Modelpublic

swtx/Erlangshen-Roberta-110M-Similarity

sourceHugging Faceapache-2.0updated 4y agoView on Hugging Face

1likes14downloads

Model Card

Erlangshen-Roberta-110M-Similarity, model (Chinese)，one model of Fengshenbang-LM.

We collect 20 paraphrace datasets in the Chinese domain for finetune, with a total of 2773880 samples. Our model is mainly based on roberta

Usage

python

from transformers import BertForSequenceClassification
from transformers import BertTokenizer
import torch

tokenizer=BertTokenizer.from_pretrained('IDEA-CCNL/Erlangshen-Roberta-110M-Similarity')
model=BertForSequenceClassification.from_pretrained('IDEA-CCNL/Erlangshen-Roberta-110M-Similarity')

texta='今天的饭不好吃'
textb='今天心情不好'

output=model(torch.tensor([tokenizer.encode(texta,textb)]))
print(torch.nn.functional.softmax(output.logits,dim=-1))

Scores on downstream chinese tasks（The dev datasets of BUSTM and AFQMC may exist in the train set）

Model	BQ	BUSTM	AFQMC
Erlangshen-Roberta-110M-Similarity	85.41	95.18	81.72
Erlangshen-Roberta-330M-Similarity	86.21	99.29	93.89
Erlangshen-MegatronBert-1.3B-Similarity	86.31	-	-

Citation

If you find the resource is useful, please cite the following website in your paper.

@misc{Fengshenbang-LM,
  title={Fengshenbang-LM},
  author={IDEA-CCNL},
  year={2021},
  howpublished={\url{https://github.com/IDEA-CCNL/Fengshenbang-LM}},
}