hellonlp/simcse-roberta-base-zh

SimCSE(sup)

Data List

The following datasets are all in Chinese.

Data	size(train)	size(valid)	size(test)
ATEC	62477	20000	20000
BQ	100000	10000	10000
LCQMC	238766	8802	12500
PAWSX	49401	2000	2000
STS-B	5231	1458	1361
SNLI	146828	2699	2618
MNLI	122547	2932	2397

Model List

The evaluation dataset is in Chinese, and we used the same language model RoBERTa base on different methods. In addition, considering that the test set of some datasets is small, which may lead to a large deviation in evaluation accuracy, the evaluation data here uses train, valid and test at the same time, and the final evaluation result adopts the weighted average (w-avg) method.

Model	STS-B(w-avg)	ATEC	BQ	LCQMC	PAWSX	Avg.
BERT-Whitening	65.27	-	-	-	-	-
SimBERT	70.01	-	-	-	-	-
SBERT-Whitening	71.75	-	-	-	-	-
BAAI/bge-base-zh	78.61	-	-	-	-	-
hellonlp/simcse-base-zh(sup)	80.96	-	-	-	-	-

Uses

You can use our model for encoding sentences into embeddings

import torch
from transformers import BertTokenizer
from transformers import BertModel
from sklearn.metrics.pairwise import cosine_similarity

# model
simcse_sup_path = "hellonlp/simcse-roberta-base-zh"
tokenizer = BertTokenizer.from_pretrained(simcse_sup_path)
MODEL = BertModel.from_pretrained(simcse_sup_path)

def get_vector_simcse(sentence):
    """
    预测simcse的语义向量。
    """
    input_ids = torch.tensor(tokenizer.encode(sentence)).unsqueeze(0)
    output = MODEL(input_ids)
    return output.last_hidden_state[:, 0].squeeze(0)

embeddings = get_vector_simcse("武汉是一个美丽的城市。")
print(embeddings.shape)
#torch.Size([768])

You can also compute the cosine similarities between two sentences

def get_similarity_two(sentence1, sentence2):
    vec1 = get_vector_simcse(sentence1).tolist()
    vec2 = get_vector_simcse(sentence2).tolist()
    similarity_list = cosine_similarity([vec1], [vec2]).tolist()[0][0]
    return similarity_list

sentence1 = '你好吗'
sentence2 = '你还好吗'
result = get_similarity_two(sentence1,sentence2)
print(result) #0.7996
#(1.0, '你好吗')
#(0.8247, '你好不好')
#(0.8217, '你现在好吗')
#(0.7976, '你还好吗')
#(0.7918, '你好不好呢')
#(0.712, '你过的好吗')
#(0.6986, '你怎么样')
#(0.6693, '你')
#(0.5442, '你好个鬼')
#(0.4516, '你吃饭了吗')
#(0.4, '我好开心啊')
#(0.29, '我不开心')
#(0.2782, '我吃了一个苹果')