Official Code for "Consistency Training by Synthetic Question Generation for Conversational Question Answering" ACL 2024
Note: I made sure that the code is not buggy right now and that it runs smoothly, but it may also be! Please note that I will add the code for training the Conversational Question Generation Module in the next few days.
For Running CoTaH may simply run (change parameters by your will):
python Cotah_Bert.py --model_name CoTaH-BERT \
--S 2 \
--kl_ratio 2. \
--tau 6 \
--batch_size 6 \
--accumulation_steps 1 \
--seed 1000 \
--dist_type uniform \
--use_sim_threshold True \
--source_directory .
Of course, They have a default value so may simply run:
python Cotah_Bert.py
For Running BERT:
python Bert.py --model_name BERT \
--batch_size 6 \
--accumulation_steps 1 \
--seed 1000 \
--source_directory .