A smart assistant that is capable of taking actions through voice.
It features running a background process for wake-word detection, with samples for training the model based on my voice. The model architecture is a hybrid between CNN and GRU, where the CNN is for feature extraction, and the GRU is for sequenctial features.