The repository contains several Pytorch model implementations for text semantic matching. The implemented models are as followed,
- ESIM Enhanced LSTM for Natural Language Inference
- CAFE Compare, Compress and Propagate: Enhancing Neural Architectures with Alignment Factorization for Natural Language Inference
- RE2 Simple and Effective Text Matching with Richer Alignment Features
cd setup
bash setup_snli.sh
python train.py --model="esim" # run ESIM model
or
python train.py --model="cafe" # run CAFE model
or
python train.py --model="re2" # run RE2 model
We report the performance of ESIM, CAFE, RE2 in the SNLI dataset
| Models | Accuracy |
|---|---|
| ESIM | 88.325 |
| CAFE | 87.320 |
| RE2 | 88.458 |