A Simple PoC (Proof of Concept) of Hate-speech (Toxic content) Detector API Server using model from detoxify and/or custom traditional machine learning model. Detoxify (unbiased model) achieves AUC score of 93.74% compared to top leaderboard score with AUC 94.73% in Jigsaw Unintended Bias in Toxicity Classification. To those who are interested in training custom machine learning model based on Jigsaw Unintended Bias in Toxicity Classification can take a look at our Jupyter Notebook.
hate-speech-detector-api is a core dependency of nostr-filter-relay.
A demo gradio showcase is available on HuggingFace Spaces - https://huggingface.co/spaces/rifatramadhani/hate-speech-detector. There is no guarantee for the uptime, but feel free to test.
Python 3.9 or Python 3.10 is required to run the app. There is bug/issue for Python 3.11 or higher version affecting detoxify library.
You can start by cloning this repository to run or modify it locally
git clone https://github.com/atrifat/hate-speech-detector-api
cd hate-speech-detector-api
Create virtual environment using venv, pyenv, or conda. This is an example using venv to create and activate the environment:
python3 -m venv venv
source venv/bin/activate
install its dependencies
pip install -U -r requirements.txt
and run it using command
python3 app.py
You can also copy .env.example to .env file and change the environment value based on your needs before running the app.
There is also Dockerfile available if you want to build docker image locally. If you don't want to build docker image locally, you can use the published version in ghcr.io/atrifat/hate-speech-detector-api.
Run it:
docker run --init --env-file .env -p 7860:7860 -it ghcr.io/atrifat/hate-speech-detector-api
or run it in the background (daemon):
docker run --init --env-file .env -p 7860:7860 -it --name hate-speech-detector-api -d ghcr.io/atrifat/hate-speech-detector-api
If you want to test the API server, you can use GUI tools like Postman or using curl.
curl --header "Content-Type: application/json" \
--request POST \
--data '{"api_key":"your_own_api_key_if_you_set_them", "q":"hello world good morning"}' \
http://localhost:7860/predict
The result of classification will be shown as follow (Example: using unbiased-small model):
{
"identity_attack":0.0,
"insult":0.0,
"obscene":0.0,
"severe_toxicity":0.0,
"sexual_explicit":0.0,
"threat":0.0,
"toxicity":0.0010000000474974513
}
MIT License
Copyright (c) 2023 Rif'at Ahdi Ramadhani
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
Rif'at Ahdi Ramadhani (atrifat)