Skip to content

Latest commit

 

History

History
54 lines (46 loc) · 1.3 KB

README.md

File metadata and controls

54 lines (46 loc) · 1.3 KB

Experiments with the CoAE baseline

Installation

See INSTALL.md

Setup datasets

cd ${OS2D_ROOT}/baselines/CoAE/data
ln -s ${OS2D_ROOT}/data/grozi grozi
ln -s ${OS2D_ROOT}/data/dairy dairy
ln -s ${OS2D_ROOT}/data/paste paste
ln -s ${OS2D_ROOT}/data/instre instre

Preparations

cd ${OS2D_ROOT}/baselines/CoAE
conda activate os2d
export PYTHONPATH=$OS2D_ROOT:$PYTHONPATH

Train models

# training runs
python experiments/launcher_coae_grozi_train.py
python experiments/launcher_coae_instre_train.py

View training logs

# Convert text logs to the OS2D binary format
python experiments/parse_logs_to_pkl.py --log_path output/grozi
python experiments/parse_logs_to_pkl.py --log_path output/instre

# View in Visdom
python ../../os2d/utils/plot_visdom.py --log_path output/grozi
python ../../os2d/utils/plot_visdom.py --log_path output/instre

Run evaluation

# evaluation of the best models (selection of the best model on the validation set was done manually)
python experiments/launcher_coae_grozi_eval.py
python experiments/launcher_coae_instre_eval.py

View results

# Create tables
# Table 3
python experiments/launcher_coae_grozi_eval_collect.py
# Table 4
python experiments/launcher_coae_instre_eval_collect.py