embedding.py 文件源码

def closest(self, w, n=10):
        """
        Assumes the vectors have been normalized.
        """
        scores = self.m.dot(self.represent(w))
        return heapq.nlargest(n, zip(scores, self.iw))