Investigating Factual Accuracy in Abstractive Summarization Models
pip3 install -r requirements.txt
python3 -m spacy download en_core_web_sm
python3 -m spacy download en_core_web_lg
In data/pointer-generator, download data folder "test_output" from here
ls articles | cat > names_articles.txt
python clean_point-gen_data.py names_articles.txt
In data/bottom-up, download data files from here and here
cp test.txt.src.tagged.shuf.400words articles.txt
cp bottom_up_cnndm_015_threshold.out decoded.txt
cp test.txt.tgt.tagged.shuf.noslash reference.txt
In data/fast-abs-RL, download data folder "acl18_results" from here
ls references | cat > names_references.txt
python clean_fast-abs-RL_data.py names_references.txt