1 paper
Han Guo, Songlin Yang, Tarushii Goel +3
The attention mechanism in Transformers is an important primitive for accurate and scalable sequence modeling. Its quadratic-compute and linear-memory complexity however remain sig…