A tool to help people avoid online sexual harassment by filtering out unwanted comments from Facebook or Twitter.
tensorflow 1.12.2, keras 2.2.4, numpy, pandas, flask, matplotlib, scikit-learn
To create an environment in conda run: conda create --name stop-it tensorflow keras numpy pandas flask matplotlib scikit-learn pytest
python model.py
or using flask
python deploy.py and in the browser
127.0.0.1:5000/predict?sentence="your sentence to analyse" (without quotation marks). Be sure to include spaces in the senteces. Spaces can be explicity set using the UTF-8 encoded "%20".
To work on the browser extension run: cd extension npm install
And to run the test: npm test
For further details about the project and topic as a whole, please see the wiki
- The Problem : Outlines the issue
- The Solution : Describes the browser extension that we are aiming to build and it's target functionality
- Flow in TensorFlow : The current workings of the model
- The Language of Misogyny and Sexual Aggression : A discussion of the language of cyber sexual harassment/sexist hate speech and the surrounding literature
Icons from : http://www.aha-soft.com/iconsets.htm