BiMPM

Re-implementation of BIMPM(Bilateral Multi-Perspective Matching for Natural Language Sentences)

Training

• On Snli data: bash script/run_SNLI.py
• On Chinese sentence Similarity task such as BQ_corpus: bash script/run_BQ.py. BQ_corpus is not public data set, you can apply from here: http://icrc.hitsz.edu.cn/info/1037/1162.html
• On other data: Modify the code from data_loader.py and other related code.

Evaluation

Accuracy on Snli data set（%）：

data set	Accuracy（%）
dev set	85.74
test set	84.91

Accuracy on the article is 86.0(test set).

Prediction

Upload soon!