paraphrase_detection_pytorch

training

CUDA_VISIBLE_DEVICES=0 python train.py train --ddir /path/to/data/ --data-cache-dir /path/to/cache/ --savedir /path/to/dump/ --bsize 128 --ft_path /path/to/ft/bin --use_cuda --epoch 10 --lr 0.01 --seed 1111

evaluation

CUDA_VISIBLE_DEVICES=0 python eval.py eval --ddir /path/to/data --data-cache-dir /dir/to/cache --model_path ./test/model.pth --ft-path /path/to/fasttext.bin --use_cuda

reference

swem.py: https://github.com/yagays/swem