en2cn/item_embedding/all_lang_emb.py

from sentence_transformers import SentenceTransformer, util

from my_config import LANG_EMB_MODEL


model = SentenceTransformer(LANG_EMB_MODEL["all"])
# print("model load time:{}".format(time.time()-s))
# # Our sentences we like to encode


def item2emb_all(items_list):
    """
     items= ["已知集合点集,集合点集,求交集",
                 "求集合点集与集合点集的交集",
                 "求函数根式复合是一次的定义域",
                 "函数是奇函数,在区间上单调递增,求参数的取值范围"]
     :param items_list:
     :return:
    """
    if isinstance(items_list, str):
        items_list = [items_list]
    # # Sentences are encoded by calling model.encode()
    item_embeddings = model.encode(items_list)
    return item_embeddings


if __name__ == '__main__':
    # ss = item2emb_all("已知集合点集,集合点集,求交集")
    # print(ss)
    # print(ss.shape),
    # ss = item2emb_all(["已知集合点集,集合点集,求交集",
    #                    "求集合点集与集合点集的交集",
    #                    "求函数根式复合是一次的定义域",
    #                    "函数是奇函数,在区间上单调递增,求参数的取值范围"])
    ss = item2emb_all(["广泛地阅读", "泛读"])  # ["visual", "scene", "situation"]
    a = util.cos_sim(ss, ss)
    # print(ss.shape)
    # # b = util.cos_sim(ss[:1],ss[1:])
    print(a)
    # print(b)
    # res = similarity(ss)
    # print(res)