The code repo for Two Heads Are Better Than One: Exploiting Both Sequence and Graph Models in AMR-To-Text Generation.
cd pre-training
pip install -r requirements.txt
pip install amrlib
change the path in get_data.sh
sh get_data.sh
silver data in silver_data/training/
available
cd fairseq
change the path in preprocess.sh
sh preprocess.sh
cd fairseq
change the path in pre-train.sh
sh pre-train.sh
cd fairseq
change the path in fine-tune.sh
sh fine-tune.sh
cd fairseq
change the path in train.sh
sh train.sh