Wenpeng Lu, Rui Yu, Shoujin Wang, Can Wang, Ping Jian, Heyan Huang
The paper has been accepted by ACM Transaction on Internet Technology.
python 3.6
numpy==1.16.4
pandas==0.22.0
tensorboard==1.12.0
tensorflow-gpu==1.12.0
keras==2.2.4
gensim==3.0.0
Run 3DSSM.py
python3 3DSSM.py
We used two datasets: BQ & LCQMC.
- "The BQ Corpus: A Large-scale Domain-specific Chinese Corpus For Sentence Semantic Equivalence Identification", https://www.aclweb.org/anthology/D18-1536/.
- "LCQMC: A Large-scale Chinese Question Matching Corpus", https://www.aclweb.org/anthology/C18-1166/.
Due to the differences between the two data sets, some parameters adopted by 3DSSM are different. Therefore, we provide two versions of the code for the two data sets.