The evaluation dataset is in Chinese, and we used the same language model RoBERTa Large 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.
BAAI/bge-large-zh	78.61	-	-	-	-	-
BAAI/bge-large-zh-v1.5	79.07	-	-	-	-	-
hellonlp/simcse-large-zh	81.32	-	-	-	-	-
hellonlp/promcse-large-zh	81.63	-	-	-	-	-

Uses

To use the tool, first install the promcse package from PyPI

pip install promcse

After installing the package, you can load our model by two lines of code

python

from promcse import PromCSE
model = PromCSE("hellonlp/promcse-bert-large-zh", "cls", 10)

Then you can use our model for encoding sentences into embeddings

python

embeddings = model.encode("武汉是一个美丽的城市。")
print(embeddings.shape)
#torch.Size([1024])

Compute the cosine similarities between two groups of sentences

python

sentences_a = ['你好吗']
sentences_b = ['你怎么样','我吃了一个苹果','你过的好吗','你还好吗','你',
               '你好不好','你好不好呢','我不开心','我好开心啊', '你吃饭了吗',
               '你好吗','你现在好吗','你好个鬼']
similarities = model.similarity(sentences_a, sentences_b)
print(similarities)
# [(1.0, '你好吗'),
#  (0.9324, '你好不好'),
#  (0.8945, '你好不好呢'),
#  (0.8845, '你还好吗'),
#  (0.8382, '你现在好吗'),
#  (0.8072, '你过的好吗'),
#  (0.7648, '你怎么样'),
#  (0.6736, '你'),
#  (0.5706, '你吃饭了吗'),
#  (0.5417, '你好个鬼'),
#  (0.3747, '我好开心啊'),
#  (0.0777, '我不开心'),
#  (0.0624, '我吃了一个苹果')]