can create .npy file from dataset zihu and composite file. Can you help me how to preprocess this data. thank you so much.

[TobiasLee]

Sorry for the late response. The code is not available now, and I hope that you can create the .npy file following the suggestions:

• A detailed description for each .npy file in the issue #1 .
• A text pre-processing notes written by myself, and I did:
  • Build Vocabulary by keeping the most frequent words, convert the other words to UNK
  • Padding the length to max-length
  • Convert the word to index according to the vocabulary built before, and save it as .npy
  • Prepare topic2label dictionary and convert the label to corresponding label indexes.

Hope this can help you, good luck!