pip install git+git://github.com/iesl/grinch.git
pip install git+git://github.com/epfml/sent2vec.git
conda install pytorch==1.2.0 torchvision==0.4.0 cudatoolkit=9.2 -c pytorch
`
python
import nltk
nltk.download('stopwords')
nltk.download('punkt')
`
Set the following files / values:
[DATABASE]
granted_patent_database =
pregrant_database =
[DATABASE]
host =
username =
password =
Download pre-trained models for disambiguation, place in resources folder: Google Drive
python -m pv.disambiguation.inventor.build_assignee_features_sql
python -m pv.disambiguation.inventor.build_coinventor_features_sql
python -m pv.disambiguation.inventor.build_title_map_sql
python -m pv.disambiguation.inventor.build_canopies_sql
wandb sweep bin/inventor/run_all.yaml
wandb agent $sweep_id
# or using slurm
sh bin/launch_sweep.sh $sweep_id
python -m pv.disambiguation.inventor.finalize
This command trains an inventor disambiguation model.
python -m pv.disambiguation.inventor.train_model --training_data data/inventor-train/eval_common_characteristics.train --dev_data data/inventor-train/eval_common_characteristics.dev --max_num_dev_canopies 200
python -m pv.disambiguation.assignee.build_assignee_name_mentions_sql
wandb sweep bin/assignee/run_all.yaml
wandb agent $sweep_id
# or using slurm
sh bin/launch_sweep.sh $sweep_id
python -m pv.disambiguation.assignee.finalize
python -m pv.disambiguation.location.build_assignee_location_canopies
python -m pv.disambiguation.location.build_inventor_location_canopies
python -m pv.disambiguation.location.build_assignee_location_mentions
python -m pv.disambiguation.location.build_inventor_location_mentions
wandb sweep bin/location/run_all.yaml
wandb agent $sweep_id
# or using slurm
sh bin/launch_sweep.sh $sweep_id
python -m pv.disambiguation.location.finalize