A neural machine translation framework based on transformer architecture, with most of the code refered to Li Mu's 《Dive into deep learning》, has completely removed its dependence on D2L.
star to me, if you like it. :)
一个基于transformer架构的神经机器翻译框架，大部分代码借鉴于李沐的动手学深度学习，已经完全去除了对d2l的依赖。（中科大深度学习第五次作业，自己学的时候没有找到合适的代码，故此开源）