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
When Kaggle trains the algorithm it only goes through 5 epochs due to earlystopping. Unfortunately, this means that it's accuracy is often sub-70 on SwDa datasets. Is it possible for kaggle to create longer checkpoints?
When Kaggle trains the algorithm it only goes through 5 epochs due to earlystopping. Unfortunately, this means that it's accuracy is often sub-70 on SwDa datasets. Is it possible for kaggle to create longer checkpoints?