Cross_Utterance_Clean

This directory contains different LSTM-based cross-utterance language models (LMs)

---

1. Train/pre-train the first-level LM using vanillaLSTM.sh The model will be saved in Flvmodel directory
2. Train the cross-utterance LM using jointatten.sh The model will be saved in atten_<exp_no> directory