We've developed a simple meta-learning algorithm called Reptile which works by repeatedly sampling a task, performing stochastic gradient ...
確定! 回上一頁