BehaviourRecognition Multi-modal Dual 1D-CNN with attention (Used Focal loss not Cross entropy) CNN-LSTM with attention Data is very skewed
