Fast MoE ===
An easy-to-use but efficient implementation of the Mixture of Experts (MoE) model for PyTorch
