GAHs

Guided Attention Heads

steps

Put your datset in the folder datasets/your_DataSet_name/your_data_file, then update the load in preprocessor/raw_data_loader.py
Preprocess your data (semantic annotation): python main.py run_mode preprocess --dataset TREC --splits train,test

-- the processed data will be in datasets/your_DataSet_name/** -- splits means how man files you have

Name		Name	Last commit message	Last commit date
Latest commit History 26 Commits
backs		backs
best_models		best_models
config		config
datasets		datasets
models		models
newenv		newenv
preprocessor		preprocessor
.gitignore		.gitignore
Note		Note
README.md		README.md
ablation_large.txt		ablation_large.txt
ablation_record.txt		ablation_record.txt
ablation_record2.txt		ablation_record2.txt
data_helper.py		data_helper.py
data_larger.py		data_larger.py
en2de_main.py		en2de_main.py
main.py		main.py
mask.py		mask.py
pre.sh		pre.sh
sbatchscript.sh		sbatchscript.sh
selection.txt		selection.txt
slurm-1410875.out		slurm-1410875.out
slurm-1410981.out		slurm-1410981.out
slurm-1426120.out		slurm-1426120.out
slurm-1427611.out		slurm-1427611.out
slurm-1428303.out		slurm-1428303.out
temp_ablation_record.txt		temp_ablation_record.txt
temp_record.txt		temp_record.txt
test.py		test.py
train.py		train.py
train_large.py		train_large.py
util.py		util.py