PyTorch implementation of the paper "Dialogue Act Classification with Context-Aware Self-Attention" for dialogue act classification with a generic dataset class and PyTorch-Lightning trainer
Very basic question, apologies if I am missing something.
Looking at your data (switchboard_train.csv), I am having a hard time seeing how the utterances are arranged in conversations to achieve the conversation-level context representation. It seems to be a list of utterances, with no hierarchical structure. Using your data generator (DADataset) with this input file yields 193K samples, which is equal to the number of utterances reported for the dataset (Table 2). Is this expected?
Hi there,
Very basic question, apologies if I am missing something.
Looking at your data (switchboard_train.csv), I am having a hard time seeing how the utterances are arranged in conversations to achieve the conversation-level context representation. It seems to be a list of utterances, with no hierarchical structure. Using your data generator (DADataset) with this input file yields 193K samples, which is equal to the number of utterances reported for the dataset (Table 2). Is this expected?