Skip to content
Branch: master
Find file History

Latest commit

Fetching latest commit…
Cannot retrieve the latest commit at this time.

Files

Permalink
Type Name Latest commit message Commit time
..
Failed to load latest commit information.
10_cased_unredadtions_with_Bernstein.txt
9_cased_unredactions.md
Mueller_Report_UnRedacted_with_BERT_simplified_incomplete.ipynb
README.md
__init__.py
download_bert.sh
find_redactions.py
load_and_predict.py
requirements.txt
uncased_unredactions.md

README.md

Muellerbot

Unredacts the mueller report (inaccurately) using BERT.

Quickstart

git clone https://github.com/manceps/tfw
cd tfw
conda env create --name tfw --file environment.yml
conda activate tfw
pip install -e .
cd examples/muellerbot
source download_bert.sh
python load_and_predict.py

Usage

At the Text: You can type or paste any text containing unk tokens as the redaction markers and muellerbot will try to fill in the blanks. At the Redaction marker: You can type or paste any short text without spaces, like '[MASK]' or '[HOM]' to be used as the marker. Markers are assumed to hide a single word. So your text should contain multiple contiguous markers (like "unk unk unk") to predict/unredact multiple words. If you used unk (the default redaction marker) then you can just hit enter.

You can’t perform that action at this time.