A Tensorflow implementation of the transformer deep learning model, based on the research paper "Attention Is All You Need" (Vaswani et al., 2017). The official implementation can be found here.
The implementation follows the model architecture from the original work:
The attention mechanism is shown as below:
