cdparser.py 文件源码

def load_embeddings(self, filename, xz=False):
        if not os.path.isfile(filename):
            print(filename, "does not exist")
            return self

        if xz:
            f = lzma.open(filename, "rt", encoding="utf-8", errors="ignore")
        else:
            f = open(filename, "r")
        found_set = set()
        for line in f:
            l = line.split()
            word = strong_normalize(l[0])
            vec = [float(x) for x in l[1:]]
            if word in self._vocab:
                found_set.add(word)
                self._word_lookup.init_row(self._vocab[word], vec)
        f.close()
        print("Loaded embeddings from", filename)
        print(len(found_set), "hits with vocab size of", len(self._vocab))

        return self