similarities 实现了多种相似度计算、匹配搜索算法，支持文本、图像，python3开发。
安装
python
pip3 install torch # conda install pytorch
pip3 install -U similarities或
python
git clone https://github.com/shibing624/similarities.git
cd similarities
python3 setup.py install报错
ChineseCLIPProcessor
Traceback (most recent call last): File "xx\similarity_test1.py",
line 9, in
from similarities import BertSimilarity File "xx\lib\site-packages\similarities_init _.py", line 28, in
from similarities.clip_similarity import ClipSimilarity File "xx\lib\site-packages\similarities\clip_similarity.py", line 16, in
from similarities.clip_module import ClipModule File "xx\lib\site-packages\similarities\clip_module.py", line 18, in
from transformers import ChineseCLIPProcessor, ChineseCLIPModel, CLIPProcessor, CLIPModel ImportError: cannot import name
'ChineseCLIPProcessor' from 'transformers'
(xx\lib\site-packages\transformers_init_.py)
报这个错的原因是transformers版本太低，升级下版本就可以了。
python
pip install --upgrade transformerspydantic
另外还缺少pydantic：
python
pip install pydantic样例
python
# -*- coding: utf-8 -*-
"""
@author:XuMing(xuming624@qq.com)
@description: 文本语义相似度计算和文本匹配搜索
"""
import sys
sys.path.append('..')
from similarities import BertSimilarity
# 1.Compute cosine similarity between two sentences.
sentences = ['如何更换花呗绑定银行卡',
'花呗更改绑定银行卡']
corpus = [
'花呗更改绑定银行卡',
'我什么时候开通了花呗',
'俄罗斯警告乌克兰反对欧盟协议',
'暴风雨掩埋了东北部；新泽西16英寸的降雪',
'中央情报局局长访问以色列叙利亚会谈',
'人在巴基斯坦基地的炸弹袭击中丧生',
]
model = BertSimilarity(model_name_or_path="shibing624/text2vec-base-chinese")
print(model)
similarity_score = model.similarity(sentences[0], sentences[1])
print(f"{sentences[0]} vs {sentences[1]}, score: {float(similarity_score):.4f}")
print('-' * 50 + '\n')
# 2.Compute similarity between two list
similarity_scores = model.similarity(sentences, corpus)
print(similarity_scores.numpy())
for i in range(len(sentences)):
for j in range(len(corpus)):
print(f"{sentences[i]} vs {corpus[j]}, score: {similarity_scores.numpy()[i][j]:.4f}")
print('-' * 50 + '\n')
# 3.Semantic Search
model.add_corpus(corpus)
res = model.most_similar(queries=sentences, topn=3)
print(res)
for q_id, id_score_dict in res.items():
print('query:', sentences[q_id])
print("search top 3:")
for corpus_id, s in id_score_dict.items():
print(f'\t{model.corpus[corpus_id]}: {s:.4f}')
print('-' * 50 + '\n')
print(model.search(sentences[0], topn=3))结果：
Similarity: BertSimilarity, matching_model: <SentenceModel: shibing624/text2vec-base-chinese, encoder_type: MEAN, max_seq_length: 256, emb_dim: 768>
2024-03-07 20:12:46.481 | DEBUG | text2vec.sentence_model:__init__:80 - Use device: cpu
如何更换花呗绑定银行卡 vs 花呗更改绑定银行卡, score: 0.8551
--------------------------------------------------
[[0.8551465 0.72119546 0.14502521 0.21666759 0.25171342 0.08089039]
[0.9999997 0.6807433 0.17136583 0.21621695 0.27282682 0.12791349]]
如何更换花呗绑定银行卡 vs 花呗更改绑定银行卡, score: 0.8551
如何更换花呗绑定银行卡 vs 我什么时候开通了花呗, score: 0.7212
如何更换花呗绑定银行卡 vs 俄罗斯警告乌克兰反对欧盟协议, score: 0.1450
如何更换花呗绑定银行卡 vs 暴风雨掩埋了东北部；新泽西16英寸的降雪, score: 0.2167
如何更换花呗绑定银行卡 vs 中央情报局局长访问以色列叙利亚会谈, score: 0.2517
如何更换花呗绑定银行卡 vs 人在巴基斯坦基地的炸弹袭击中丧生, score: 0.0809
花呗更改绑定银行卡 vs 花呗更改绑定银行卡, score: 1.0000
花呗更改绑定银行卡 vs 我什么时候开通了花呗, score: 0.6807
花呗更改绑定银行卡 vs 俄罗斯警告乌克兰反对欧盟协议, score: 0.1714
花呗更改绑定银行卡 vs 暴风雨掩埋了东北部；新泽西16英寸的降雪, score: 0.2162
花呗更改绑定银行卡 vs 中央情报局局长访问以色列叙利亚会谈, score: 0.2728
花呗更改绑定银行卡 vs 人在巴基斯坦基地的炸弹袭击中丧生, score: 0.1279
--------------------------------------------------
2024-03-07 20:13:03.429 | INFO | similarities.bert_similarity:add_corpus:108 - Start computing corpus embeddings, new docs: 6
Batches: 100%|██████████| 1/1 [00:10<00:00, 10.45s/it]
2024-03-07 20:13:13.889 | INFO | similarities.bert_similarity:add_corpus:120 - Add 6 docs, total: 6, emb len: 6
{0: {0: 0.8551465272903442, 1: 0.7211954593658447, 4: 0.25171342492103577}, 1: {0: 0.9999997019767761, 1: 0.6807432770729065, 4: 0.27282682061195374}}
query: 如何更换花呗绑定银行卡
search top 3:
花呗更改绑定银行卡: 0.8551
我什么时候开通了花呗: 0.7212
中央情报局局长访问以色列叙利亚会谈: 0.2517
query: 花呗更改绑定银行卡
search top 3:
花呗更改绑定银行卡: 1.0000
我什么时候开通了花呗: 0.6807
中央情报局局长访问以色列叙利亚会谈: 0.2728
--------------------------------------------------
{0: {0: 0.8551465272903442, 1: 0.7211954593658447, 4: 0.25171342492103577}}
相关链接
https://github.com/shibing624/similarities
https://huggingface.co/shibing624/text2vec-base-chinese
Compute similarity score Demo
Semantic Search Demo